Siddharth Vashishtha

PhD Student

Pronouns: he/him

Hi! I am Siddharth (Sid). I am a fourth-year Computer Science Ph.D. student at the University of Rochester advised by Aaron White.

My research focus is in Natural Language Understanding and Information Extraction. I work on building computational models of semantics in natural language text. In my previous work, I have worked on building models for extracting semantic information related to events – event triggers, event chronology, event duration etc. I also collaborate with Ben Van Durme and his group at Johns Hopkins University on projects related to Information Extraction.

My papers are listed on the publications page. Full list of publications can be found at Semantic Scholar/Google Scholar

PhD Internships


Amazon Alexa AI	Verneek.ai	Google
Applied Scientist Intern	NLP Research Intern	Research Intern
Summer 2020	Summer 2021	Summer 2022

selected publications

ACL

Fine-Grained Temporal Relation Extraction

Siddharth Vashishtha, Benjamin Van Durme, and Aaron Steven White

In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, Jul 2019

Abs PDF Code Website

We present a novel semantic framework for modeling temporal relations and event durations that maps pairs of events to real-valued scales. We use this framework to construct the largest temporal relations dataset to date, covering the entirety of the Universal Dependencies English Web Treebank. We use this dataset to train models for jointly predicting fine-grained temporal relations and event durations. We report strong results on our data and show the efficacy of a transfer-learning approach for predicting categorical relations.
EMNLP Findings

Temporal Reasoning in Natural Language Inference

Siddharth Vashishtha, Adam Poliak, Yash Kumar Lal, and 2 more authors

In Findings of the Association for Computational Linguistics: EMNLP, Nov 2020

Abs PDF Code

We introduce five new natural language inference (NLI) datasets focused on temporal reasoning. We recast four existing datasets annotated for event duration—how long an event lasts—and event ordering—how events are temporally arranged—into more than one million NLI examples. We use these datasets to investigate how well neural models trained on a popular NLI corpus capture these forms of temporal reasoning.