Photo taken at Lake Ruhondo, Rwanda. May 2023.
Thiviyan Thanapalasingam
I am currently doing a research internship at Sony AI in Barcelona.
PhD candidate at the University of Amsterdam, advised by Paul Groth and Peter Bloem.
I am interested in Neurosymbolic Learning and Reasoning.
INDElab at Science Park, Amsterdam
News
I have started a research internship at Sony AI in Barcelona 🇪🇸.
A new preprint on GRAPES: Learning to Sample Graphs for Scalable Graph Neural Networks is now available on arXiv. GRAPES is a GFlowNet-based sampling approach that selects crucial nodes for a given downstream task, enabling GNNs to scale to large graphs.
A new preprint on IntelliGraphs: Datasets for Benchmarking Knowledge Graph Generation is now available on arXiv.
Last week, I attended the NeSy 2023 workshop which was hosted at the beautiful Certosa di Pontignano in Siena, Italy 🇮🇹. It was a perfect blend of networking and diving deeper into the fascinating world of NeSy research.
We hosted a hybrid workshop about Neurosymbolic Generative Models (NeSy-GeMs) at the ICLR conference in Kigali, Rwanda 🇷🇼. It was great to see so many interesting presentations. Check out the accepted papers on the NeSy-GeMs website.
A new preprint on A-NeSI: A Scalable Approximate Method for Probabilistic Neurosymbolic Inference is now available on arXiv. A-NESI efficiently combines neural networks with probabilistic reasoning using approximate inference, allowing it to perform reasoning tasks at scale. (Accepted at NeurIPS 2023 🎉)
Preprints
IntelliGraphs: Datasets for Benchmarking Knowledge Graph Generation [ PDF ]new
Thiviyan Thanapalasingam, Emile van Krieken, Peter Bloem, Paul Groth.
arXiv preprint. 2023.
GRAPES: Learning to Sample Graphs for Scalable Graph Neural Networks [ PDF ]new
Taraneh Younesian, Thiviyan Thanapalasingam, Emile van Krieken, Daniel Daza, Peter Bloem.
arXiv preprint. 2023.
Also presented at the Graph Learning Frontiers workshop at NeurIPS 2023
Conference papers
A-NeSI: A Scalable Approximate Method for Probabilistic Neurosymbolic Inference [ PDF ]new
Emile van Krieken, Thiviyan Thanapalasingam, Jakub M. Tomczak, Frank van Harmelen, Annette ten Teije.
Neural Information Processing Systems (NeurIPS) 2023.
Also presented at the Neurosymbolic Generative Models workshop at ICLR 2023
The CSO Classifier: Ontology-Driven Detection of Research Topics in Scholarly Articles [ PDF ] [ PyPI Python Package ]
Angelo Salatino, Francesco Osborne, Thiviyan Thanapalasingam and Enrico Motta.
In International Conference on Theory and Practice of Digital Libraries (TPDL) 2019.
Ontology-Based Recommendation of Editorial Products [ PDF ]
Thiviyan Thanapalasingam, Francesco Osborne, Aliaksandr Birukou and Enrico Motta.
Industry Track at International Semantic Web Conference (ISWC) 2018.
The Computer Science Ontology: A Large-Scale Taxonomy of Research Areas [ PDF ]
Angelo Salatino, Thiviyan Thanapalasingam, Andrea Mannocci, Francesco Osborne and Enrico Motta.
Resource Track at International Semantic Web Conference (ISWC) 2018.
Nominated for the Best Resource Paper Award 🎉
Supporting Springer Nature Editors by means of Semantic Technologies [ PDF ]
Francesco Osborne, Angelo Salatino, Aliaksandr Birukou, Thiviyan Thanapalasingam and Enrico Motta.
Industry Track at International Semantic Web Conference (ISWC) 2017.
Technical reports
Prompting as Probing: Using Language Models for Knowledge Base Construction [ PDF ] [ Full report ]
Joint work with Dimitrios Alivanistos, Selene Báez Santamaría, Michael Cochez, Jan-Christoph Kalo, Emile van Krieken.
Knowledge Base Construction from Pre-trained Language Models (LM-KBC) challenge at International Semantic Web Conference (ISWC) in Bertinoro, Italy. 2022.
Won 1st prize for the best-performing model 🏆
A Decentralized Approach to Validating Personal Data Using a Combination of Blockchains and Linked Data [ PDF ] [ Full report ]
Mentored by John Domingue. Joint work with Cristina-Iulia Bucur, Fiorela Ciroku, Tatiana Makhalova, Ettore Rizza, Dalia Varanka, Michael Wolowyk.
At International Semantic Web Summer School (ISWS) in Bertinoro, Italy. 2018.
Won 1st prize for best research report 🏆 and 1st prize best presentation 🏆
A Theoretical and Experimental Study of Pure and Mixed Water Clusters using Infrared Spectroscopy [ PDF ]
Advised by Professor Andrew Ellis at the University of Leicester, UK.
Masters thesis in Physical and Theoretical Chemistry. 2016.
Awarded a First Class Honours Degree 🎓
Workshop Organisation
💎 Neurosymbolic Generative Models (NeSy-GeMS) Hybrid Workshop
Workshop Website
Organising Committee:
- Thiviyan Thanapalasingam (University of Amsterdam, NL)
- Emile van Krieken (Vrije Universiteit Amsterdam, NL)
- Halley Young (University of Pennsylvania, USA)
- Disha Shrivastava (MILA, Canada)
- Jakub M. Tomczak (Eindhoven University of Technology, NL)
- Kevin Ellis (Cornell University, USA)
Journal articles
Relational Graph Convolutional Networks: A Closer Look [ PDF ]
Thiviyan Thanapalasingam, Lucas van Berkel, Peter Bloem and Paul Groth.
PeerJ Computer Science (2022) 8:e1073.
The Computer Science Ontology: A Comprehensive Automatically-Generated Taxonomy of Research Areas [ PDF ]
Angelo Salatino, Thiviyan Thanapalasingam, Andrea Mannocci, Aliaksandr Birukou, Francesco Osborne and Enrico Motta.
Data Intelligence (2019) 2 (3) : p379–416.
Posters & Demos
The Smart Book Recommender: An Ontology-Driven Application for Recommending Editorial Products [ PDF ]
Thiviyan Thanapalasingam, Francesco Osborne, Aliaksandr Birukou and Enrico Motta.
Demo Track at International Semantic Web Conference (ISWC) 2018.
Classifying Research Papers with the Computer Science Ontology [ PDF ] [ Code ]
Angelo Salatino, Thiviyan Thanapalasingam, Andrea Mannocci, Francesco Osborne and Enrico Motta.
Demo Track at International Semantic Web Conference (ISWC) 2018.
Smart Book Recommender: A Semantic Recommendation Engine for Editorial Products [ PDF ]
Francesco Osborne, Thiviyan Thanapalasingam, Angelo Salatino, Aliaksandr Birukou and Enrico Motta.
Poster Track at International Semantic Web Conference (ISWC) 2017.