Search Results for author: Sarthak Dash

Found 9 papers, 1 papers with code

Permutation Invariant Strategy Using Transformer Encoders for Table Understanding

no code implementations Findings (NAACL) 2022 Sarthak Dash, Sugato Bagchi, Nandana Mihindukulasooriya, Alfio Gliozzo

Existing methods that leverage pretrained Transformer encoders range from a simple construction of pseudo-sentences by concatenating text across rows or columns to complex parameter-intensive models that encode table structure and require additional pretraining.

Column Type Annotation Entity Linking +4

Type Prediction Systems

no code implementations2 Apr 2021 Sarthak Dash, Nandana Mihindukulasooriya, Alfio Gliozzo, Mustafa Canim

Inferring semantic types for entity mentions within text documents is an important asset for many downstream NLP tasks, such as Semantic Role Labelling, Entity Disambiguation, Knowledge Base Question Answering, etc.

Entity Disambiguation Knowledge Base Question Answering +2

Open Knowledge Graphs Canonicalization using Variational Autoencoders

1 code implementation EMNLP 2021 Sarthak Dash, Gaetano Rossiello, Nandana Mihindukulasooriya, Sugato Bagchi, Alfio Gliozzo

In this work, we propose Canonicalizing Using Variational Autoencoders (CUVA), a joint model to learn both embeddings and cluster assignments in an end-to-end approach, which leads to a better vector representation for the noun and relation phrases.

Clustering Knowledge Graphs

Hypernym Detection Using Strict Partial Order Networks

no code implementations23 Sep 2019 Sarthak Dash, Md. Faisal Mahbub Chowdhury, Alfio Gliozzo, Nandana Mihindukulasooriya, Nicolas Rodolfo Fauceglia

This paper introduces Strict Partial Order Networks (SPON), a novel neural network architecture designed to enforce asymmetry and transitive properties as soft constraints.

Populating Web Scale Knowledge Graphs using Distantly Supervised Relation Extraction and Validation

no code implementations21 Aug 2019 Sarthak Dash, Michael R. Glass, Alfio Gliozzo, Mustafa Canim

In addition to that, the system uses a deep learning approach for knowledge base completion by utilizing the global structure information of the induced KG to further refine the confidence of the newly discovered relations.

Knowledge Base Completion Knowledge Graphs +1

Distributional Negative Sampling for Knowledge Base Completion

no code implementations16 Aug 2019 Sarthak Dash, Alfio Gliozzo

State-of-the-art approaches for Knowledge Base Completion (KBC) exploit deep neural networks trained with both false and true assertions: positive assertions are explicitly taken from the knowledge base, whereas negative ones are generated by random sampling of entities.

Knowledge Base Completion

Cannot find the paper you are looking for? You can Submit a new open access paper.