IBM
http://www.semanlink.net/tag/ibm
Documents tagged with IBMTopics | IBM Research
http://www.semanlink.net/doc/2024/03/topics_%7C_ibm_research
2024-03-07T16:10:39ZOn the Surprising Behavior of Distance Metrics in High Dimensional Space (Aggarwal 2001)
http://www.semanlink.net/doc/2024/03/on_the_surprising_behavior_of_d
> in high dimensional space, the concept of proximity, distance
or nearest neighbor may not even be qualitatively meaningful.
2024-03-03T21:33:43ZBen Hoover sur X : "Energy Transformer (ET): A novel architecture combining 3 prominent ideas in AI..."
http://www.semanlink.net/doc/2023/12/ben_hoover_sur_x_%E2%9A%A1%EF%B8%8Fenergy_tr
- Transformers: mix tokens with attention
- Energy-based Models: inference descends a tractable energy function
- Associative Memory: inference performs error correction
2023-12-02T17:13:49Z[2306.04640] ModuleFormer: Modularity Emerges from Mixture-of-Experts
http://www.semanlink.net/doc/2023/09/2306_04640_moduleformer_modu
> a new neural network architecture, ModuleFormer, that leverages modularity to improve the efficiency and flexibility of large language models.
[GitHub](https://github.com/IBM/ModuleFormer)
2023-09-16T00:15:56ZIBM/zshot: Zero and Few shot named entity & relationships recognition
http://www.semanlink.net/doc/2022/12/ibm_zshot_zero_and_few_shot_na
2022-12-23T01:00:31Z[2212.01340] Moving Beyond Downstream Task Accuracy for Information Retrieval Benchmarking
http://www.semanlink.net/doc/2022/12/2212_01340_moving_beyond_down
2022-12-06T19:27:25Z[2210.13952] KnowGL: Knowledge Generation and Linking from Text
http://www.semanlink.net/doc/2022/11/2210_13952_knowgl_knowledge_
How to fine-tune PLMs to read a sentence and
generate the corresponding full set of semantic annotations
that are compliant with the terminology of a KG?
> we propose a framework able to convert text into
a set of Wikidata statements
2022-11-13T10:48:17ZText classification by labeling words | Proceedings of the 19th national conference on Artifical intelligence (2004)
http://www.semanlink.net/doc/2022/11/text_classification_by_labeling
2022-11-08T18:37:01ZZshot: Zero and Few shot named entity & relationships recognition
http://www.semanlink.net/doc/2022/10/zshot_zero_and_few_shot_named_
2022-10-01T20:13:51Z[2010.00711] A Survey of the State of Explainable AI for Natural Language Processing
http://www.semanlink.net/doc/2022/09/2010_00711_a_survey_of_the_st
2022-09-08T09:30:14ZActive Learning: A Survey (C. Aggarwal 2014)
http://www.semanlink.net/doc/2022/09/active_learning_a_survey_c_a
2022-09-06T18:33:24ZActive Learning for BERT: An Empirical Study - ACL Anthology
http://www.semanlink.net/doc/2022/09/active_learning_for_bert_an_em
> The use of Actice Learning (AL)
with deep pre-trained models has so far received
little consideration.
>
> We study the
potential of (i) various AL strategies; (ii) in conjunction
with BERT, (iii) within a highly challenging
– yet common – real-world scenario of
class imbalance and scarce labeled data.
focused on binary classification
> AL can boost BERT performance, especially in the most realistic scenario in which the initial set of labeled examples is created using keyword-based queries, resulting in a biased sample of the minority class.
[Github](https://github.com/IBM/low-resource-text-classification-framework)
2022-09-02T16:08:49Z[2203.10581] Cluster & Tune: Boost Cold Start Performance in Text Classification
http://www.semanlink.net/doc/2022/04/2203_10581_cluster_tune_bo
[Leshem Choshen sur Twitter : "Labelled data is scarce, what can we do?..."](doc:2022/04/leshem_choshen_sur_twitter_l)
> **One-sentence Summary**: we suggest adding an unsupervised intermediate classification step, before finetunning and after pretraining BERT, and show it improves performance for data-constrained cases.
> for text classification cold start (when labeled
data is scarce), **add an intermediate unsupervised
classification task**, between the pretraining
and fine-tuning phases:
> perform clustering and
train the pre-trained model on predicting the
cluster labels.
> this additional
classification phase can significantly improve
performance, mainly for **topical classification**
tasks
> we use an efficient clustering technique,
that relies on simple Bag Of Words (BOW)
representations, to partition the unlabeled training
data into relatively homogeneous clusters of text
instances.
>
> Next, we treat these clusters as labeled
data for an intermediate text classification task, and
train the pre-trained model – with or without additional
MLM pretraining – with respect to this
multi-class problem, prior to the final fine-tuning
over the actual target-task labels
> The underlying
intuition is that inter-training the model
over a related text classification task would be more
beneficial compared to MLM inter-training, which
focuses on different textual entities, namely predicting
the identity of a single token.
2022-04-06T01:22:32ZLeshem Choshen sur Twitter : "Labelled data is scarce, what can we do?..."
http://www.semanlink.net/doc/2022/04/leshem_choshen_sur_twitter_l
> We can MLM on the unlabeled data, but You can do better: Cluster & Tune - **finetune on clusters as labels**
[github](https://github.com/IBM/intermediate-training-using-clustering) ; Paper: [[2203.10581] Cluster & Tune: Boost Cold Start Performance in Text Classification](doc:2022/04/2203_10581_cluster_tune_bo)
2022-04-06T01:18:22Z[2108.13934] Robust Retrieval Augmented Generation for Zero-shot Slot Filling
http://www.semanlink.net/doc/2022/01/2108_13934_robust_retrieval_a
> "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)
2022-01-19T17:14:49ZA Survey of Text Clustering Algorithms - C. C. Aggarwal (2012)
http://www.semanlink.net/doc/2021/04/a_survey_of_text_clustering_alg
2021-04-20T01:08:01Z[2009.07938] Type-augmented Relation Prediction in Knowledge Graphs
http://www.semanlink.net/doc/2020/09/2009_07938_type_augmented_rel
2020-09-19T10:00:31ZIBM Research addressing Enterprise NLP challenges in 2020
http://www.semanlink.net/doc/2020/06/ibm_research_addressing_enterpr
2020-06-12T09:41:21ZMatching Resumes to Jobs via Deep Siamese Network | Companion Proceedings of the The Web Conference 2018
http://www.semanlink.net/doc/2020/02/matching_resumes_to_jobs_via_de
Siamese adaptation of CNN, using contrastive loss. The document embedding of resumes and job descriptions
(dim 200) are generated using [#Doc2Vec](/tag/doc2vec.html) and are given as
inputs to the network.
2020-02-10T13:43:44ZAdvancing Natural Language Processing (NLP) for Enterprise Domains
http://www.semanlink.net/doc/2020/01/advancing_natural_language_proc
Reviews 4 papers by IBM research.
Introductive remark: the specificities of search in enterprises when compared to the web:
content stored in silos with much less repetition of key information,
intricate questions expecting detailed answers,
reluctance to blackbox.
Regarding NLP: silos, incomplete data, small data, changing environment.
-> 3 themes of research at IBM Research to improve NLP for enterprises:
- systems that can work with small data, external knowledge and use neurosymbolic approaches to language
- explainability on how a system reached a conclusion
- scaling to allow continuous adaptation
2020-01-07T12:05:46ZProject Debater - IBM Research AI
http://www.semanlink.net/doc/2019/11/project_debater_ibm_research_
2019-11-06T01:12:43Zwatson-developer-cloud/speech-javascript-sdk: IBM Watson Speech Services for Web Browsers
http://www.semanlink.net/doc/2019/10/watson_developer_cloud_speech_j
2019-10-30T00:47:08ZGetting robots to listen: Using Watson's Speech to Text service - Watson
http://www.semanlink.net/doc/2019/10/getting_robots_to_listen_using
Include python sample code using WebSockets: streaming audio to the Watson Speech to Text service while also getting responses back at the same time.
2019-10-30T00:10:58ZIBM Cloud Speech to Text : Références de recherche
http://www.semanlink.net/doc/2019/10/ibm_cloud_speech_to_text_refe
2019-10-29T17:57:39ZSpeech to Text - IBM Cloud API Docs
http://www.semanlink.net/doc/2019/10/speech_to_text_ibm_cloud_api_
2019-10-28T10:57:07ZA Postman Collection for Training IBM Watson Speech to Text
http://www.semanlink.net/doc/2019/10/a_postman_collection_for_traini
2019-10-24T23:37:39Z[1909.04120] Span Selection Pre-training for Question Answering
http://www.semanlink.net/doc/2019/09/_1909_04120_span_selection_pre
> a **new pre-training task inspired by reading
comprehension** and an **effort to avoid encoding general knowledge in the transformer network itself**
Current transformer architectures store general knowledge -> large models, long pre-training time. Better to offload the requirement of general knowledge to a sparsely activated network.
"Span selection" as an additional auxiliary task: the query is a sentence drawn from a corpus
with a term replaced with a special token: [BLANK]. The term replaced by the blank is the answer term. The passage is
relevant as determined by a BM25 search, and answer-bearing (containing the answer
term). Unlike BERT’s cloze task, where the answer must be drawn from the model itself, the answer is found in a passage
using language understanding.
> **We hope to progress to a model of general purpose language modeling that uses an indexed long
term memory to retrieve world knowledge, rather than holding it in the densely activated transformer encoder layers.**
2019-09-18T17:26:33ZSpeech to Text Demo - Watson
http://www.semanlink.net/doc/2019/06/speech_to_text_demo_watson
2019-06-11T11:04:08ZWord Mover's Embedding: From Word2Vec to Document Embedding (2018)
https://aclanthology.coli.uni-saarland.de/papers/D18-1482/d18-1482
unsupervised embeddings of sentences of variable length from pre-trained word embeddings (better on short length text).
(Builds on the word mover's distance, but using ideas borrowed from kernel methods approximation, gets a representation of sentences, instead of just a distance between them)
2018-11-10T15:38:38ZQuantum Computing - IBM Q
https://www.research.ibm.com/ibm-q/
IBM Q is an industry-first initiative to build commercially available universal quantum computers for business and science.
2018-08-21T09:01:17ZWatson : l’Intelligence artificielle en ses limites | InternetActu
http://internetactu.blog.lemonde.fr/2017/10/07/watson-lintelligence-artificielle-en-ses-limites/
2017-10-07T21:50:55ZWatson: Alchemy Language v1 API Explorer
https://watson-api-explorer.mybluemix.net/apis/alchemy-language-v1
The AlchemyLanguage API uses natural language processing technology and machine learning algorithms to extract semantic meta-data from content, such as information on people, places, companies, topics, facts, relationships, authors, and languages.
2017-07-18T18:04:05ZIBM SPSS Text Analytics for Surveys
https://www.ibm.com/us-en/marketplace/spss-text-analytics-for-surveys
2017-07-13T10:38:21ZSurvey results analysis - Analytics Exchange
https://console.ng.bluemix.net/data/exchange/public/entry/view/ac418581e657fc785fe9573c1013c3a6
Use this storybook to analyze results of surveys from online tools such as SurveyMonkey
2017-06-08T14:06:31ZAnalyzing survey text: a brief overview
http://www.besmart.company/wp-content/uploads/2014/11/briefoverview01.pdf
Learn how IBM SPSS Text Analytics for Surveys gives you greater insight
2017-06-08T00:46:32ZWatson API Explorer
https://watson-api-explorer.mybluemix.net/
2017-06-06T11:57:28ZIBM Watson Developer Cloud
https://www.ibm.com/watson/developercloud/
2017-06-06T11:53:54ZWhat IBM, the Semantic Web Company, and Siemens are doing with semantic technologies | ZDNet
http://www.zdnet.com/article/a-little-semantics-goes-a-long-way/
2016-09-17T15:06:38ZIBM and Microsoft Will Let You Roll Your Own Blockchain | WIRED
http://www.wired.com/2016/02/ibm-and-microsoft-will-let-you-roll-your-own-blockchain/
2016-02-17T23:50:01ZIBM Watson APIs hold key to broader cognitive computing use
http://searchdatamanagement.techtarget.com/news/4500269406/IBM-Watson-APIs-hold-key-to-broader-cognitive-computing-use
2015-12-30T20:19:54ZIBM's 'Rodent Brain' Chip Could Make Our Phones Hyper-Smart | WIRED
http://www.wired.com/2015/08/ibms-rodent-brain-chip-make-phones-hyper-smart/
2015-08-18T14:07:58ZBBC News - IBM's Watson in Africa to help solve problems
http://www.bbc.co.uk/news/technology-26065991
2014-02-17T23:56:52ZReflections on a Year spent developing with RDF and JSON (Software Development on the Cloud Exploration)
https://www.ibm.com/developerworks/community/blogs/c06ef551-0127-483d-a104-cdd02b1cee31/entry/february_3_2014_1_47_pm?lang=en
2014-02-05T14:04:09ZHow IBM's Watson Will Change The Way We Work - Forbes
http://www.forbes.com/sites/gregsatell/2013/10/27/how-ibms-watson-will-change-the-way-we-work/
2013-12-18T15:58:54ZIBM to offer Watson supercomputer as cloud development platform | ITworld
http://www.itworld.com/software/382730/ibm-offer-watson-supercomputer-cloud-development-platform
2013-11-18T09:57:23ZGoogle in Jeopardy: What If IBM's Watson Dethroned the King of Search? | Wired Opinion | Wired.com
http://www.wired.com/opinion/2013/10/google-in-jeopardy-what-if-watson-beat-the-search-giant/
2013-10-05T23:56:29ZElementary, My Dear Watson - Will IBM’s quiz show champion outgrow humankind?
http://www.geekexchange.com/elementary-my-dear-watson-will-ibms-quiz-show-champion-outgrow-humankind-73517.html
2013-07-31T09:53:53ZWatson Goes Back to School - And what it tells us about the evolving role of semantic technology
http://semtechbiznyc2012.semanticweb.com/sessionPop.cfm?confid=68&proposalid=5022
In the traditional vision of AI, understanding flowed from perception through language to knowledge. It had always been envisioned that this understanding would be in some precise and unambiguous knowledge representation, and that all meaning processing would happen in this representation. This is the root of all semantic technology today. However, over time, the failure of the AI community to achieve this end-to-end vision made many, especially those in NLP, question the endpoint. In other words, to doubt the value of semantic technology. In this talk, we show that it was the vision, not the technology, that deserved to be doubted. Semantic technology has significant value in accomplishing tasks that require understanding, but it is not the endpoint.
2012-07-30T23:59:28ZSearch smarter with Apache Solr, Part 2: Solr for the enterprise
http://www.ibm.com/developerworks/java/library/j-solr2/
2012-05-15T17:00:01ZSearch smarter with Apache Solr, Part 1: Essential features and the Solr schema
http://www.ibm.com/developerworks/java/library/j-solr1/
2012-05-15T16:55:20ZIBM’s Watson Computer Gets a Wall Street Job - Bloomberg
http://www.bloomberg.com/news/2012-03-05/ibm-s-watson-computer-gets-wall-street-job-one-year-after-jeopardy-win.html
2012-03-06T17:40:01ZToward a Basic Profile for Linked Data
http://www.ibm.com/developerworks/rational/library/basic-profile-linked-data/index.html
A collection of best practices and a simple approach for a Linked Data architecture
2011-12-27T19:17:06ZIBM's Five Predictions for the Next Five Years - BusinessWeek
http://www.businessweek.com/technology/ibms-five-predictions-for-the-next-five-years-12192011.html
2011-12-21T23:11:52ZThe Semantic Web, Linked Data and Drupal, Part 2: Combine linked datasets with Drupal 7 and SPARQL Views
http://www.ibm.com/developerworks/library/wa-datasets/
2011-09-16T16:04:55ZThe Semantic Web, Linked Data and Drupal, Part 1: Expose your data using RDF
http://www.ibm.com/developerworks/web/library/wa-rdf/
2011-09-15T13:55:44ZImprove your taxonomy management using the W3C SKOS standard
http://www.ibm.com/developerworks/xml/library/x-skostaxonomy/index.html
2011-05-12T21:53:51ZSubject classification with DITA and SKOS
http://www.ibm.com/developerworks/xml/library/x-dita10/
2011-02-15T11:55:39ZIntegrate disparate data sources with Semantic Web technology
http://www.ibm.com/developerworks/library/x-disprdf/index.html
Different sets of RDF data are much easier to combine than different sets of data in other common formats. You can easily convert disparate non-RDF data sets to RDF and then combine them to create new content. In this article, learn how to integrate spreadsheet data, CSV data from a web service, and fielded data from a website into a single report.
2010-09-30T23:06:24ZSemantic Enterprise: What Are The Gorillas Doing? (Oracle, IBM, HP, Cisco, Microsoft and SAP) - Semantic Web
http://www.semanticweb.com/on/semantic_enterprise_what_are_the_gorillas_doing_oracle_ibm_hp_cisco_microsoft_and_sap_168973.asp
2010-09-17T18:28:49ZSmarter Than You Think - I.B.M.'s Supercomputer to Challenge 'Jeopardy!' Champions - NYTimes.com
http://www.nytimes.com/2010/06/20/magazine/20Computer-t.html
2010-06-23T00:28:34ZHow IBM Plans to Win Jeopardy!
http://www.technologyreview.com/computing/22702/?a=f
IBM's Watson will showcase the latest tricks in natural-language processing.
2010-06-23T00:26:50ZExtracting Enterprise Vocabularies Using Linked Open Data
http://data.semanticweb.org/pdfs/iswc/2009/in-use/paper143.pdf
A common vocabulary is vital to smooth business operation, yet codifying and maintaining an enterprise vocabulary is an arduous, manual task. We describe a process to automatically extract a domain specific vocabulary (terms and types) from unstructured data in the en- terprise guided by term definitions in Linked Open Data (LOD). We validate our techniques by applying them to the IT (Information Tech- nology) domain, taking 58 Gartner analyst reports and using two specific LOD sources – DBpedia and Freebase. We show initial findings that ad- dress the generalizability of these techniques for vocabulary extraction in new domains, such as the energy industry.
<br/>IBM Watson Research Center
2010-05-31T12:06:48ZIBM plans 'brain-like' computers
http://newsvote.bbc.co.uk/mpapps/pagetools/print/news.bbc.co.uk/2/hi/science/nature/7740484.stm?ad=1
IBM has announced it will lead a US government-funded collaboration to make electronic circuits that mimic brains.
2008-11-24T13:35:09ZCreate test cases for Web applications
http://www.ibm.com/developerworks/java/library/j-jwebunit/
2008-11-18T11:34:32ZalphaWorks : Scalable Highly Expressive Reasoner : Overview
http://www.alphaworks.ibm.com/tech/sher
Scalable Highly Expressive Reasoner (SHER) is a breakthrough technology that provides ontology analytics over highly expressive ontologies (OWL-DL without nominals). SHER does not do any inferencing on load; hence it deals better with quickly changing data (the downside is, of course, that reasoning is performed at query time). The tool can reason on approximately seven million triples in seconds, and it scales to data sets with 60 million triples, responding to queries in minutes. It has been used to semantically index 300 million triples from medical literature. SHER tolerates logical inconsistencies in the data, and it can quickly point you to these inconsistencies in the data and help you clean up inconsistencies before issuing semantic queries. The tool explains (or justifies) why a particular result set is an answer to the query; this explanation is useful for validation by domain experts.
2008-07-19T18:14:06ZXML processing in Ajax, Part 2: Two Ajax and XSLT approaches
http://www.ibm.com/developerworks/xml/library/x-xmlajaxpt2/
2008-03-18T16:21:15Z"Mastering Ajax" - developerWorks : Web development : Technical library view
http://www.ibm.com/developerworks/views/web/libraryview.jsp?search_by=Mastering+Ajax
2007-12-06T00:50:14ZAjax and XML: Ajax for media
http://www.ibm.com/developerworks/library/x-ajaxxml7/?ca=dgr-lnxw01AjaxMedia
Use Ajax techniques to show movies and slide shows
2007-11-08T16:45:05ZSearch RDF data with SPARQL (and Jena)
http://www-128.ibm.com/developerworks/xml/library/j-sparql/
2007-04-20T20:58:42ZThe ultimate mashup -- Web services and the semantic Web
http://www-128.ibm.com/developerworks/edu/x-dw-x-ultimashup1.html
In addition to single-service applications, developers are creating mashups, applications that combine data from multiple services to create something new. This series chronicles the creation of the ultimate mashup, an application that not only stores data from different mashups but uses semantic technology to enable users to create their own mashups by swapping services, or even by picking and choosing data. It uses Java™ programming and a combination of servlets, JSP, software from the open source Jena project, and DB2's new native XML capabilities.
2006-08-29Mashups: The new breed of Web app
http://www-128.ibm.com/developerworks/library/x-mashups.html?ca=dgr-lnxw16MashupChallenges
An introduction to mashups
2006-08-19Display XML with Cascading Stylesheets, Part 1: Use Cascading Stylesheets to display XML
https://www6.software.ibm.com/developerworks/education/x-xmlcss/
2006-06-25BBC NEWS - Salvage prospect for 'junk' DNA
http://news.bbc.co.uk/1/hi/sci/tech/4940654.stm
A mathematical analysis of the human genome suggests that so-called "junk DNA" might not be so useless after all.
2006-05-05alphaWorks : IBM Web Ontology Manager
http://www.alphaworks.ibm.com/tech/wom?open&S_TACT=105AGX59&S_CMP=GR
IBM Web Ontology Manager is a lightweight, Web-based tool for managing ontologies expressed in Web Ontology Language (OWL). With this technology, users can browse, search, and submit ontologies to an ontology repository. This technology includes a Web interface for easy uploading of ontologies in an .owl format by any user of the system. It also includes an interface for generating (using Jastor) Java™ APIs from uploaded ontology files.
<br/>
IBM Web Ontology Manager differs from IBM Ontology Management System (a former alphaWorks technology, now merged into the IBM Integrated Ontology Development Toolkit) in that it does not include a statement repository. Instead, based on the ontologies visible to the system, it can generate Java classes for accessing any Jena-compatible RDF statement repository.
2006-04-29Recommended PHP reading list
http://www-128.ibm.com/developerworks/opensource/library/os-php-read/
2006-03-28Mastering Ajax, Part 3: Advanced requests and responses in Ajax
http://www-128.ibm.com/developerworks/web/library/wa-ajaxintro3/?ca=dgr-lnxw01MasterAJAX3
2006-02-17Call SOAP Web services with Ajax
http://www-128.ibm.com/developerworks/webservices/library/ws-wsajax/?ca=dgr-lnxw03SOAP-AJAX
Implement a Web browser-based SOAP Web services client using the Asynchronous JavaScript and XML (Ajax) design pattern.
2006-02-05The future of the Web is Semantic
http://www-128.ibm.com/developerworks/web/library/wa-semweb/
2006-01-30Joho the Blog: IBM shows del.icio.us for the enterprise, and more
http://www.hyperorg.com/blogger/mtarchive/ibm_shows_delicious_for_the_en.html
2005-11-10Writing multithreaded Java applications
http://www-128.ibm.com/developerworks/java/library/j-thread.html
2005-10-29Mission to build a simulated brain begins
http://www.newscientist.com/article.ns?id=dn7470&print=true
2005-06-06