Semanlink - Neural networks AND Microsoft

> the paper LoRA: Low-Rank Adaptation of Large Language Models proposes to decompose the weight changes, ΔW, into a lower-rank representation. (To be technically correct, LoRA does not decompose the matrices directly, but it learns the decomposed matrices via backpropagation).
>
> suppose ΔW is the weight update for an A × B weight matrix. Then, we can decompose the weight update matrix into two smaller matrices: ΔW = WA WB, where WA is an an A × r-dimensional matrix, and WB is an an r × B-dimensional matrix.

LORA and LLama :

> Lit-LLaMA repository a simple, readable reimplementation of Meta’s popular LLaMA model. Besides code for training and running LLaMA itself (with the original Meta LLaMA weights), it also contains code for finetuning LLaMA using LLaMA-Adapter and LoRA.

2023-07-27 A propos

What is low-rank adaptation (LoRA)? - TechTalks

Tags:

2023-07-27 A propos

LoRA for semantic similarity tasks

Tags:

2023-07-21 A propos

[2307.08621] Retentive Network: A Successor to Transformer for Large Language Models

Tags:

2023-07-20 A propos

Introducing Azure OpenAI Service On Your Data in Public Preview - Microsoft Community Hub

Tags:

2023-06-21 A propos

How to use Alpaca-LoRA to fine-tune a model like ChatGPT

Tags:

2023-05-26 A propos

State of GPT (Andrej Karpathy at Microsoft Build)

Tags:

2023-05-24 A propos

Bringing the power of AI to Windows 11

Tags:

Windows Copilot

2023-05-24 A propos

Text Translation cognitive skill - Azure Cognitive Search | Microsoft Learn

Tags:

2023-05-09 A propos

Google "We Have No Moat, And Neither Does OpenAI"

Tags:

> low-cost public involvement was enabled
by a vastly cheaper mechanism for fine tuning called low
rank adaptation ()[LoRA](tag:lora)

> **Part of what makes LoRA so effective is that ... it’s stackable.**
>
> By contrast, training giant models from scratch not only
throws away the pretraining, but also any iterative
improvements that have been made on top.

> LoRA updates are very cheap to produce (~$100) for the
most popular model sizes.

> Many of these projects are saving time by training on
small, highly curated datasets...
> These
datasets are built using synthetic methods (e.g. filtering
the best responses from an existing model) and
scavenging from other projects

> Directly Competing With Open Source
Is a Losing Proposition

> Paradoxically, the one clear winner in all of this is Meta.
Because the leaked model was theirs ([LLaMA](tag:llama)), they have
effectively garnered an entire planet's worth of free labor.
Since most open source innovation is happening on top of
their architecture, there is nothing stopping them from
directly incorporating it into their products.

2023-05-04 A propos

Azure-Samples/azure-search-openai-demo: A sample app for the Retrieval-Augmented Generation pattern running in Azure

Tags:

2023-04-14 A propos

Physics of AI - YouTube

Tags:

- Intelligence has emerged: why? how?
- Let's study this with *controlled experiments* and *toy models*
- Clean and clear insights that peer slightly behind the magic curtain

emergence; rise of scale. Phase transition: fct(compute time for training: nb of params, dataset size)

(12:25) Transformers: the jump when compared to traditional NN: instead on oparating on single input X, operates on a set of inputs. (cf. training to say whether an image contains 2 elements of the same kind). **The aptitude to compare elements in the input sequence enables analogies** - the essence of reasoning.(Classical NN treats one high dimensional vector, tranformers hanlde sets of vectors. Attention layer replaces learned filters W of classical NN by the other elements in the sequence. Attention module compares part of input against other part of input. "Relative machines rather than absolute". Sets are a very powerful level of abstraction)

emergence
> The truth is that nobody has a clue what is going on"
>
> "**Something unknown is doing we don't know what**" (A. Eddington)

being inspired by physics' methodology: controled experiments and toy mathematical models

2023-03-28 A propos

[2106.09685] LoRA: Low-Rank Adaptation of Large Language Models

Tags:

2023-03-21 A propos

Microsoft launches Teams Premium with features powered by OpenAI - The Verge

Tags:

2023-02-07 A propos

BioGPT

Tags:

2023-02-02 A propos

[2002.01808] K-Adapter: Infusing Knowledge into Pre-Trained Models with Adapters

Tags:

2023-01-12 A propos

Talking to Models: Stanford U & Microsoft Method Enables Developers to Correct Model Bugs via Natural Language Patches | Synced

Tags:

2022-11-27 A propos

[2008.12813] HittER: Hierarchical Transformers for Knowledge Graph Embeddings

Tags:

2022-06-30 A propos

Unveiling Transformers with LEGO - YouTube

Tags:

2022-06-30 A propos

Lenka Zdeborova sur Twitter : "Amazing lecture of @SebastienBubeck giving intuition on transformers and how to approach them theoretically..."

Tags:

2022-06-30 A propos

[2004.05119] Beyond Fine-tuning: Few-Sample Sentence Embedding Transfer

Tags:

2022-03-31 A propos

[2106.13474] Adapt-and-Distill: Developing Small, Fast and Effective Pretrained Language Models for Domains

Tags:

2021-10-21 A propos

[2007.15779] Domain-Specific Language Model Pretraining for Biomedical Natural Language Processing

Tags:

2021-04-11 A propos

[2001.09522] TaxoExpan: Self-supervised Taxonomy Expansion with Position-Enhanced Graph Neural Network

Tags:

how to add a set of new concepts to an existing taxonomy.

[Tweet](https://twitter.com/mickeyjs6/status/1253772146142216194?s=20) [GitHub](https://github.com/mickeystroller/TaxoExpan)

> we study the taxonomy expansion task: given an
existing taxonomy and a set of new emerging concepts, we aim
to automatically expand the taxonomy to incorporate these new
concepts (without changing the existing relations in the given taxonomy).

> To the best of our knowledge, this is the first study on **how to
expand an existing directed acyclic graph (as we model a taxonomy
as a DAG) using self-supervised learning**.

Self-supervised framework, the existing taxonomy being used as training data: it learns a model to predict whether a query concept is the direct hyponym of an anchor concept.

> 2 techniques:
>
> 1. a **position-enhanced graph neural network that encodes the local structure of an anchor concept** in the existing taxonomy,
> 2. a noise-robust training objective that enables the learned model to be insensitive to the label noise in the self-supervision data.

Regarding 1: uses [GNN](/tag/graph_neural_networks.html) to model the "ego network" of concepts (potential “siblings”
and “grand parents” of the query concept).

> Regular
GNNs fail to distinguish nodes with different relative positions to
the query (i.e., some nodes are grand parents of the query while
the others are siblings of the query). To address this limitation, we
present a simple but effective enhancement to inject such position
information into GNNs using position embedding. We show that
such embedding can be easily integrated with existing GNN architectures
(e.g., [GCN](/tag/graph_convolutional_networks) and GAT) and significantly boosts the
prediction performance

Regarding point 2: uses InfoNCE loss, cf. [Contrastive Predictive Coding](/doc/?uri=https%3A%2F%2Farxiv.org%2Fabs%2F1807.03748)

> Instead of predicting
whether each individual ⟨query concept, anchor concept⟩ pair
is positive or not, we first group all pairs sharing the same query
concept into a single training instance and learn a model to select
the positive pair among other negative ones from the group.

(Hum, ça me rappelle quelque chose)

> assume each concept (in existing taxonomy + set of new concepts) has an initial embedding
vector learned from some text associated with this concept.

To keep things tractable, only attempts to find a single parent node of each new concept.

2020-04-25 A propos

[2002.02925] BERT-of-Theseus: Compressing BERT by Progressive Module Replacing

Tags:

2020-02-10 A propos

Canwen Xu sur Twitter : "WTF? We brutally dismember BERT and replace all his organs?"

Tags:

2020-02-10 A propos

Neural Models for Information Retrieval (2017)

Tags:

2019-08-18 A propos

[1602.01137] A Dual Embedding Space Model for Document Ranking

Tags:

2019-07-17 A propos

Robust Language Representation Learning via Multi-task Knowledge Distillation - Microsoft Research

Tags:

2019-05-19 A propos

[1901.11504] Multi-Task Deep Neural Networks for Natural Language Understanding

Tags:

2019-02-17 A propos

Knowledge Graph and Text Jointly Embedding (2014)

Tags:

2018-01-05 A propos

Learning Deep Structured Semantic Models for Web Search using Clickthrough Data - Microsoft Research (2013)

Tags:

2017-12-30 A propos

DSSM ("Deep Semantic Similarity Model") - Microsoft Research

Tags:

DSSM (Deep Semantic Similarity Model)

2017-12-30 A propos

Enriching Word Embeddings Using Knowledge Graph for Semantic Tagging in Conversational Dialog Systems - Microsoft Research (2015)

Tags:

> new simple, yet effective approaches to
learn domain specific word embeddings.

## Intro

> Adapting word embeddings, such as jointly capturing
syntactic and semantic information, can further enrich semantic
word representations for several tasks, e.g., sentiment
analysis (Tang et al. 2014), named entity recognition
(Lebret, Legrand, and Collobert 2013), entity-relation extraction
(Weston et al. 2013), etc. (Yu and Dredze 2014)
has introduced a lightly supervised word embedding learning
extending word2vec. They incorporate prior information to the objective
function as a regularization term considering synonymy relations
between words from Wordnet (Fellbaum 1999).

> In this work, we go one step further and investigate if
enriching the word2vec word embeddings trained on unstructured/
unlabeled text with domain specific semantic relations
obtained from knowledge sources (e.g., knowledge
graphs, search query logs, etc.) can help to discover relation
aware word embeddings. Unlike earlier work, **we encode the
information about the relations between phrases, thereby,
entities and relation mentions are all embedded into a low dimensional
vector space**.

## Related work (Learning Word Embeddings with Priors)

- word2vec
- Relational Constrained Model (RTM) (Yu and Dredze 2014)
While CBOW learns lexical word embeddings from provided text, the RTM learns embeddings of words based on their similarity to other words provided by a knowledge resource (eg. wordnet)
- Joint model  (Yu and Dredze 2014)
combines CBOW and RTM through linear combination

2017-10-02 A propos

Vectorland: Brief Notes from Using Text Embeddings for Search

Tags:

2017-09-18 A propos

Using Text Embeddings for Information Retrieval

Tags:

2017-09-18 A propos

Scientists See Advances in Deep Learning, a Part of Artificial Intelligence - NYTimes.com

Tags:

2012-11-30 A propos