Semanlink - Search

Parents:

Related Tags:

Descendants

Cognitive Search
- Azure Cognitive Search
- Goldfire
- Sinequa
Concept Search
Desktop search
Enterprise Search
- Goldfire
- Sinequa
  Nirant sur Twitter : "1/ As vector databases continue raising funds, let's reevaluate the landscape. Scale is important, but FAISS did billion-vectors in 2017. What do we use them for? Private Search Engines..."
  Tags:
  > reason we need Search vs fine-tuned models: freshness and cost > - Challenges in private search; > 1. Too few sources: Need great recall > 2. In-domain context > Most users find off-the-shelf OpenAI embeddings & cosine-ranking to be subpar beyond initial wow factor. > To improve search, we can: > > 1. Fine-tune off-the-shelf embeddings within domain > 2. Incorporate lexical (keyword) search e.g. BM25 > 3. Utilize hybrid approaches like Reciprocal Rank Fusion faiss vs Pinecone, Qdrant, Weaviate, Chroma, Vespa, @LangchainAI... Elastic
  
  2023-04-08 About
  Sinequa Brings the Power of ChatGPT to the Enterprise With Neural Search
  Tags:
  > Sinequa’s GPT summarizes the information gathered from Sinequa’s Neural Search into more rapidly digestible and reusable formats tailored to the specific needs of the employee. Sinequa’s GPT also allows an interactive dialogue, so that employees can ask deeper questions, refine the search, or refine the response. In this way, employees can converse with their content and have dialogue with their data, all using natural language.
  
  2023-04-07 About
  Unlocking the Power of Vector Search in Enterprise
  Tags:
  > we've developed a method for fine-tuning embeddings to the unique language of our clients
  
  2023-02-17 About
  Chau Tran sur Twitter : "Some "in the trenches" learnings from integrating vector search into an enterprise search system..."
  Tags:
  Blog post: [Unlocking the Power of Vector Search in Enterprise](doc:2023/02/unlocking_the_power_of_vector_s) > 1. As of Feb 2023, open source text embedding models on @huggingface (E5-large, Instructor-XL, and MPNet) are > to other commercial providers > 2. on out-of-domain data (enterprise search being an extreme case of this)... finetuning embedding models extremely helpful > 3. Vector search, while helpful, is not the whole story! We still need traditional keyword search and personalization
  
  2023-02-17 About
  Sinequa : Enterprise Search Platform
  Tags:
  2021-05-19 About
  The Enterprise Search Market – What should be on your radar? Keynote IKS 2012 Workshop | IKS Blog – The Semantic CMS Community
  Tags:
  2012-07-04 About
Information retrieval
- BEIR
- Cognitive Search
- Course Material on IR
- Cross-Modal Retrieval
- Embeddings in Information Retrieval
  - DSSM (Deep Semantic Similarity Model)
- Hybrid Search
  - Vespa
- Information retrieval: techniques
  - Embeddings in Information Retrieval
  - Probabilistic relevance model
    - BM25
    - TF-IDF
  - Query Expansion
  - Ranking (information retrieval)
    - BM25
    - Discounted cumulative gain
    - Google ranking
      - PageRank
        
        Linking Entities with Knowledge Graphs | by Sigurd Berglann | Strise | Medium
        
        Tags:
        
        Entity linking
        Knowledge Graphs
        PageRank
        Random walk
        
        2021-02-05 About
        
        Article recommendation with Personalized PageRank and Full Text Search
        
        Tags:
        
        Neo4j
        PageRank
        Recommender Systems
        
        2019-02-28 About
        
        #PageRank-based #Wikidata autocomplete powered by #Apache #Solr
        
        Tags:
        
        PageRank
        Solr
        Solr - autocomplete
        Wikidata
        
        2017-10-21 About
    - Learning to rank
      - Ranking SVM
        
        A Ranking Approach to Keyphrase Extraction - Microsoft Research (2009)
        
        Tags:
        
        Keyword/keyphrase extraction
        Learning to rank
        Microsoft Research
        NLP@Microsoft
        Ranking SVM
        
        Previously, automatic keyphrase extraction was formalized as classification and learning methods for classification were utilized. This paper points out that it is more essential to **cast the keyphrase extraction problem as ranking** and employ a **learning to rank** method to perform the task. As example, it employs Ranking SVM, a state-of-art method of learning to rank, in keyphrase extraction
        
        2017-06-27 About
      - TextRank
        
        Efficient unsupervised keywords extraction using graphs
        
        Tags:
        
        Neo4j
        TextRank
        
        2017-10-04 About
        
        What is a simple but detailed explanation of Textrank? - Quora
        
        Tags:
        
        TextRank
        
        2017-07-12 About
        
        How does Textrank work? (slides)
        
        Tags:
        
        TextRank
        
        2017-07-12 About
        
        TextRank: Bringing Order into Texts (2004)
        
        Tags:
        
        Rada Mihalcea
        TextRank
        
        paper, description of the TextRank algorithm
        
        2017-06-14 About
        
        Automated Keyword Extraction – TF-IDF, RAKE, and TextRank (Less Than Dot - Blog)
        
        Tags:
        
        Keyword/keyphrase extraction
        NLP sample code
        Python sample code
        RAKE
        TextRank
        
        2017-06-13 About
        
        Text Summarization with Gensim
        
        Tags:
        
        gensim
        TextRank
        Text Summarization
        
        2017-06-02 About
    - Probabilistic relevance model
    - Ranked Entities in Search Results
    - Reranking
    - Unsupervised ranker
  - Retriever
    - Dense retriever
    - Retriever-Reader
      - BLINK
        
        [1911.03814] Scalable Zero-shot Entity Linking with Dense Entity Retrieval
        
        Tags:
        
        Arxiv Doc
        BERT-based entity linking
        BLINK
        Entity Retrieval
        NLP@Facebook
        Retriever-Reader
        Zero-shot Entity Linking
        
        > a two stage approach, based on fine-tuned BERT architectures. In the first stage, we do retrieval in a dense space defined by a bi-encoder that independently embeds the mention context and the entity descriptions (Humeau et al., 2019; Gillick et al., 2019). Each retrieved candidate is then examined more carefully with a cross-encoder that concatenates the mention and entity text,
        
        2020-05-02 About
        
        Sebastian Riedel sur Twitter : Happy to introduce BLINK, the @facebookai open-source entity linker!...
        
        Tags:
        
        BLINK
        Tweet
        
        > BLINK outperforms previous linkers, is easy to use and very fast
        
        2020-05-02 About
      - RAG (Retrieval-Augmented Generation)
        LLM Fine-tuning vs RAG
        
        Knowledge Graphs & LLMs: Fine-Tuning Vs. Retrieval-Augmented Generation | by Tomaz Bratanic | Neo4j Developer Blog (June 6 2023)
        
        Tags:
        
        LLM + KG
        LLM Fine-tuning vs RAG
        LM Fine-tuning
        Neo4j
        RAG (Retrieval-Augmented Generation)
        RAG + KG
        
        > What are the limitations of LLMs, and how to overcome them? > We haven’t found any recent articles describing **how to use a knowledge graph to prepare good question-answer pairs that can be used to fine-tune an LLM**. This is an area that we plan to explore during the [NaLLM project](doc:2023/08/github_neo4j_nallm_repositor) > we recommend updating the internal knowledge of an LLM through fine-tuning techniques only for slowly changing or updating data.
        
        2023-08-13 About
        
        RAG: Agents
        
        Jerry Liu sur X : "4 Levels of Agents for RAG..."
        
        Tags:
        
        Jerry Liu
        RAG: Agents
        RAG (Retrieval-Augmented Generation)
        Twitter thread
        
        2024-01-23 About
        
        RAG: chunking
        
        Nils Reimers sur X : "Embeddings can store only 1 aspect/topic per embedding well."
        
        Tags:
        
        Nils Reimers
        RAG: chunking
        Text Embeddings
        Tips
        Wikipedia
        
        > On Wikipedia, one paragraph typically focuses on one topic. So this gives you a good chunking for Wikipeda
        
        2024-03-13 About
        
        Jerry Liu sur X : "To better augment LLMs with context, it makes a lot of sense to organize context not just as a flat list of text chunks, but as a hierarchy of high-level to low-level details. RAPTOR..."
        
        Tags:
        
        Jerry Liu
        RAG: chunking
        RAPTOR
        Twitter thread
        
        > To better augment LLMs with context, it makes a lot of sense to organize context not just as a flat list of text chunks, but as a hierarchy of high-level to low-level details. > > RAPTOR is a super simple but neat idea towards this direction. Hierarchically cluster and summarize the text into a tree (the clustering is important, allows semantically related concepts to be grouped together and doesn't purely rely on spatial positioning!). During query-time dynamically retrieve the most relevant context to the question.
        
        2024-03-03 About
        
        LlamaIndex: "Instead of using a global fixed chunk size for RAG, try splitting based on the semantics of the text"
        
        Tags:
        
        Jerry Liu
        LlamaIndex
        RAG: chunking
        Tweet
        
        2024-01-13 About
        
        Jerry Liu sur X : "Naive chunking strategies cause poor RAG performance...."
        
        Tags:
        
        Jerry Liu
        RAG: chunking
        Tweet
        
        > **“small-to-big” chunking/retrieval strategies**
        
        2023-12-06 About
        
        Rohan sur X : "smaller chunks are good for capturing semantic meaning and larger ones are good for providing better context. @llama_index AutoMergingRetriever takes it one step further..."
        
        Tags:
        
        LlamaIndex
        RAG: chunking
        Tree structure
        Tweet
        
        2023-09-30 About
        
        Jerry Liu sur X : "A simple trick to improve retrieval for RAG 💡: Embed “references” to each text chunk instead of the chunk itself (e.g. smaller chunks, summaries)..."
        
        Tags:
        
        Jerry Liu
        LlamaIndex
        RAG: chunking
        RAG (Retrieval-Augmented Generation)
        Tweet
        
        2023-09-06 About
        
        Jerry Liu sur X : "This might be the first time ChatGPT (+@jxnlco) helped us come up with a better retrieval algorithm for RAG..."
        
        Tags:
        
        ChatGPT
        Jerry Liu
        RAG: chunking
        RAG (Retrieval-Augmented Generation)
        
        <https://gpt-index.readthedocs.io/en/latest/examples/retrievers/auto_merging_retriever.html> 1. Create a hierarchy/graph of “parent chunks” -> smaller chunks. Also link adjacent chunks together. 2. During query-time, first retrieve smaller chunks with embedding similarity. 3. Merge leaves: If any subset of these chunks is a major portion of a larger chunk, return the parent chunk instead.
        
        2023-08-28 About
        
        Jerry Liu sur Twitter : "Some critical data considerations that you must take into account to make your LLM application production-ready"
        
        Tags:
        
        Jerry Liu
        LLM
        RAG: chunking
        RAG (Retrieval-Augmented Generation)
        Slides
        Tweet
        
        > Using naive RAG techniques (naive text chunking, simple top-k retrieval -> LLM) is fine for hackathons, but will lead to lots of failure cases. [slides](https://docs.google.com/presentation/d/1wTEt3sy7ZHk3rYO3nFYhPZEFrfpG70l6WzY12wIaycE/edit#slide=id.p) among the points: - Good parser - Augmenting chunks with context. Eg. keeping page num with chunk allows for inline citation - Right indexes over your data - Using LLMs for Automatic Metadata Extraction
        
        2023-07-23 About
        
        RAG + KG
        
        Enhancing RAG-based application accuracy by constructing and leveraging knowledge graphs
        
        Tags:
        
        LangChain
        Neo4j
        RAG + KG
        
        2024-03-16 About
        
        GraphRAG: Unlocking LLM discovery on narrative private data - Microsoft Research
        
        Tags:
        
        LLM generated Knowledge Graphs
        Microsoft Research
        NLP@Microsoft
        RAG + KG
        
        > GraphRAG uses **LLM generated knowledge graphs** to provide substantial improvements in question-and-answer performance when conducting document analysis of complex information. > power of **prompt augmentation** when performing **discovery** on private datasets (data that the LLM is not trained on and has never seen before, such as an enterprise’s proprietary research, business documents..) > GraphRAG uses the LLM to **create a knowledge graph based on the private dataset**. This graph is then used alongside graph machine learning to perform **prompt augmentation** at query time. > the GraphRAG approach [can] **discover entities in the query**. This allows the LLM to ground itself in the graph and results in superior answer that contains provenance through links to the original supporting text GraphRAG can answer queries such as "**what are the top five themes in the data?**"
        
        2024-03-07 About
        
        Knowledge Graph RAG Query Engine - LlamaIndex
        
        Tags:
        
        LlamaIndex
        LLM + KG
        RAG (Retrieval-Augmented Generation)
        RAG + KG
        
        > Typically, this is to build context based on entities’ SubGraph related to the task.
        
        2023-08-17 About
        
        Jerry Liu on X: “Graph RAG” in @llama_index: a new method of augmenting LLMs with context from a graph database
        
        Tags:
        
        Graph database
        Jerry Liu
        LlamaIndex
        LLM + KG
        RAG (Retrieval-Augmented Generation)
        RAG + KG
        Tweet
        
        [Knowledge Graph RAG Query Engine - LlamaIndex](doc:2023/08/knowledge_graph_rag_query_engin)
        
        2023-08-17 About
        
        Knowledge Graphs & LLMs: Fine-Tuning Vs. Retrieval-Augmented Generation | by Tomaz Bratanic | Neo4j Developer Blog (June 6 2023)
        
        Tags:
        
        LLM + KG
        LLM Fine-tuning vs RAG
        LM Fine-tuning
        Neo4j
        RAG (Retrieval-Augmented Generation)
        RAG + KG
        
        > What are the limitations of LLMs, and how to overcome them? > We haven’t found any recent articles describing **how to use a knowledge graph to prepare good question-answer pairs that can be used to fine-tune an LLM**. This is an area that we plan to explore during the [NaLLM project](doc:2023/08/github_neo4j_nallm_repositor) > we recommend updating the internal knowledge of an LLM through fine-tuning techniques only for slowly changing or updating data.
        
        2023-08-13 About
        
        RAG evaluation
        
        Jerry Liu sur X : "evaluating RAG: purely evaluating retrieval metrics (MRR, precision) isn’t the whole picture - you need end-to-end response evals..."
        
        Tags:
        
        Jerry Liu
        RAG evaluation
        Tweet
        
        2023-09-26 About
        
        RAG isn't easy
        
        Vahid Panjganj sur X : "The problem with RAG is that the retrieval is based on words used in the query. For instance, if I vectorise a cook book. And then ask for recipes for a lactose-intolerant person..."
        
        Tags:
        
        RAG isn't easy
        
        2024-01-13 About
        
        WalkingRAG
        
        Hrishi sur X : "RAPTOR is... one of the very few [RAG architectures] that actively presumes and uses the structure in a document...."
        
        Tags:
        
        Document structure
        RAPTOR
        Twitter thread
        WalkingRAG
        
        (thread by the person of [WalkingRAG](tag:walkingrag)) > The similarities between WalkingRAG and RAPTOR are that both attempt to capture relationships in the data into a higher structure using LLMs... This is a tree in RAPTOR's case, with WalkingRAG it's a graph.
        
        2024-03-09 About
        
        Hrishi sur X : "WalkingRAG is finally out!..."
        
        Tags:
        
        Demo
        UI
        WalkingRAG
        
        2024-03-09 About
        
        LlamaIndex sur X : "Save Memory (and Money) in RAG pipeline with @Cohere 's Int8 and Binary Embeddings..."
        
        Tags:
        
        Cohere.ai
        Embedding Quantization
        Money
        RAG (Retrieval-Augmented Generation)
        Vector database
        
        2024-03-30 About
        
        LlamaIndex 🦙 sur X : "An emerging technique to better represent your data for RAG/LLM applications is to only chunk the data, but also hierarchically cluster and index it..."
        
        Tags:
        
        Clustering
        LlamaIndex
        RAG (Retrieval-Augmented Generation)
        RAPTOR
        Tweet
        
        2024-03-30 About
        
        LlamaIndex 🦙 sur X : "RAFT - Retrieval Augmented Fine Tuning..."
        
        Tags:
        
        Fine-tuning
        LlamaIndex
        RAG (Retrieval-Augmented Generation)
        Tweet
        
        2024-03-28 About
        
        Command-R: RAG at Production Scale
        
        Tags:
        
        Cohere.ai
        LLM
        RAG (Retrieval-Augmented Generation)
        
        2024-03-13 About
        
        RAG makes LLMs better and equal | Pinecone
        
        Tags:
        
        LLM
        Pinecone
        RAG (Retrieval-Augmented Generation)
        
        > The study demonstrates that RAG significantly improves LLM performance, **even on questions within their training domain**. > RAG could enable smaller, less costly, or private models to deliver high-quality results in tasks requiring simple factual reasoning
        
        2024-03-13 About
        
        What you should know about RAG (from beginner to advanced) | by Jonathan Nguyen | Medium
        
        Tags:
        
        RAG (Retrieval-Augmented Generation)
        
        2024-03-11 About
        
        RAG CLI - LlamaIndex
        
        Tags:
        
        LlamaIndex
        RAG (Retrieval-Augmented Generation)
        Tools
        
        CLI tool to ingest local files into a local vector database that is then used for a Chat Q&A repl within your terminal.
        
        2024-03-10 About
        
        Akshay 🚀 sur X : "Let's build a "Chat with your code" RAG application, step-by-step"
        
        Tags:
        
        LlamaIndex
        LLM + Code
        RAG (Retrieval-Augmented Generation)
        Sample code
        Twitter thread
        
        2024-03-09 About
        
        How to Build a RAG System With LlamaIndex, OpenAI, and MongoDB Vector Database | MongoDB
        
        Tags:
        
        MongoDB
        RAG (Retrieval-Augmented Generation)
        
        2024-03-03 About
        
        Ravi Theja sur X : "𝐀𝐜𝐭𝐢𝐯𝐞𝐑𝐀𝐆: 𝐑𝐞𝐯𝐞𝐚𝐥𝐢𝐧𝐠 𝐭𝐡𝐞 𝐓𝐫𝐞𝐚𝐬𝐮𝐫𝐞𝐬 𝐨𝐟 𝐊𝐧𝐨𝐰𝐥𝐞𝐝𝐠𝐞 𝐯𝐢𝐚 𝐀𝐜𝐭𝐢𝐯𝐞 𝐋𝐞𝐚𝐫𝐧𝐢𝐧𝐠..."
        
        Tags:
        
        LLMs and basic Renault Knowledge
        LM + KG
        RAG (Retrieval-Augmented Generation)
        
        2024-02-25 About
        
        Jerry Liu sur X : "a big step towards better RAG... is to just have a really nice PDF parser. It’s so important because a good parser unlocks way more interesting indexing/retrieval strategies…"
        
        Tags:
        
        Chat over your PDFs
        Jerry Liu
        pdf parsing
        RAG (Retrieval-Augmented Generation)
        
        2024-02-23 About
        
        Jerry Liu sur X : "Build a RAG backend over any website in a single CLI command ..."
        
        Tags:
        
        Jerry Liu
        LlamaIndex
        RAG (Retrieval-Augmented Generation)
        Tweet
        
        2024-02-03 About
        
        bclavie/RAGatouille
        
        Tags:
        
        Benjamin Clavié
        ColBERT
        GitHub project
        RAG (Retrieval-Augmented Generation)
        RAGatouille
        
        > RAGatouille's purpose is make it easy to use state-of-the-art methods in your RAG pipeline, without having to worry about the details or the years of literature! At the moment, RAGatouille focuses on making ColBERT simple to use. [Using ColBERT in-memory: Index-Free Encodings & Search](https://github.com/bclavie/RAGatouille/blob/0.0.5b1/examples/06-index_free_use.ipynb) ``` from ragatouille import RAGPretrainedModel RAG = RAGPretrainedModel.from_pretrained "colbert-ir/colbertv2. 0" ) # Your documents, a plain old list of chunked strings. documents = [...] # In-memory indexing supports metadata too! meta = ['attribute': ' really cool value'}...] # All the magic happens here RAG.encode documents, document_metadatas=meta) # Query your in-memory index RAG. search_encoded_docs(query = "A great question", k=3) # All further encode() calls add to the existing documents... RAG.encode(extra_documents, document_metadatas=extra_meta) # ... until you clear them RAG.clear_encoded ```
        
        2024-01-26 About
        
        LlamaIndex 🦙 sur X : "Use RAG to build advanced text-to-SQL..."
        
        Tags:
        
        LlamaIndex
        RAG (Retrieval-Augmented Generation)
        Text to SQL
        Twitter thread
        
        2024-01-24 About
        
        Jerry Liu sur X : "4 Levels of Agents for RAG..."
        
        Tags:
        
        Jerry Liu
        RAG: Agents
        RAG (Retrieval-Augmented Generation)
        Twitter thread
        
        2024-01-23 About
        
        Ravi Theja sur X : "Propositions as Chunks to Improve RAG Performance with @llama_index..."
        
        Tags:
        
        LlamaIndex
        RAG (Retrieval-Augmented Generation)
        Tweet
        
        2024-01-02 About
        
        Full-stack RAG with Jina Embeddings v2 and LlamaIndex
        
        Tags:
        
        LlamaIndex
        RAG (Retrieval-Augmented Generation)
        Sample code
        
        2023-12-23 About
        
        [2312.10997] Retrieval-Augmented Generation for Large Language Models: A Survey
        
        Tags:
        
        Arxiv Doc
        RAG (Retrieval-Augmented Generation)
        Survey / Review
        
        2023-12-23 About
        
        Jerry Liu sur X : "There's different ways you can parse embedded tables for RAG..."
        
        Tags:
        
        Jerry Liu
        pdf parsing
        RAG (Retrieval-Augmented Generation)
        Tables
        
        2023-12-02 About
        
        LlamaIndex 🦙 sur X : "A big issue with RAG is that stuffing more context ≠ better...
        
        Tags:
        
        LlamaIndex
        RAG (Retrieval-Augmented Generation)
        Tweet
        
        2023-11-07 About
        
        Jerry Liu sur X : "Advanced RAG architectures need to implement dynamic retrieval..."
        
        Tags:
        
        Jerry Liu
        RAG (Retrieval-Augmented Generation)
        
        > A *big* downside with naive top-k RAG is the fact that retrieval is static: > - Always retrieves a fixed number (k) regardless of query > - What if query requires more context (e.g. summarization) or less context (e.g. b/c retriever is bad?)? > dynamic retrieval in two conceptual stages: > 1. Pre-retrieval: decide which retriever is suitable for the use case > 2. Post-retrieval: decide whether to rerank or filter results
        
        2023-10-22 About
        
        Self-RAG: Learning to Retrieve, Generate and Critique through Self-Reflection
        
        Tags:
        
        Allen Institute for AI (AI2)
        Discuté avec Raphaël
        RAG (Retrieval-Augmented Generation)
        Tool-Augmented LLMs
        
        2023-10-20 About
        
        L. Dietz @deeds@mastodon.social @dietz@idf.social sur X : "How to evaluate RAG / RACAS (Retrieval-Augmented-Clustering-Augmented-Summarization) systems for Article Generation..."
        
        Tags:
        
        RAG (Retrieval-Augmented Generation)
        SIGIR
        Text Summarization
        
        2023-10-20 About
        
        Jerry Liu sur X : "Using LLMs for automated decision making (routing) is a huge part of an advanced RAG/agent app But LLMs are expensive and slow... Fine-tune embeddings instead"
        
        Tags:
        
        Jerry Liu
        LlamaIndex
        RAG (Retrieval-Augmented Generation)
        
        2023-10-19 About
        
        [2310.03025] Retrieval meets Long Context Large Language Models
        
        Tags:
        
        Arxiv Doc
        Discuté avec Raphaël
        LMs: context length
        RAG (Retrieval-Augmented Generation)
        Retrieval augmented LM
        
        > Retrieval-augmentation versus long context window, which one is better for downstream tasks? > we find that > - LLM with 4K context window using simple retrieval-augmentation at generation can achieve comparable performance to finetuned LLM with 16K context window via positional interpolation on long context tasks, while taking much less computation. > - retrieval can significantly improve the performance of LLMs regardless of their extended context window sizes.
        
        2023-10-07 About
        
        Jerry Liu sur X : "The idea of “small-to-big” retrieval: - retrieve on smaller pieces, expand into more context for LLM synthesis..."
        
        Tags:
        
        Jerry Liu
        LlamaIndex
        RAG (Retrieval-Augmented Generation)
        
        It looks like the paradigm is turning upside down - shifting from content splitting towards composing content blocks
        
        2023-10-02 About
        
        Rohan sur X : "Multi Document Agent architecture (v0) in @llama_index... allows answering broader set of questions over multiple documents..."
        
        Tags:
        
        LlamaIndex
        RAG (Retrieval-Augmented Generation)
        Tweet
        
        2023-10-02 About
        
        Fine-Tuning Your Embedding Model to Maximize Relevance Retrieval in RAG Pipeline | by Wenqi Glantz | Sep, 2023 | Better Programming
        
        Tags:
        
        Embedding Model Fine-tuning
        RAG (Retrieval-Augmented Generation)
        
        see also [Jerry Liu sur X : "One major way to improve your RAG system is to fine-tune your embedding model"](doc:2023/08/jerry_liu_sur_x_one_major_wa)
        
        2023-09-13 About
        
        How to Optimize Retrieval-Augmented Generation
        
        Tags:
        
        RAG (Retrieval-Augmented Generation)
        
        > We all know that RAG is the killer application for LLMs but did you know that it doesn't work (out of the box)?
        
        2023-09-08 About
        
        Jerry Liu sur X : "A simple trick to improve retrieval for RAG 💡: Embed “references” to each text chunk instead of the chunk itself (e.g. smaller chunks, summaries)..."
        
        Tags:
        
        Jerry Liu
        LlamaIndex
        RAG: chunking
        RAG (Retrieval-Augmented Generation)
        Tweet
        
        2023-09-06 About
        
        Building Performant RAG Applications for Production - LlamaIndex
        
        Tags:
        
        LlamaIndex
        RAG (Retrieval-Augmented Generation)
        
        4 core techniques to improve the performance of RAG pipeline
        
        2023-08-30 About
        
        Jerry Liu sur X : "how do you get a good retrieval dataset w/o spending a ton of money on human labels? Answer: use LLMs"
        
        Tags:
        
        Jerry Liu
        LM evaluation
        RAG (Retrieval-Augmented Generation)
        Tweet
        
        2023-08-30 About
        
        Jerry Liu sur X : "This might be the first time ChatGPT (+@jxnlco) helped us come up with a better retrieval algorithm for RAG..."
        
        Tags:
        
        ChatGPT
        Jerry Liu
        RAG: chunking
        RAG (Retrieval-Augmented Generation)
        
        <https://gpt-index.readthedocs.io/en/latest/examples/retrievers/auto_merging_retriever.html> 1. Create a hierarchy/graph of “parent chunks” -> smaller chunks. Also link adjacent chunks together. 2. During query-time, first retrieve smaller chunks with embedding similarity. 3. Merge leaves: If any subset of these chunks is a major portion of a larger chunk, return the parent chunk instead.
        
        2023-08-28 About
        
        Jerry Liu sur X : "One major way to improve your RAG system is to fine-tune your embedding model"
        
        Tags:
        
        Embedding Model Fine-tuning
        Jerry Liu
        RAG (Retrieval-Augmented Generation)
        Synthetic Data
        Tweet
        Unsupervised Sentence Embedding Learning
        
        > We’ve created a full repo/guide on fine-tuning embeddings over any unstructured text (no labels needed) Hum, // TODO check See [Improving Search Ranking with Few-Shot Prompting of LLMs | Vespa Blog](doc:2023/07/improving_search_ranking_with_f)
        
        2023-08-25 About
        
        Yi Ding -- prod/acc sur X : "First is best for retrieval when using ChatGPT models."
        
        Tags:
        
        ChatGPT
        RAG (Retrieval-Augmented Generation)
        
        2023-08-23 About
        
        Comparing Methods for Structured Retrieval (Auto-Retrieval vs. Recursive Retrieval) - LlamaIndex 🦙 0.8.6
        
        Tags:
        
        LlamaIndex
        RAG (Retrieval-Augmented Generation)
        
        2023-08-23 About
        
        Jerry Liu sur X : "8 key considerations for building *production-grade* LLM apps over your data (RAG)..."
        
        Tags:
        
        Jerry Liu
        LLMs over your data
        RAG (Retrieval-Augmented Generation)
        Tweet
        
        (not everything clear to me - eg. point 2, see ["decoupling embedding representations from raw text chunks"](https://twitter.com/jerryjliu0/status/1693290236363919840?s=20)?)
        
        2023-08-20 About
        
        Knowledge Graph RAG Query Engine - LlamaIndex
        
        Tags:
        
        LlamaIndex
        LLM + KG
        RAG (Retrieval-Augmented Generation)
        RAG + KG
        
        > Typically, this is to build context based on entities’ SubGraph related to the task.
        
        2023-08-17 About
        
        Jerry Liu on X: “Graph RAG” in @llama_index: a new method of augmenting LLMs with context from a graph database
        
        Tags:
        
        Graph database
        Jerry Liu
        LlamaIndex
        LLM + KG
        RAG (Retrieval-Augmented Generation)
        RAG + KG
        Tweet
        
        [Knowledge Graph RAG Query Engine - LlamaIndex](doc:2023/08/knowledge_graph_rag_query_engin)
        
        2023-08-17 About
        
        Knowledge Graphs & LLMs: Fine-Tuning Vs. Retrieval-Augmented Generation | by Tomaz Bratanic | Neo4j Developer Blog (June 6 2023)
        
        Tags:
        
        LLM + KG
        LLM Fine-tuning vs RAG
        LM Fine-tuning
        Neo4j
        RAG (Retrieval-Augmented Generation)
        RAG + KG
        
        > What are the limitations of LLMs, and how to overcome them? > We haven’t found any recent articles describing **how to use a knowledge graph to prepare good question-answer pairs that can be used to fine-tune an LLM**. This is an area that we plan to explore during the [NaLLM project](doc:2023/08/github_neo4j_nallm_repositor) > we recommend updating the internal knowledge of an LLM through fine-tuning techniques only for slowly changing or updating data.
        
        2023-08-13 About
        
        LlamaIndex: Harnessing the Power of Text2SQL and RAG to Analyze Product Reviews | by Ravi Theja | LlamaIndex Blog | Aug, 2023 | Medium
        
        Tags:
        
        LlamaIndex
        LLM+SQL
        RAG (Retrieval-Augmented Generation)
        Text to SQL
        Users’ reviews
        
        > The LLM stack over structured data is typically just text-to-SQL. > > But SQL tables can contain unstructured text. So you can actually combine text-to-SQL with retrieval augmentation for sophisticated LLM QA [src](https://twitter.com/jerryjliu0/status/1690389529897979904?s=20)
        
        2023-08-13 About
        
        Jerry Liu sur Twitter : "Some critical data considerations that you must take into account to make your LLM application production-ready"
        
        Tags:
        
        Jerry Liu
        LLM
        RAG: chunking
        RAG (Retrieval-Augmented Generation)
        Slides
        Tweet
        
        > Using naive RAG techniques (naive text chunking, simple top-k retrieval -> LLM) is fine for hackathons, but will lead to lots of failure cases. [slides](https://docs.google.com/presentation/d/1wTEt3sy7ZHk3rYO3nFYhPZEFrfpG70l6WzY12wIaycE/edit#slide=id.p) among the points: - Good parser - Augmenting chunks with context. Eg. keeping page num with chunk allows for inline citation - Right indexes over your data - Using LLMs for Automatic Metadata Extraction
        
        2023-07-23 About
        
        Jerry Liu sur Twitter : "Using cross-encoding as a reranking step can dramatically speed up LLM inference time AND improve accuracy!"
        
        Tags:
        
        Cross-Encoders
        Jerry Liu
        RAG (Retrieval-Augmented Generation)
        Top-k
        Tweet
        
        (speedup inference, because you can pass less nodes to the context) > We use an [MSMarco SBERT cross-encoder from @huggingface](https://www.sbert.net/docs/pretrained-models/ce-msmarco.html) ``` from sentence_transformers import CrossEncoder model = CrossEncoder('model_name', max_length=512) scores = model.predict([('Query', 'Paragraph1'), ('Query', 'Paragraph2') , ('Query', 'Paragraph3')]) ``` (cf. https://www.sbert.net/docs/pretrained-models/ce-msmarco.html
        
        2023-07-20 About
        
        Jerry Liu sur Twitter : Hot take: if you want to... deliver technical differentiation, you will need to learn LLM development in a “bottoms-up” fashion
        
        Tags:
        
        Jerry Liu
        LangChain
        RAG (Retrieval-Augmented Generation)
        Tweet
        
        > DON’T just use out of the box RAG (e.g. default VectorStoreIndex in @llama_index, RetrieverQAChain in langchain,…
        
        2023-07-16 About
        
        Jerry Liu sur Twitter : "If you’re building “chat over your PDFs” with LLMs, you need to deal with the pesky issue of how to parse embedded tables/diagrams..."
        
        Tags:
        
        Chat over your PDFs
        Jerry Liu
        RAG (Retrieval-Augmented Generation)
        Tables
        Tweet
        
        > Native text splitting + top-k on your tables == bad results! > A nuanced, hierarchical data representation over your PDF can help
        
        2023-07-07 About
        
        Jerry Liu sur Twitter : "The DSP project carries a lot of insights for improving RAG..."
        
        Tags:
        
        DSPy (Demonstrate-Search-Predict)
        J'y étais
        Jerry Liu
        Omar Khattab
        RAG (Retrieval-Augmented Generation)
        Tweet
        Webinar
        
        > - value of few-shot ex’s > - declarative modules > - compile an optimized system with distilled LM’s
        
        2023-06-18 About
        
        Build Industry-Specific LLMs Using Retrieval Augmented Generation | by Skanda Vivek | May, 2023 | Towards Data Science
        
        Tags:
        
        Cosine similarity
        RAG (Retrieval-Augmented Generation)
        
        > Cosine similarity is the typical metric used, but you might want to weight different metrics. Such as including a keyword metric... For example, **you might want to weight contexts that contain the words “abstract” or “summary” when you ask the question to an LLM to summarize a document**.
        
        2023-06-16 About
        
        Retrieval-augmented LMs, and prompting
        
        Tags:
        
        Prompt engineering
        Prompting/In-context learning
        RAG (Retrieval-Augmented Generation)
        Retrieval augmented LM
        Slides
        
        slides CS 685, Spring 2023 Advanced Natural Language Processing, Mohit Iyyer, College of Information and Computer Sciences University of Massachusetts Amherst
        
        2023-06-16 About
        
        Retrieval Augmented Generation (RAG) | Prompt Engineering Guide
        
        Tags:
        
        Prompt engineering
        Prompting/In-context learning
        RAG (Retrieval-Augmented Generation)
        
        good intro to RAG and RAG papers
        
        2023-06-16 About
        
        [2207.06300] Re2G: Retrieve, Rerank, Generate
        
        Tags:
        
        Arxiv Doc
        Discuté avec Raphaël
        Knowledge distillation
        Michael Glass
        Question Answering
        RAG (Retrieval-Augmented Generation)
        REALM
        Retriever-Reader
        Slot filling
        Zero shot
        
        > Recent models such as RAG and REALM have introduced retrieval into conditional generation. These models incorporate neural initial retrieval from a corpus of passages. We build on this line of research, proposing Re2G, which combines both neural initial retrieval and reranking into a BART-based sequence-to-sequence generation. Our reranking approach also permits merging retrieval results from sources with incomparable scores, enabling an ensemble of BM25 and neural initial retrieval. > > To train our system end-to-end, we introduce a novel variation of knowledge distillation to train the initial retrieval, reranker, and generation using only ground truth on the target sequence output. > > Large gains in four diverse tasks: zero-shot slot filling, question answering, fact checking and dialog, with relative gains of 9% to 34% over the previous SotA on the KILT leaderboard. [Code]( ibm/kgi-slot-filling)
        
        2022-07-14 About
        
        [2108.13934] Robust Retrieval Augmented Generation for Zero-shot Slot Filling
        
        Tags:
        
        Arxiv Doc
        Discuté avec Raphaël
        EMNLP 2021
        Hard negative mining
        Knowledge Extraction
        Michael Glass
        NLP@IBM
        RAG (Retrieval-Augmented Generation)
        Retrieval-based language models
        Slot filling
        Zero shot
        
        > "Knowledge Graph Induction", a system for slot filling based on advanced training strategies for both Dense Passage Retrieval (DPR) and Retrieval Augmented Generation (RAG) see [[1909.04120] Span Selection Pre-training for Question Answering](doc:2019/09/_1909_04120_span_selection_pre) (same first author) [GitHub](https://github.com/IBM/kgi-slot-filling)
        
        2022-01-19 About
        
        [2005.11401] Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
        
        Tags:
        
        Arxiv Doc
        Discuté avec Raphaël
        Knowledge-Intensive NLP Tasks
        NLP@Facebook
        RAG (Retrieval-Augmented Generation)
        
        > We introduce RAG models where the parametric memory is a pre-trained seq2seq model and the non-parametric memory is a dense vector index of Wikipedia, accessed with a pre-trained neural retriever. > > [REALM](tag:realm) and ORQA, two recently introduced models that combine masked language models with a differentiable retriever... have only explored open-domain extractive question answering. Here, we bring hybrid parametric and non-parametric memory to the “workhorse of NLP,” i.e. sequence-to-sequence (seq2seq) models. > > RAG models use the input sequence x to retrieve text documents z and use them as additional context when generating the target sequence > > **A key feature of our memory is that it is comprised of raw text rather distributed representations**, which makes the memory both (i) human-readable, lending a form of interpretability to our model, and (ii) human-writable, enabling us to dynamically update the model’s memory by editing the document index
        
        2022-01-19 About
        
        Hugging Face sur Twitter : "Transformers release of the Retrieval-Augmented Generation model in collaboration with @facebookai!"
        
        Tags:
        
        huggingface / transformers
        NLP@Facebook
        Not Encoding Factual Knowledge in Language Model
        RAG (Retrieval-Augmented Generation)
        Retrieval augmented LM
        Tweet
        
        > the **RAG model is trained end-to-end for retrieval-in-the-loop generation**, a new paradigm that allows a model to go find useful information in a text corpus when generating. **No need to try to encode all of that knowledge in a trillion parameters any more ;)**
        
        2021-02-23 About
        
        Retrieval Augmented Generation with Huggingface Transformers and Ray | Distributed Computing with Ray
        
        Tags:
        
        huggingface / transformers
        RAG (Retrieval-Augmented Generation)
        Retrieval augmented LM
        
        > Huggingface Transformers recently added the Retrieval Augmented Generation (RAG) model, a new NLP architecture that leverages external documents (like Wikipedia) to augment its knowledge and achieve state of the art results on knowledge-intensive tasks [Hugging Face sur Twitter : "Transformers release of the Retrieval-Augmented Generation model in collaboration with @facebookai!"](doc:2021/02/hugging_face_sur_twitter_tra)
        
        2021-02-11 About
      - Retriever-Ranker
        
        Stanford NLP Group sur Twitter : "YONO: You Only Need One Model for Open-domain Question Answering..."
        
        Tags:
        
        NLP@Stanford
        Open Domain Question Answering
        Retriever-Ranker
        Tweet
        
        2022-12-05 About
    - Sparse retrieval
      - SPLADE
        
        raphaelsty/neural-cherche: Neural Search
        
        Tags:
        
        Cherche (Raphaël)
        ColBERT
        GitHub project
        Neural Search
        Raphaël Sourty
        SPLADE
        
        > a library to fine-tune neural search models such as Splade, ColBERT, and SparseEmbed on a specific dataset
        
        2023-11-17 About
        
        SPLADE for Sparse Vector Search Explained | Pinecone
        
        Tags:
        
        Good
        Pinecone
        SPLADE
        
        2023-09-22 About
        
        SPLADE - a sparse bi-encoder BERT-based model achieves effective and efficient first-stage ranking - Naver Labs Europe
        
        Tags:
        
        SPLADE
        
        > SPLADE learns how to predict irrelevant and missing terms in queries and documents as well as term importance
        
        2023-08-23 About
        
        SPLADE models - Naver Labs Europe
        
        Tags:
        
        SPLADE
        
        2023-08-23 About
        
        raphaelsty/sparsembed: Sparse Information Retrieval with Transformers
        
        Tags:
        
        GitHub project
        Raphaël Sourty
        Sparse+Dense
        Sparse retrieval
        SPLADE
        
        > unofficial replication of the research papers: > - [SPLADE: Sparse Lexical and Expansion Model for First Stage Ranking](doc:2023/05/2107_05720_splade_sparse_lex) > - [SPLADE v2: Sparse Lexical and Expansion Model for Information Retrieval](doc:2023/07/2109_10086_splade_v2_sparse_) > - [SparseEmbed: Learning Sparse Lexical Representations with Contextual Embeddings for Retrieval](doc:2023/07/sparseembed_learning_sparse_le)
        
        2023-08-21 About
        
        [2109.10086] SPLADE v2: Sparse Lexical and Expansion Model for Information Retrieval
        
        Tags:
        
        Arxiv Doc
        Benjamin Piwowarski
        Sparse retrieval
        SPLADE
        
        cf. [[2107.05720] SPLADE: Sparse Lexical and Expansion Model for First Stage Ranking](doc:2023/05/2107_05720_splade_sparse_lex)
        
        2023-07-26 About
        
        [2107.05720] SPLADE: Sparse Lexical and Expansion Model for First Stage Ranking
        
        Tags:
        
        Arxiv Doc
        Benjamin Piwowarski
        Ranking
        Retriever
        Sparse retrieval
        SPLADE
        
        2023-05-18 About
  - Shingles
  - Similarity queries / Vector search
    - Approximate nearest-neighbor
      - faiss
        
        Karl Higley sur Twitter : "Many ANN search tools (e.g. FAISS, ScaNN) allow you to provide multiple points as part of the same query..."
        
        Tags:
        
        faiss
        Neural Search
        Retriever-Reader
        Tweet
        
        > Puzzled why more retrieval models don’t take advantage of this. Give me 100 neighbors of ten points, not 1000 neighbors of one point! (Then score and order them.)
        
        2022-08-20 About
        
        How to Build a Semantic Search Engine With Transformers and Faiss | by Kostas Stathoulopoulos | Towards Data Science
        
        Tags:
        
        faiss
        NLP sample code
        Semantic Search
        
        2022-01-29 About
        
        James Briggs sur Twitter : *free* course on vector similarity search and Faiss..."
        
        Tags:
        
        Cosine similarity
        faiss
        Nearest neighbor search
        Tutorial
        
        2021-10-16 About
        
        Finding similar documents with transformers · Codegram
        
        Tags:
        
        faiss
        huggingface / transformers
        Identification of similar documents
        NLP and Search
        NLP sample code
        
        2020-07-10 About
        
        facebookresearch/faiss: A library for efficient similarity search and clustering of dense vectors.
        
        Tags:
        
        Facebook FAIR
        faiss
        GitHub project
        KD-MKB biblio
        Locality Sensitive Hashing
        Nearest neighbor search
        
        Algorithms that search in sets of vectors of any size, up to ones that possibly do not fit in RAM. Faiss is written in C++ with complete wrappers for Python/numpy. [paper](https://arxiv.org/abs/1702.08734)
        
        2020-06-09 About
      - HNSW
        
        Understanding Hierarchical Navigable Small Worlds (HNSW)
        
        Tags:
        
        HNSW
        
        2024-03-08 About
    - Cosine similarity
    - DSSM (Deep Semantic Similarity Model)
    - Locality Sensitive Hashing
    - Nearest neighbor search
      - Approximate nearest-neighbor
      - ElasticSearch: nearest neighbor(s)
        
        MinHash token filter | Elasticsearch Reference
        
        Tags:
        
        ElasticSearch: nearest neighbor(s)
        MinHash
        
        One solution to make similarity search more practical and computationally feasible involves hashing of documents, in a way that similar documents are more likely to produce the same hash code (locality sensitive hashing, LSH). Depending on what constitutes the similarity between documents, various LSH functions have been proposed. For Jaccard similarity, a popular LSH function is MinHash.
        
        2020-02-14 About
        
        ElastiK Nearest Neighbors
        
        Tags:
        
        ElasticSearch: nearest neighbor(s)
        Locality Sensitive Hashing
        
        Combining Locality-Sensitive Hashing and Elasticsearch for Scalable Online K-Nearest Neighbors Search. [Github](https://github.com/alexklibisz/elastik-nearest-neighbors), [Improved version](https://github.com/alexklibisz/elastiknn)
        
        2020-02-13 About
        
        Text similarity search in Elasticsearch using vector fields | Elastic Blog
        
        Tags:
        
        Approximate nearest-neighbor
        ElasticSearch: nearest neighbor(s)
        NLP and Search
        Text Similarity
        Using word embeddings
        
        > How text embeddings and Elasticsearch’s dense_vector type **could be** used to support similarity search. > In practice, sentence embeddings often don’t generalize well to large sections of text. They are not commonly used to represent text longer than a short paragraph. Example of use: search similar question in a collection of Q/A Sample code is given to rank search results (TensorFlow + Google's universal sentence encoder + cosineSimilarity) Current limitation of vector similarity in Elasticsearch: vectors can be used for scoring documents, but not in the initial retrieval step. ([Ongoing work about approximate nearest neighbours search](https://github.com/elastic/elasticsearch/issues/42326). Will be a licensed feature of ES). > Conclusions: Using vectors for search is an important and **nuanced** area
        
        2020-01-10 About
      - faiss
      - HNSW
      - k-nearest neighbors
        
        k-Nearest Neighbor lookups on embeddings: kNN vs SVM
        
        Tags:
        
        Andrej Karpathy
        GitHub project
        k-nearest neighbors
        Similarity queries / Vector search
        SVM
        
        > A very common workflow is to index some data based on its embeddings and then given a new query embedding retrieve the most similar examples with k-Nearest Neighbor search. > in my experience it **~always works better to use an SVM instead of kNN**, if you can afford the slight computational hit." > can accommodate a number of positives not just one > works because SVM ranking considers the unique aspects of your query w.r.t. data. > [Fun weekend hack: awesome-movies.life...](doc:2023/04/andrej_karpathy_sur_twitter__1) [Tweet](https://twitter.com/karpathy/status/1647025230546886658) ; in [LangChain](https://twitter.com/hwchase17/status/1647328542529843200?s=20)
        
        2023-04-15 About
        
        Uri Alon sur Twitter : "K-nearest neighbors language models (kNN-LMs; @ukhndlwl et al., ICLR'2020) improve the perplexity of standard LMs... Why?"
        
        Tags:
        
        k-nearest neighbors
        Retrieval augmented LM
        Tweet
        
        2023-01-11 About
        
        [2203.08913] Memorizing Transformers
        
        Tags:
        
        Arxiv Doc
        k-nearest neighbors
        Language Model
        LMs: context length
        Memory in deep learning
        NLP@Google
        
        [tweet](https://twitter.com/LiamFedus/status/1522605777961119745?s=20&t=Jt9GBjNcFw6TqeqYvz_BRA): Memorizing Transformers which increases context length up to 262k by an external memory of (keys, values) for that document. - Matches quality of Transformers 5x larger - Can fine-tune a prior pre-trained models to use it > Language models typically need to be trained or finetuned in order to acquire new knowledge, which involves updating their weights. We instead envision language models that can simply read and memorize new data at inference time, thus acquiring new knowledge immediately
        
        2022-05-07 About
        
        A novel multi-label classification algorithm based on K-nearest neighbor and random walk - Zhen-Wu Wang, Si-Kai Wang, Ben-Ting Wan, William Wei Song, 2020
        
        Tags:
        
        k-nearest neighbors
        KNN in Multi-label classification
        Multi-label classification
        Random walk
        
        2020-08-11 About
        
        [1703.03129] Learning to Remember Rare Events
        
        Tags:
        
        Arxiv Doc
        Few-shot learning
        Google Brain
        k-nearest neighbors
        Memory in deep learning
        Rare events
        
        > a large-scale life-long memory module for use in deep learning. The module exploits fast nearest-neighbor algorithms for efficiency and thus scales to large memory sizes. Except for the nearest-neighbor query, the module is fully differentiable and trained end-to-end with no extra supervision. It operates in a life-long manner, i.e., without the need to reset it during training. > Our memory module can be easily added to any part of a supervised neural network
        
        2018-10-23 About
        
        ML-knn: A lazy learning approach to multi-label learning (2007)
        
        Tags:
        
        Frequently cited paper
        k-nearest neighbors
        KNN in Multi-label classification
        Multi-label classification
        NLP 4 Semanlink
        
        a lazy learning approach derived from the traditional [k-nearest neighbors algorithm](tag:k_nearest_neighbors_algorithm) > for each unseen instance, its K nearest neighbors in the training set are firstly identified. After that, based on statistical information gained from the label sets of these neighboring instances, i.e. the number of neighboring instances belonging to each possible class, [Maximum a posteriori (MAP)](tag:maximum_a_posteriori_estimation) principle is utilized to determine the label set for the unseen instance. Implemented in [scikit-multilearn](http://scikit.ml/api/skmultilearn.adapt.mlknn.html), in [java](https://github.com/lefman/mulan-extended/blob/master/mulan/src/mulan/classifier/lazy/MLkNN.java) > the first lazy approach proposed specifically for multi-label classification. This is also a binary relevance approach which considers each label independently as a binary classification problem. Instead of a standard k-NN method, however, MLkNN uses the maximum a-posteriori (MAP) (Kelleher et al., 2015) approach combined with k-NN. [src](https://pdfs.semanticscholar.org/af9b/33da37d290c063cd826ab5923d96892a9767.pdf)
        
        2018-03-18 About
      - KNN in Multi-label classification
        
        SLEEC: Sparse Local Embeddings for Extreme Multi-label Classification (2015)
        
        Tags:
        
        Extreme Multi-label Classification
        KNN in Multi-label classification
        Manik Varma
        NLP@Microsoft
        
        Embedding style **algorithm that performs a clustering of the training points and applies learning procedure in each of the cluster separately**. > The main technical contribution in SLEEC is a formulation for learning a small ensemble of local distance preserving embeddings which can accurately predict infrequently occurring (tail) labels. This allows SLEEC to break free of the traditional low-rank assumption and boost classification accuracy by **learning embeddings which preserve pairwise distances between only the nearest label vectors**. > "Tail labels" : > The critical assumption made by embedding methods, that the training label matrix is low-rank, is violated in almost all real world applications. [Python implementation](https://github.com/xiaohan2012/sleec_python)
        
        2020-08-11 About
        
        A novel multi-label classification algorithm based on K-nearest neighbor and random walk - Zhen-Wu Wang, Si-Kai Wang, Ben-Ting Wan, William Wei Song, 2020
        
        Tags:
        
        k-nearest neighbors
        KNN in Multi-label classification
        Multi-label classification
        Random walk
        
        2020-08-11 About
        
        ML-knn: A lazy learning approach to multi-label learning (2007)
        
        Tags:
        
        Frequently cited paper
        k-nearest neighbors
        KNN in Multi-label classification
        Multi-label classification
        NLP 4 Semanlink
        
        a lazy learning approach derived from the traditional [k-nearest neighbors algorithm](tag:k_nearest_neighbors_algorithm) > for each unseen instance, its K nearest neighbors in the training set are firstly identified. After that, based on statistical information gained from the label sets of these neighboring instances, i.e. the number of neighboring instances belonging to each possible class, [Maximum a posteriori (MAP)](tag:maximum_a_posteriori_estimation) principle is utilized to determine the label set for the unseen instance. Implemented in [scikit-multilearn](http://scikit.ml/api/skmultilearn.adapt.mlknn.html), in [java](https://github.com/lefman/mulan-extended/blob/master/mulan/src/mulan/classifier/lazy/MLkNN.java) > the first lazy approach proposed specifically for multi-label classification. This is also a binary relevance approach which considers each label independently as a binary classification problem. Instead of a standard k-NN method, however, MLkNN uses the maximum a-posteriori (MAP) (Kelleher et al., 2015) approach combined with k-NN. [src](https://pdfs.semanticscholar.org/af9b/33da37d290c063cd826ab5923d96892a9767.pdf)
        
        2018-03-18 About
      - kNN-Transformers
        
        neulab/knn-transformers: PyTorch code for the RetoMaton paper: "Neuro-Symbolic Language Modeling with Automaton-augmented Retrieval" (ICML 2022), including an implementation of kNN-LM and kNN-MT
        
        Tags:
        
        GitHub project
        kNN-Transformers
        
        > a Hugging Face's transformers implementation of k-nearest-neighbor-based language models and machine translation models, designed to be easy and useful in research, and for experimenting with new ideas in kNN-based models. cf. - [[1911.00172] Generalization through Memorization: Nearest Neighbor Language Models](doc:2019/12/_1911_00172_generalization_thr) - [[2201.12431] Neuro-Symbolic Language Modeling with Automaton-augmented Retrieval](doc:2022/07/2201_12431_neuro_symbolic_lan)
        
        2022-07-21 About
        
        Uri Alon sur Twitter : "New repo: kNN-Transformers..."
        
        Tags:
        
        kNN-Transformers
        Tweet
        
        2022-07-21 About
        
        [1911.00172] Generalization through Memorization: Nearest Neighbor Language Models
        
        Tags:
        
        Arxiv Doc
        Dan Jurafsky
        Generalization
        kNN-Transformers
        Knowledge-augmented Language Models
        NLP@Facebook
        NLP@Stanford
        
        extend LMs with nearest neighbor search in embedding space > kNN-LM, an approach that extends a pre-trained LM by linearly interpolating its next word distribution with a k-nearest neighbors (kNN) model > > This approach allows rare patterns to be memorized explicitly, rather than implicitly in model parameters > The kNN-LM involves augmenting such a pre-trained LM with a nearest neighbors retrieval mechanism, without any additional training (the representations learned by the LM remain unchanged). This can be done with a single forward pass over a text collection (potentially including the original LM training set), where the resulting context-target pairs are stored in a key-value datastore that is queried during inference
        
        2019-12-20 About
    - Probabilistic relevance model
    - Semantic Hashing
  - TF-IDF
  - Vector Retrieval
    - Similarity queries / Vector search
      [2401.09350] Foundations of Vector Retrieval
      
      Tags:
      
      Algorithmes
      Arxiv Doc
      Survey / Review
      Vector Retrieval
      
      2024-01-18 About
  - Vector space model
    - gensim
    - Latent Semantic Analysis
    - TF-IDF
      [1910.03524] Beyond Vector Spaces: Compact Data Representation as Differentiable Weighted Graphs
      
      Tags:
      
      Arxiv Doc
      Geometry of language embeddings
      Graphs+Machine Learning
      Poincaré Embeddings
      Vector space model
      Yandex
      
      > In this paper, we aim to eliminate the inductive bias imposed by the embedding space geometry. Namely, we propose to map data into more general non-vector metric spaces: a weighted graph with a shortest path distance. By design, such graphs can model arbitrary geometry with a proper configuration of edges and weights. Our main contribution is PRODIGE (Probabilistic Differentiable Graph Embeddings): a method that learns a weighted graph representation of data end-to-end by gradient descent. [Github](https://github.com/stanis-morozov/prodige)
      
      2019-10-09 About
      [1706.00957] Semantic Vector Encoding and Similarity Search Using Fulltext Search Engines
      
      Tags:
      
      Arxiv Doc
      ElasticSearch
      Semantic Search
      Vector space model
      
      > The ubiquity of semantic vector space modeling raises the challenge of efficient searching in dense, high-dimensional vector spaces. We would naturally want to take advantage of the design and optimizations behind modern fulltext engines like Elasticsearch so as to meet the scalability and robustness demands of modern IR applications. This is the research challenge addressed in this paper. > The paper describes novel ways of encoding dense vectors into text documents, allowing the use of traditional inverted index engines. [blog post](https://rare-technologies.com/semantic-search-fulltext-engine-acl-2017/)
      
      2017-11-11 About
      From Frequency to Meaning: Vector Space Models of Semantics (2010)
      
      Tags:
      
      Good
      Survey / Review
      Vector space model
      
      Good survey of VSMs, of their 3 classes (based either on term-document, word-context, or pair-pattern matrices), and of their applications. A detailed look at a specific open source project in each category.
      
      2017-07-10 About
- Learning sparse representations
  - SPLADE
    SparseEmbed: Learning Sparse Lexical Representations with Contextual Embeddings for Retrieval
    Tags:
    
    Discuté avec Raphaël
    Embeddings in Information Retrieval
    Information retrieval
    Learning sparse representations
    NLP@Google
    SIGIR 2023
    Sparse+Dense
    Sparse retrieval
    retrieval model that learns sparse lexical representations with contextual embeddings > we **combine the strengths of both the sparse and dense representations** for first-stage retrieval. > > Compared with [SPLADE](tag:splade), our model leverages the contextual embeddings to improve model expressiveness. Compared with [ColBERT](tag:colbert), our sparse representations are trained end-to-end to optimize both efficiency and effectiveness.
    
    2023-07-26 About
- Neural Search
  - BEIR
  - Cherche (Raphaël)
  - ColBERT
  - Dense Passage Retrieval
    - Dual encoders (IR)
  - Embeddings in Information Retrieval
  - Extractive Search
    - SPIKE-KBC
  - Haystack
    - Haystack Annotation Tool
  - Hybrid Search
  - Multi-Vector Retrieval
    - ColBERT
      [2211.01267] Multi-Vector Retrieval as Sparse Alignment
      
      Tags:
      
      Arxiv Doc
      Multi-Vector Retrieval
      
      2023-04-07 About
      [2304.01982] Rethinking the Role of Token Retrieval in Multi-Vector Retrieval
      
      Tags:
      
      Arxiv Doc
      ColBERT
      Multi-Vector Retrieval
      Neural Search
      NLP@Google
      
      > Multi-vector retrievers like [ColBERT](tag:colbert) are powerful, but they come at the cost of complicated inference. In this paper, we ask: "can token retrieval alone achieve great performance in multi-vector retrieval?" [tweet](https://twitter.com/leejnhk/status/1643632578824396805?s=20) > The key insight of XTR is that the token-retrieval in multi-vector models should be **trained to retrieve the most salient and informative document tokens**, so that the score between a query and document can be computed using only the retrieved information, just like how single-vector retrieval models work > This is an *amazing* way to re-engineer the scoring mechanism of late interaction / ColBERT retrievers! [src: ColBERT's author Omar Khattab](https://twitter.com/lateinteraction/status/1643439889902637056?s=20) - scoring using only retrieved document terms - imputing missing token scores using their upper bound
      
      2023-04-05 About
  - Neural Ranking Models
  - Pinecone
  - REALM
  - txtai
  - Vespa
  - Weaviate
- NLP based IR
  - Neural Search
    stanfordnlp/dspy: 𝗗𝗦𝗣: Demonstrate-Search-Predict. A framework for composing retrieval and language models for knowledge-intensive NLP.
    Tags:
    
    DSPy (Demonstrate-Search-Predict)
    GitHub project
    Knowledge-Intensive NLP Tasks
    LM + Search
    NLP@Stanford
    NLP based IR
    Omar Khattab
    Prompting/In-context learning
    Retriever-Reader
    (initially called DSP, rebranded as DSPy) > The DSP framework provides a programming abstraction for building grounded AI systems. In a few lines of code, a DSP program expresses rich interactions between retrieval models (RMs) and language models (LMs) to tackle difficult knowledge-intensive NLP tasks (e.g., complex question answering or conversational search). > DSP discourages ["prompt engineering"](tag:prompted_models), which we view much the same way as hyperparameter tuning in traditional ML [@matei_zaharia](https://twitter.com/matei_zaharia/status/1626705622585716737?s=20): >Who are the World Cup champions? I knew ChatGPT would get it wrong when it launched, but it's surprising that all the new search+LLM engines do too. > > **Combining retrieval+LMs won't just be a matter of prompting**. That's why we've been building tools like DSP at Stanford to do it.
    
    2023-02-18 About
    Alex sur Twitter : how you can leverage both @Cohere and @Pinecone libraries to quickly build a POC Arxiv search
    Tags:
    
    Arxiv
    Cohere.ai
    NLP based IR
    Pinecone
    Tweet
    2022-12-01 About
    NLP Based Information Retrieval System | by Nikhil Sharma | Towards Data Science
    Tags:
    
    NLP based IR
    NLP sample code
    2021-01-27 About
- Pseudo relevance feedback
- Ranking (information retrieval)
- RAPTOR
- Search Engines
  - Baidu
    - NLP@Baidu
  - Bing
    - Bing+ChatGPT
  - Cliqz
  - DuckDuckGo
  - ElasticSearch
    - ElasticSearch: annotated text field
    - ElasticSearch: nearest neighbor(s)
      ChatGPT and Elasticsearch: A plugin to use ChatGPT with your Elastic data | Elastic Blog
      
      Tags:
      
      ChatGPT Over Your Data
      ChatGPT plugins
      ElasticSearch
      
      2023-07-07 About
      How to improve Elasticsearch search relevance with boolean queries | Elastic Blog
      
      Tags:
      
      ElasticSearch
      
      2020-12-02 About
      Recherche multilingue grâce à la détection de la langue dans Elasticsearch | Elastic Blog
      
      Tags:
      
      ElasticSearch
      Language Identification
      
      2020-11-26 About
      Construire un bon analyzer français pour Elasticsearch
      
      Tags:
      
      ElasticSearch
      
      2020-10-23 About
      Building a Faster and Accurate Search Engine on Custom Dataset with Transformers 🤗 | by Shivanand Roy | Analytics Vidhya | Sep, 2020 | Medium
      
      Tags:
      
      ElasticSearch
      NLP and Search
      Transformers
      
      2020-10-22 About
      Elastic Transformers. Making BERT stretchy — Scalable… | by Mihail Dungarov | Sep, 2020 | Medium
      
      Tags:
      
      ElasticSearch
      huggingface / transformers
      
      2020-09-08 About
      Searching with Shingles | Elastic Blog
      
      Tags:
      
      ElasticSearch
      Shingles
      
      word n-grams. > Shingles effectively give you the ability to **pre-bake phrase matching**
      
      2020-09-03 About
      Elasticsearch meets BERT: Building Search Engine with Elasticsearch and BERT
      
      Tags:
      
      BERT
      ElasticSearch
      GitHub project
      NLP and Search
      Text Similarity
      
      - Links to [this ES blog post](/doc/2020/01/text_similarity_search_in_elast) - [somewhat related](/doc/2020/01/building_a_search_engine_with_b)
      
      2020-01-10 About
      Meetup Paris #40 : Beyond plain text: elasticsearch’s annotated text field type (en anglais) - YouTube
      
      Tags:
      
      ElasticSearch
      ElasticSearch: annotated text field
      Named Entity Recognition
      
      Elasticsearch has a new field type which allows **structured data to be added into the text of documents as annotations**. This offers powerful new search features: for ex. NER (Named Entity Recognition) tools like OpenNLP and Spacy can be used to tag people, places and organisations mentioned in text. Sample code: [rss indexer](/doc/2019/11/elasticsearch_rss_feed_indexer_)
      
      2019-11-24 About
      Building a Full-Text Search App Using Docker and Elasticsearch
      
      Tags:
      
      Docker
      ElasticSearch
      Sample code
      Text Search
      
      2019-11-10 About
      Supercharging Elasticsearch for extended Knowledge Graph use cases: Data science + business analytics training: Strata Data Conference
      
      Tags:
      
      ElasticSearch
      Giovanni Tummarello
      Knowledge Graphs
      
      2019-05-16 About
      Introducing the Elastic Enterprise Search Beta: Search Everything, Anywhere | Elastic Blog
      
      Tags:
      
      ElasticSearch
      
      2019-05-09 About
      Open Semantic Search: Your own search engine for documents, images, tables, files, intranet & news
      
      Tags:
      
      ElasticSearch
      Semantic Web : Tools
      Semantic Web search engine
      
      2018-04-18 About
      Custom Similarity for ElasticSearch - Algorithms for Big Data
      
      Tags:
      
      ElasticSearch
      Similarity queries / Vector search
      
      2018-04-03 About
      [1706.00957] Semantic Vector Encoding and Similarity Search Using Fulltext Search Engines
      
      Tags:
      
      Arxiv Doc
      ElasticSearch
      Semantic Search
      Vector space model
      
      > The ubiquity of semantic vector space modeling raises the challenge of efficient searching in dense, high-dimensional vector spaces. We would naturally want to take advantage of the design and optimizations behind modern fulltext engines like Elasticsearch so as to meet the scalability and robustness demands of modern IR applications. This is the research challenge addressed in this paper. > The paper describes novel ways of encoding dense vectors into text documents, allowing the use of traditional inverted index engines. [blog post](https://rare-technologies.com/semantic-search-fulltext-engine-acl-2017/)
      
      2017-11-11 About
      Similarity module | Elasticsearch Reference
      
      Tags:
      
      ElasticSearch
      Text Similarity
      
      2017-07-19 About
      Document Similarity Analysis Using ElasticSearch and Python - Data Science Central
      
      Tags:
      
      ElasticSearch
      NLP sample code
      Python sample code
      Text Similarity
      
      2017-07-19 About
      Dealing with Human Language | Elasticsearch: The Definitive Guide [master]
      
      Tags:
      
      ElasticSearch
      NLP tools
      
      2017-07-18 About
      When should we not use Elasticsearch? - Quora
      
      Tags:
      
      ElasticSearch
      
      2017-06-18 About
      Use Elasticsearch in your Java applications
      
      Tags:
      
      ElasticSearch
      Spark (Java web framework)
      Tutorial
      
      2017-05-15 About
      A plugin for language detection in Elasticsearch using Nakatani Shuyo's language detector
      
      Tags:
      
      ElasticSearch
      
      2017-04-28 About
      More Like This Query | Elasticsearch Reference
      
      Tags:
      
      ElasticSearch
      
      2017-04-28 About
      Semantic search with NLP and elasticsearch - Stack Overflow
      
      Tags:
      
      ElasticSearch
      NLP
      Semantic Search
      Stack Overflow
      
      2017-04-28 About
      Locality-sensitive hashing - Elasticsearch - Stack Overflow
      
      Tags:
      
      ElasticSearch
      Locality Sensitive Hashing
      Stack Overflow
      
      2017-04-27 About
      Uncoiling the Data in DNA: Elasticsearch as a BioInformatics Research Tool | Elastic
      
      Tags:
      
      ADN
      Bioinformatics
      ElasticSearch
      
      2017-04-27 About
      Loading JSON-LD into ES - Elasticsearch - Discuss the Elastic Stack
      
      Tags:
      
      ElasticSearch
      JSON-LD
      
      2017-04-21 About
      Serving autocomplete suggestions fast!
      
      Tags:
      
      ElasticSearch
      Solr - autocomplete
      
      2015-06-15 About
      search - Solr vs. ElasticSearch - Stack Overflow
      
      Tags:
      
      ElasticSearch
      Solr
      Stack Overflow
      
      2015-03-14 About
      Apache Solr vs ElasticSearch - the Feature Smackdown!
      
      Tags:
      
      ElasticSearch
      Solr
      
      2015-03-11 About
      ElasticSearch vs Solr
      
      Tags:
      
      ElasticSearch
      Solr
      
      2015-03-11 About
  - Exalead
  - Generative Search Engines
    - ChatGPT+Search
      - Bing+ChatGPT
      - ChatGPT Retrieval Plugin
        
        openai/chatgpt-retrieval-plugin: The ChatGPT Retrieval Plugin
        
        Tags:
        
        ChatGPT Retrieval Plugin
        GitHub project
        
        > The ChatGPT Retrieval Plugin lets you easily search and find personal or work documents by asking questions in everyday language. > uses OpenAI's text-embedding-ada-002 embeddings model to generate embeddings of document chunks, and then stores and queries them using a vector database on the backend. As an open-source and self-hosted solution, developers can deploy their own Retrieval Plugin and register it with ChatGPT. The Retrieval Plugin supports several vector database providers, allowing developers to choose their preferred one from a list. > Memory Feature; capacity to provide ChatGPT with memory.
        
        2023-04-13 About
      - Knowledge (Raphaël)
        
        raphaelsrty sur X : "My personal Knowledge Base made to the front page of HackerNews today..."
        
        Tags:
        
        Knowledge (Raphaël)
        Personal Knowledge Base
        Twitter thread
        
        2024-03-23 About
        
        raphaelsty.github.io/knowledge demo
        
        Tags:
        
        Demo
        Knowledge (Raphaël)
        
        2023-03-15 About
      - paper-QA
        
        whitead/paper-qa: LLM Chain for answering questions from documents with citations
        
        Tags:
        
        ChatGPT+Search
        LangChain
        LLM
        paper-QA
        Question Answering
        Research papers
        
        2023-03-29 About
      - perplexity.ai
        
        Perplexity AI
        
        Tags:
        
        perplexity.ai
        
        2023-01-20 About
  - Google ranking
  - Intent detection
    - Intent classification and slot filling
  - Knowledge graph search engine
  - Personal search engine
    - Knowledge (Raphaël)
      Building Monocle, a universal personal search engine for life | thesephist.com
      
      Tags:
      
      Personal-information management
      Personal Knowledge Management
      Personal search engine
      Semanlink related
      thesephist.com (Linus)
      
      [Incremental note-taking | thesephist.com](doc:2023/02/incremental_note_taking_%7C_these)
      
      2023-02-25 About
  - Private Search Engines
  - Ranked Entities in Search Results
  - Semantic Web search engine
    - Datao
    - sindice
    - Swoogle
  - SEO
    - Google: SEO
      - Google Rich Snippets
        
        Google Product Rich Snippets for Multiple Products on a Page | The Hepp Research Blog on Marketing with Data
        
        Tags:
        
        Google Rich Snippets
        Martin Hepp
        schema.org
        
        2015-07-09 About
        
        Google Structured Data Testing Tool
        
        Tags:
        
        Google Rich Snippets
        schema.org
        Validator
        Webmasters @ Google
        
        2013-06-13 About
        
        RDFa format data: what is its impact and when should you implement?
        
        Tags:
        
        Google Rich Snippets
        RDFa
        
        2013-01-30 About
        
        Why is Google not showing rich snippets for my pages? - GoodRelations Wiki
        
        Tags:
        
        GoodRelations
        Google Rich Snippets
        Wiki
        
        2012-10-15 About
        
        Rich snippets - Products
        
        Tags:
        
        Google Rich Snippets
        Product description
        
        2012-04-25 About
        
        Google Author Rich Snippets | Internet Alchemy
        
        Tags:
        
        Google Rich Snippets
        Ian Davis
        
        2011-08-22 About
        
        Marking up products for rich snippets
        
        Tags:
        
        Google Rich Snippets
        
        2011-02-09 About
        
        Rich snippets: Shopping and products
        
        Tags:
        
        Google Rich Snippets
        
        2011-02-09 About
        
        goodrelations-based rich snippet example in google query - Semantic Overflow
        
        Tags:
        
        GoodRelations
        Google Rich Snippets
        Semantic Overflow
        
        2011-02-09 About
        
        Rich snippets (microdata, microformats, RDFa)
        
        Tags:
        
        Google Rich Snippets
        
        2011-02-09 About
        
        Official Google Webmaster Central Blog: Introducing Rich Snippets - Welcome Google, really.
        
        Tags:
        
        fps' post
        Google Rich Snippets
        
        It looks like "Welcome Google, really" is (was) not in Google's index.
        
        2009-05-14 About
        
        Official Google Webmaster Central Blog: Introducing Rich Snippets
        
        Tags:
        
        Google Rich Snippets
        Guha
        
        2009-05-13 About
      - Google Structured Data Testing Tool
        
        category:structured-data jsonld (in Webmaster Central Help Forum) - Google Product Forums
        
        Tags:
        
        Google Structured Data Testing Tool
        JSON-LD
        
        2017-09-26 About
        
        Re: Google Structured Data Testing Tool fails on valid JSON-LD from Dan Brickley on 2016-04-07 (public-schemaorg@w3.org from April 2016)
        
        Tags:
        
        Dan Brickley
        Google Structured Data Testing Tool
        JSON-LD
        Martynas Jusevicius
        
        > The Google Structured Data Testing Tool in its current incarnation is largely oriented towards vocabulary that Google understands, even though it is perfectly harmless to include other information in your structured data
        
        2017-09-26 About
    - Google Rich Snippets
    - Google Structured Data Testing Tool
    - Semantic SEO
      - fps@EC-Web'14
        
        EC-WEB 2014 "Automotive ranges as e-commerce data"
        
        Tags:
        
        fps@EC-Web'14
        fpservant@slideshare
        
        2017-04-12 About
        
        Automotive range as e-commerce data (EC-WEB 2014)
        
        Tags:
        
        Automobile
        EC-Web
        fps@EC-Web'14
        
        2014-08-04 About
      - GoodRelations
        Car Options Ontology
        
        The Volkswagen Vehicles Ontology (VVO)
        
        Tags:
        
        Automotive and web technologies
        Car Options Ontology
        GoodRelations
        Volkswagen
        
        2011-09-23 About
        
        Volkswagen: Das Auto Company is Das Semantic Web Company! - semanticweb.com
        
        Tags:
        
        Automotive and web technologies
        Car Options Ontology
        Semantic Web
        Semantic web company
        Volkswagen
        
        2011-09-23 About
        
        The Car Options Ontology (COO)
        
        Tags:
        
        Car Options Ontology
        
        2011-03-24 About
        
        The Car Options Ontology (COO)
        
        Tags:
        
        Car Options Ontology
        
        2011-03-24 About
        
        Car Options Ontology at Schemapedia
        
        Tags:
        
        Car Options Ontology
        
        2011-03-24 About
        
        eClassOWL
        
        eClassOWL - The Web Ontology for Products and Services
        
        Tags:
        
        eClassOWL
        
        2013-03-02 About
        
        GoodRelations/Renault
        
        ANN: Renault adds GoodRelations to UK Shop
        
        Tags:
        
        GoodRelations/Renault
        Makolab
        Martin Hepp
        
        2011-10-12 About
        
        Product Types Ontology
        
        The Product Types Ontology: Use Wikipedia pages for describing products or services with GoodRelations
        
        Tags:
        
        GoodRelations
        Product Types Ontology
        
        2011-04-05 About
        
        VSO
        
        [goodrelations] vso:feature, dbPedia and productontology
        
        Tags:
        
        fps' post
        GoodRelations
        VSO
        
        2013-01-29 About
        
        The Vehicle Sales Ontology (VSO)
        
        Tags:
        
        VSO
        
        2011-03-24 About
        
        VSO / Goodrelations
        
        Tags:
        
        GoodRelations
        VSO
        
        Traditional Search Engine Optimization (SEO) tries to put you on top of all search results, but quite clearly, that can work only for one company. GoodRelations puts you on top of Web visibility for people who are looking for exactly your products or services.
        
        2011-03-24 About
        
        GoodRelations: technical report
        
        Tags:
        
        GoodRelations
        
        2014-04-21 About
        
        OPDM - Ontologies
        
        Tags:
        
        GoodRelations
        Martin Hepp
        Ontologies
        Product description
        
        2014-04-19 About
        
        schema blog: Good Relations and Schema.org
        
        Tags:
        
        GoodRelations
        Guha
        schema.org
        
        2013-02-15 About
        
        [goodrelations] vso:feature, dbPedia and productontology
        
        Tags:
        
        fps' post
        GoodRelations
        VSO
        
        2013-01-29 About
        
        [goodrelations] gr:ProductFeature
        
        Tags:
        
        fps' post
        GoodRelations
        
        2013-01-29 About
        
        The goodrelations November 2012 Archive by thread
        
        Tags:
        
        fps' post
        GoodRelations
        
        2012-11-15 About
        
        GoodRelations Fully Integrated with Schema.org - semanticweb.com
        
        Tags:
        
        GoodRelations
        schema.org
        
        2012-11-14 About
        
        Documentation/Product features - GoodRelations Wiki
        
        Tags:
        
        Configuration ontology
        GoodRelations
        Product description
        
        2012-11-12 About
        
        Why is Google not showing rich snippets for my pages? - GoodRelations Wiki
        
        Tags:
        
        GoodRelations
        Google Rich Snippets
        Wiki
        
        2012-10-15 About
        
        An Illustrated Guide To E-Commerce Markup Using GoodRelations
        
        Tags:
        
        GoodRelations
        Tutorial
        
        2012-10-13 About
        
        Google Becomes Answer Engine With Semantic Technology − Great News For Retailers
        
        Tags:
        
        GoodRelations
        Google
        
        2012-04-14 About
        
        The Volkswagen Vehicles Ontology (VVO)
        
        Tags:
        
        Automotive and web technologies
        Car Options Ontology
        GoodRelations
        Volkswagen
        
        2011-09-23 About
        
        Creating Domain-specific Products and Services Ontologies for GoodRelations
        
        Tags:
        
        GoodRelations
        Howto
        
        2011-06-28 About
        
        LOC - The Linked Open Commerce Dataspace
        
        Tags:
        
        GoodRelations
        Virtuoso
        
        2011-06-23 About
        
        GoodRelations for manufacturers
        
        Tags:
        
        GoodRelations
        
        If you are a manufacturer of commodities, you can use GoodRelations to help every small Web shop on the globe to sell your products.
        
        2011-06-10 About
        
        The Product Types Ontology: Use Wikipedia pages for describing products or services with GoodRelations
        
        Tags:
        
        GoodRelations
        Product Types Ontology
        
        2011-04-05 About
        
        Open eBusiness Ontology Usage: Investigating Community Implementation of GoodRelations
        
        Tags:
        
        GoodRelations
        LDOW2011
        Richard Cyganiak
        
        Experimental results highlight that early publishers of structured eCommerce data benefit more due to structured data being more readily search engine indexable, but the lack of available product ontologies and product master datasheets is impeding the creation of a semantically interlinked eCommerce Web.
        
        2011-04-05 About
        
        Investigating Community Implementation of the GoodRelations Ontology
        
        Tags:
        
        GoodRelations
        Richard Cyganiak
        
        2011-04-05 About
        
        Semantic SEO for Google with GoodRelations and RDFa | Hepp Research GmbH
        
        Tags:
        
        GoodRelations
        Google
        RDFa
        
        2011-03-24 About
        
        VSO / Goodrelations
        
        Tags:
        
        GoodRelations
        VSO
        
        Traditional Search Engine Optimization (SEO) tries to put you on top of all search results, but quite clearly, that can work only for one company. GoodRelations puts you on top of Web visibility for people who are looking for exactly your products or services.
        
        2011-03-24 About
        
        goodrelations-based rich snippet example in google query - Semantic Overflow
        
        Tags:
        
        GoodRelations
        Google Rich Snippets
        Semantic Overflow
        
        2011-02-09 About
        
        How Best Buy Is Using The Semantic Web - NYTimes.com
        
        Tags:
        
        GoodRelations
        RDFa
        Semantic Enterprise
        Semantic web : Use cases
        
        2010-07-30 About
        
        GoodRelations Annotator
        
        Tags:
        
        GoodRelations
        Linked Data
        Web of data
        
        GoodRelations Annotator and GoodRelations Snippet Generator: The easy way to advertise your business on the Web of Linked Data.
        
        2010-07-19 About
        
        GoodRelations vs. Google RDFa vs. Open Graph vs. hProduct/hListing: Using GoodRelations in 10 Triples
        
        Tags:
        
        GoodRelations
        
        2010-05-03 About
  - Vespa
  - Web search
    - DSSM (Deep Semantic Similarity Model)
      A Moderate Proposal for Radically Better AI-powered Web Search
      
      Tags:
      
      ColBERT
      Neural Search
      NLP@Stanford
      Web search
      
      2021-07-10 About
      Discovery Hub
      
      Tags:
      
      Linked Data Exploration
      To see
      Web search
      
      The exploratory search is a new way to search the web, not to find what you are searching, but to find what you are not searching, and might be intersting for you !
      
      2013-09-07 About
      Using SKOS vocabularies for improving Web Search
      
      Tags:
      
      Bernhard Haslhofer
      SKOS
      Web search
      WWW 2013
      
      2013-05-30 About
      Pipl - People Search
      
      Tags:
      
      Web search
      
      "The most comprehensive people search on the web"
      
      2010-04-26 About
  - Yahoo!
    - Peter Mika
    - SearchMonkey
    - Yahoo - My Web 2.0
  - Yandex
- Searching by description
- SIGIR
  - SIGIR 2022
  - SIGIR 2023
- Sparse retrieval
- Text Search
  - ElasticSearch
  - Lucene
  - Solr
    - Solr (not english only)
    - Solr + RDF
    - Solr and NLP
    - Solr - autocomplete
    - SolrCloud
    - Solr documentation
Mobile search
Multilingual search
Nearest neighbor search
NLP and Search
- LM + Search
  - ChatGPT+Search
  - DSPy (Demonstrate-Search-Predict)
- NLP based IR
- Ranked Entities in Search Results
  Semantic Scholar | AI-Powered Research Tool
  Tags:
  Semantic Scholar is a free, AI-powered research tool for scientific literature, based at the Allen Institute for AI.
  
  2020-11-24 About
  Building a Faster and Accurate Search Engine on Custom Dataset with Transformers 🤗 | by Shivanand Roy | Analytics Vidhya | Sep, 2020 | Medium
  Tags:
  2020-10-22 About
  Finding similar documents with transformers · Codegram
  Tags:
  2020-07-10 About
  Building a Search Engine with BERT and TensorFlow - Towards Data Science
  Tags:
  [somewhat related](/doc/2020/01/elasticsearch_meets_bert_build)
  
  2020-01-12 About
  Text similarity search in Elasticsearch using vector fields | Elastic Blog
  Tags:
  > How text embeddings and Elasticsearch’s dense_vector type **could be** used to support similarity search. > In practice, sentence embeddings often don’t generalize well to large sections of text. They are not commonly used to represent text longer than a short paragraph. Example of use: search similar question in a collection of Q/A Sample code is given to rank search results (TensorFlow + Google's universal sentence encoder + cosineSimilarity) Current limitation of vector similarity in Elasticsearch: vectors can be used for scoring documents, but not in the initial retrieval step. ([Ongoing work about approximate nearest neighbours search](https://github.com/elastic/elasticsearch/issues/42326). Will be a licensed feature of ES). > Conclusions: Using vectors for search is an important and **nuanced** area
  
  2020-01-10 About
  Elasticsearch meets BERT: Building Search Engine with Elasticsearch and BERT
  Tags:
  - Links to [this ES blog post](/doc/2020/01/text_similarity_search_in_elast) - [somewhat related](/doc/2020/01/building_a_search_engine_with_b)
  
  2020-01-10 About
Search Engines
Searching by description
Semantic Search
- Natural Language Semantic Search
- Searching by description
- SWSE
- Top2Vec
- txtai
  Ask like a human: Implementing semantic search on Stack Overflow - Stack Overflow Blog
  Tags:
  2023-08-11 About
  Jerry Liu sur Twitter : "Tuning top-k for semantic search is challenging... Introducing the LlamaIndex AutoRetriever for vector databases"
  Tags:
  > Tuning top-k for semantic search is challenging ; **the value can change depending on the context**. We now allow you to *infer* this value + other params during retrieval-time, using an LLM! Introducing the **LlamaIndex AutoRetriever for vector databases**.
  
  2023-05-13 About
  Alex sur Twitter : "Semantic search: how you can leverage both @CohereAI and @pinecone libraries to quickly build a POC..."
  Tags:
  2022-11-08 About
  Understanding Semantic Search and Question Answering | deepset
  Tags:
  - Haystack
  - Semantic Search
  2022-06-03 About
  How to Build a Semantic Search Engine With Transformers and Faiss | by Kostas Stathoulopoulos | Towards Data Science
  Tags:
  2022-01-29 About
  Semantic Search — Sentence-Transformers documentation
  Tags:
  **symmetric** semantic search vs **asymmetric** semantic search > - Suitable models for symmetric semantic search: Pre-Trained Sentence Embedding > - Suitable models for asymmetric semantic search: Pre-Trained MS MARCO Models
  
  2022-01-29 About
  Semantic search through a vectorized Wikipedia (SentenceBERT) with the Weaviate vector search engine
  Tags:
  2021-12-05 About
  raphaelsty/RetrieverReader: Fast API QA
  Tags:
  2021-10-04 About
  Semantic Search with S-BERT is all you need
  Tags:
  > SentenceTransformers is designed in such way that fine-tuning your own sentence / text embeddings models is easy.
  
  2021-06-05 About
  [2007.12603] IR-BERT: Leveraging BERT for Semantic Search in Background Linking for News Articles
  Tags:
  2021-04-12 About
  Weaviate Vector Search Engine | SeMI Technologies
  Tags:
  - Semantic Search
  > Weaviate is an API-based vector search engine with a graph data model that allows users to add data objects as graph nodes and (automatically or manually) add (machine learning) vectors to represent the nodes. Weaviate can be used for use cases ranging from similarity search to filtering out redundant information (i.e., deduplication) and from image search to enterprise NLP-based search.
  
  2021-03-30 About
  Towards Natural Language Semantic Code Search | GitHub Engineering
  Tags:
  - GitHub
  - Semantic Search
  2018-09-19 About
  [1706.00957] Semantic Vector Encoding and Similarity Search Using Fulltext Search Engines
  Tags:
  > The ubiquity of semantic vector space modeling raises the challenge of efficient searching in dense, high-dimensional vector spaces. We would naturally want to take advantage of the design and optimizations behind modern fulltext engines like Elasticsearch so as to meet the scalability and robustness demands of modern IR applications. This is the research challenge addressed in this paper. > The paper describes novel ways of encoding dense vectors into text documents, allowing the use of traditional inverted index engines. [blog post](https://rare-technologies.com/semantic-search-fulltext-engine-acl-2017/)
  
  2017-11-11 About
  Semantic search with NLP and elasticsearch - Stack Overflow
  Tags:
  2017-04-28 About
  Adding Semantic Search to Apache Solr
  Tags:
  - Semantic Search
  - Solr
  2014-03-30 About
  Studio Ousia Envisions A World Of Semantic Augmented Reality - Semanticweb.com
  Tags:
  - NLP: use cases
  - Semantic Search
  2014-03-26 About
  [ANN] Public release of Glimmer RDF search engine and demo from Peter Mika on 2013-06-20 (public-vocabs@w3.org from June 2013)
  Tags:
  The Semantic Search research group at Yahoo! Labs is pleased to announce the open-source code release and public demo of Glimmer, a search engine for RDF data
  
  2013-06-24 About
  The Enterprise Search Market – What should be on your radar? Keynote IKS 2012 Workshop | IKS Blog – The Semantic CMS Community
  Tags:
  2012-07-04 About
  Volkswagen adopts semantic search technology (Wired UK)
  Tags:
  - Semantic Search
  - Volkswagen
  2011-12-21 About
  Huoua Uses Semantic Search to ‘Spark’ Instant Circles - semanticweb.com
  Tags:
  2011-11-13 About
  PoolParty » SKOS Thesaurus Management – Semantic Search – Linked Data
  Tags:
  PoolParty is a thesaurus management system and a SKOS editor for the Semantic Web including text mining and linked data capabilities. The system helps to build and maintain multilingual thesauri providing an easy-to-use interface. PoolParty server provides semantic services to integrate semantic search or recommender systems into systems like CMS, DMS, CRM or Wikis </br> But, it is expensive.
  
  2010-08-31 About
  Corese
  Tags:
  - INRIA
  - Semantic Search
  COnceptual REsource Search Engine
  
  2010-05-14 About
  Europeana's semantic search engine.
  Tags:
  This is a research prototype of Europeana's semantic search engine.
  
  2009-05-05 About
  Semantic Search Arrives at the Web
  Tags:
  There are two approaches toward semantic search and both have received attention in the past months. The first approach builds on the automatic analysis of text using Natural Language Processing (NLP). The second approach uses semantic web technologies, which aims to make the web more easily searchable by allowing publishers to expose their (meta)data.
  
  2008-07-19 About
  SWSE - Semantic Search
  Tags:
  - Semantic Search
  - SWSE
  2007-07-20 About
Spatial search
Text Search
Visual search

1 Document (Long List)

[2106.04612] Neural Extractive Search

Tags:

how to extend a
search paradigm we call “**extractive search**” with
neural similarity techniques.

> some information needs require extracting
and aggregating sub-sentence information
(words, phrases, or entities) from multiple documents
(e.g. a list of all the risk factors for a specific
disease and their number of mentions, or a comprehensive
table of startups and CEOs).

> extractive search combines
document selection with information extraction. **The query is extended with capture slots**:
these are **search terms that act as variables, whose
values should be extracted**.
> The user
is then presented with the matched documents, each
annotated with the corresponding captured spans,
as well as aggregate information over the captured
spans

Conclusion :

> We presented a system for neural extractive search.
While we found our system to be useful for scientific
search, it also has clear limitations and areas
for improvement, both in terms of accuracy (only
72.2% of the returned results are relevant, both the
alignment and similarity models generalize well to
some relations but not to others), and in terms of
scale

[Video of demo](https://www.youtube.com/watch?v=TtqWi2GgB5A&t=1832s)

2021-06-23 About

Properties

sl:creationDate : 2021-05-19
sl:creationTime : 2021-05-19T01:14:37Z
rdf:type : sl:Tag
skos:prefLabel : Search