A5087419620- Topic Modeling
- Natural Language Processing Techniques
- Semantic Web and Ontologies
- Data Quality and Management
- Advanced Text Analysis Techniques
- Text and Document Classification Technologies
- Advanced Graph Neural Networks
- Biomedical Text Mining and Ontologies
- Speech and dialogue systems
- Multimodal Machine Learning Applications
- Text Readability and Simplification
- Web Data Mining and Analysis
- Expert finding and Q&A systems
- Domain Adaptation and Few-Shot Learning
- Time Series Analysis and Forecasting
- Geographic Information Systems Studies
- Rough Sets and Fuzzy Logic
- Multi-Agent Systems and Negotiation
- Speech Recognition and Synthesis
- Information Retrieval and Search Behavior
- Artificial Intelligence in Healthcare and Education
- Wikis in Education and Collaboration
- Personal Information Management and User Behavior
- Intelligent Tutoring Systems and Adaptive Learning
- Data Visualization and Analytics
IBM (United States)
2012-2023
IBM Research - Thomas J. Watson Research Center
2020-2022
Columbia University
2016
Skyline College
2012
Istituto di Scienza e Tecnologie dell'Informazione "Alessandro Faedo"
2007-2010
Ferioli & Gianotti (Italy)
2009
Istituto Centrale per la Ricerca Scientifica e Tecnologica Applicata al Mare
2001-2009
National Academies of Sciences, Engineering, and Medicine
2008
Fondazione Bruno Kessler
2007
Universidad Autónoma de Madrid
2005
Pavan Kapanipathi, Ibrahim Abdelaziz, Srinivas Ravishankar, Salim Roukos, Alexander Gray, Ramón Fernandez Astudillo, Maria Chang, Cristina Cornelio, Saswati Dana, Achille Fokoue, Dinesh Garg, Alfio Gliozzo, Sairam Gurajada, Hima Karanam, Naweed Khan, Khandelwal, Young-Suk Lee, Yunyao Li, Francois Luus, Ndivhuwo Makondo, Nandana Mihindukulasooriya, Tahira Naseem, Sumit Neelam, Lucian Popa, Revanth Gangi Reddy, Ryan Riegel, Gaetano Rossiello, Udit Sharma, G P Shrivatsa Bhargav, Mo Yu. Findings...
Michael Glass, Gaetano Rossiello, Md Faisal Mahbub Chowdhury, Ankita Naik, Pengshan Cai, Alfio Gliozzo. Proceedings of the 2022 Conference North American Chapter Association for Computational Linguistics: Human Language Technologies. 2022.
This paper explores the role of domain information in word sense disambiguation. The underlying hypothesis is that labels, such as M EDICINE , A RCHITECTURE and S PORT provide a useful way to establish semantic relations among senses, which can be profitably used during disambiguation process. Results obtained at ENSEVAL -2 initiative confirm for significant subset words disambiguate with very high level precision.
Michael Glass, Alfio Gliozzo, Rishav Chakravarti, Anthony Ferritto, Lin Pan, G P Shrivatsa Bhargav, Dinesh Garg, Avi Sil. Proceedings of the 58th Annual Meeting Association for Computational Linguistics. 2020.
Michael Glass, Mustafa Canim, Alfio Gliozzo, Saneem Chemmengath, Vishwajeet Kumar, Rishav Chakravarti, Avi Sil, Feifei Pan, Samarth Bharadwaj, Nicolas Rodolfo Fauceglia. Proceedings of the 2021 Conference North American Chapter Association for Computational Linguistics: Human Language Technologies. 2021.
Text-to-SQL is emerging as a practical interface for real world databases. The dominant paradigm cross-database or schema-independent, supporting application schemas unseen during training. schema of database defines the tables, columns, column types and foreign key connections between tables. Real can be large, containing hundreds but any particular query only small fraction will relevant. Placing entire in prompt an LLM impossible models with smaller token windows expensive even when...
In this paper we present a supervised Word Sense Disambiguation methodology, that exploits kernel methods to model sense distinctions. particular combination of functions is adopted estimate independently both syntagmatic and domain similarity. We defined function, namely the Domain Kernel, allowed us plug "external knowledge" into learning process. External knowledge acquired from unlabeled data in totally unsupervised way, it represented by means Models. evaluated our methodology on...
Yannis Katsis, Saneem Chemmengath, Vishwajeet Kumar, Samarth Bharadwaj, Mustafa Canim, Michael Glass, Alfio Gliozzo, Feifei Pan, Jaydeep Sen, Karthik Sankaranarayanan, Soumen Chakrabarti. Proceedings of the 2022 Conference North American Chapter Association for Computational Linguistics: Human Language Technologies: Industry Track. 2022.
We propose KnowGL, a tool that allows converting text into structured relational data represented as set of ABox assertions compliant with the TBox given Knowledge Graph (KG), such Wikidata. address this problem sequence generation task by leveraging pre-trained sequence-to-sequence language models, e.g. BART. Given sentence, we fine-tune models to detect pairs entity mentions and jointly generate facts consisting full semantic annotations for KG, labels, types, their relationships. To...
Many forms of word relatedness have been developed, providing different perspectives on similarity.We introduce a Bayesian probabilistic tensor factorization model for synthesizing single vector representation and per-perspective linear transformations from any number similarity matrices.The resulting vectors, when combined with the transformation, approximately recreate while also regularizing generalizing, each perspective.Our method can combine manually created semantic resources neural...
Cross-language Text Categorization is the task of assigning semantic classes to documents written in a target language (e.g. English) while system trained using labeled source Italian).In this work we present many solutions according availability bilingual resources, and show that it possible deal with problem even when no such resources are accessible. The core technique relies on automatic acquisition Multilingual Domain Models from comparable corpora.Experiments effectiveness our...
This paper introduces Strict Partial Order Networks (SPON), a novel neural network architecture designed to enforce asymmetry and transitive properties as soft constraints. We apply it induce hypernymy relations by training with is-a pairs. also present an augmented variant of SPON that can generalize type information learned for in-vocabulary terms previously unseen ones. An extensive evaluation over eleven benchmarks across different tasks shows consistently either outperforms or attains...
Extracting lexico-semantic relations as graph-structured taxonomies, also known taxonomy construction, has been beneficial in a variety of NLP applications. Recently Graph Neural Network (GNN) shown to be powerful successfully tackling many tasks. However, there no attempt exploit GNN create taxonomies. In this paper, we propose Graph2Taxo, GNN-based cross-domain transfer framework for the construction task. Our main contribution is learn latent features from existing domains guide structure...
Tahira Naseem, Srinivas Ravishankar, Nandana Mihindukulasooriya, Ibrahim Abdelaziz, Young-Suk Lee, Pavan Kapanipathi, Salim Roukos, Alfio Gliozzo, Alexander Gray. Proceedings of the 59th Annual Meeting Association for Computational Linguistics and 11th International Joint Conference on Natural Language Processing (Volume 2: Short Papers). 2021.
In a multilingual scenario, the classical monolingual text categorization problem can be reformulated as cross language TC task, in which we have to cope with two or more languages (e.g. English and Italian). this setting, system is trained using labeled examples source English), it classifies documents different target
In this paper we propose and evaluate a technique to perform semi-supervised learning for Text Categorization.In particular defined kernel function, namely the Domain Kernel, that allowed us plug "external knowledge" into supervised process.External knowledge is acquired from unlabeled data in totally unsupervised way, it represented by means of Models.We evaluated Kernel two standard benchmarks Categorization with good results, compared its performance function exploits bag-of-words feature...
This paper investigates conceptually and empirically the novel sense matching task, which requires to recognize whether senses of two synonymous words match in context. We suggest direct approaches problem, avoid intermediate step explicit word disambiguation, demonstrate their appealing advantages stimulating potential for future research.
We present an approach to ontology population based on a lexical substitution technique. It consists in estimating the plausibility of sentences where named entity be classified is substituted with ones contained training data, our case, partially populated ontology. Plausibility estimated by using Web while classification algorithm instance-based. evaluated method two different tasks. Experiments show that solution effective, outperforming existing methods, and it can applied practical problems.