Nasrin Mostafazadeh

View original post here.


I am a senior AI research scientist at Elemental Cognition where I work on the next generation of AI systems that not only comprehend stories, but also explain their reasoning and answer 'why' (we are hiring!). I am active in the AI startup community and I am a big believer in the power of sustainable startups disrupting different sectors through AI, something that is harder to achieve in big tech companies. I got my PhD at the Computer Science department of University of Rochester, with my thesis titled "From Event to Story Understanding". At Rochester, I worked under James F. Allen in Conversational Interaction and Spoken Dialogue Research Group. After graduation, I joined BenevolentAI labs in NYC as a senior research scientist to use NLP and AI for accelerating scientific discovery for major diseases. Before that, I spent about a year at Microsoft Research working on various Vision & Language problems and one summer at Google NYC working on Sentiment Analysis. I've got my bachelor's degree from Sharif University of Technology in 2012 where I was a member of NLP lab. I graduated from Farzanegan high school where I started my work in NLP and AI. For more detailes, please refer to my CV.


I'm interested in building intelligent systems that can comprehend language (in a deep manner, where they can "explain" themselves!). I mainly model language in terms of 'events' and how they interact with each other in terms of temporal and causal relations, with applications ranging from storytelling to vision & language. Deep language understanding, which enables inference and commonsense reasoning, requires systems that have large amounts of knowledge enabling them to connect natural language to the concepts of the world. A part of my work concerns developing approaches for learning knowledge bases on semantically-rich narrative structures.


New York, NY

Email: nasrinm [at] cs [dot] rochester [dot] edu


Recent Invited Talks & Interviews


Interview with This Week in Machine Learning (TwiML), "Contextual Modeling for Language and Vision", August 2018 [Interview]


Speaker at ACM NYC, "Language Understanding and Language Generation in Eventful Contexts", July 2018


Keynote Speaker at the Storytelling Workshop at NAACL 2018, "Event-centric Context Modeling: The Case of Story Comprehension and Story Generation", June 2018 [Slides]


Interview with the Computer Vision & Pattern Recognition (CVPR) Magazine, featured in the Best of CVPR, June 2018, [Interview]


Speaker and Panelist at the VQA and the Visual Dialogue Workshop at CVPR 2018, "Event-centric Multimodal Context Modeling", June 2018, [Slides]


Speaker at Cognitive AI NYC, "Language Understanding and Language Generation in Eventful Contexts", April 2018


Speaker at Columbia Lean-in Conference, "Perseverance & Personal Power", March 2018


Speaker at the NLP Seminar at the Stanford University, "Language Comprehension and Language Generation in Eventful Contexts", November 2018, [Slides]


Speaker at CS Seminar Series at the Columbia University, "Language Comprehension and Language Generation in Eventful Contexts", October 2018


Publications


Tackling The Story Ending Biases in The Story Cloze TestRishi Sharma, James Allen, Omid Bakhshandeh, Nasrin MostafazadehIn Proceedings of the 2018 Conference of the Association for Computational Linguistics (ACL), 2018[Data -coming-] [Bibtex]


From Event to Story UnderstandingNasrin MostafazadehPhD Thesis at the University of Rochester, Rochester, NY, 2017                                              [Bibtex]


Image-Grounded Conversations: Multimodal Context for Natural Question and Response GenerationNasrin Mostafazadeh, Chris Brockett, Bill Dolan, Michel Galley, Jianfeng Gao, Georgios P. Spithourakis, Lucy VanderwendeIn Proceedings of the International Joint Conference on Natural Language Processing (IJCNLP), 2017[Data] [Bibtex]


LSDSem 2017 Shared Task: The Story Cloze TestNasrin Mostafazadeh, Michael Roth, Annie Louis, Nathanael Chambers, James AllenIn Proceedings of the EACL Workshop on Linking Models of Lexical, Sentential and Discourse-level Semantics (LSDSem), 2017[Bibtex]


Generating Natural Questions About an ImageNasrin Mostafazadeh, Ishan Misra, Jacob Devlin, Margaret Mitchell, Xiaodong He, Lucy VanderwendeIn Proceedings of the 2016 Conference of the Association for Computational Linguistics (ACL), 2016[Supplementary] [Data] [Bibtex]


Story Cloze Evaluator: Vector Space Representation Evaluation by Predicting What Happens NextNasrin Mostafazadeh, Lucy Vanderwende, Wen-Tau (Scott) Yih, Pushmeet Kohli, James AllenIn Proceedings of ACL Workshop on Evaluating Vector Space Representations for NLP (RepEval), 2016[Bibtex]


A Corpus and Cloze Evaluation for Deeper Understanding of Commonsense StoriesNasrin Mostafazadeh, Nathanael Chambers, Xiaodong He, Devi Parikh, Dhruv Batra, Lucy Vanderwende, Pushmeet Kohli and James AllenIn Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL HLT), 2016[Web Page] [Presentation] [Bibtex] [Story Cloze Test is on CodaLab!]


CaTeRS: Causal and Temporal Relation Scheme for Semantic Annotation of Event StructuresNasrin Mostafazadeh, Alyson Grealish, Nathanel Chambers, James Allen and Lucy VanderwendeIn Proceedings of NAACL Event Workshop, 2016[Annotated ROCStories Dataset] [Presentation] [Bibtex]


Visual StorytellingTing-Hao Huang, Francis Ferraro, Nasrin Mostafazadeh, Ishan Misra, Jacob Devlin, Aishwarya Agrawal, Ross Girshick, Xiaodong He, Pushmeet Kohli, Dhruv Batra, Larry Zitnick, Devi Parikh, Lucy Vanderwende, Michel Galley and Margaret MitchellIn Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL HLT), 2016[Bibtex] [Data]


A Survey of Current Datasets for Vision and Language ResearchFrancis Ferraro*, Nasrin Mostafazadeh*, Ting-Hao Huang, Michel Galley, Lucy Vanderwende, Margaret Mitchell and Jacob DevlinIn Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing (EMNLP), 2015* = equal contribution [Poster] [Bibtex] [vision&language.net] [interview]


SemEval-2015 Task 5: QA TempEval - Evaluating Temporal Information Understanding with Question AnsweringHector Llorens, Nathanael Chambers, Naushad UzZaman, Nasrin Mostafazadeh, James Allen and James PustejovskyIn Proceedings of the Workshop on Semantic Evaluations (SemEval), 2015.[Slides] [Bibtex] [Data]


Learning Semantically Rich Event Inference Rules Using Definition of VerbsNasrin Mostafazadeh and James F. AllenIn Proceedings of Computational Linguistics and Intelligent Text Processing (CICLing), 2015.[Slides] [Bibtex]


Formal Verification of Temporal Questions In the Context of Query-Answering Text SummarizationNasrin Mostafazadeh, Omid Bakhshandeh, Gholamreza Ghassem-SaniIn Proceedings of Canadian Conference on Artificial Intelligence (AI'2012), 2012.[Bibtex]


Pazesh: A Graph-Based Approach to Increase Readability of Automatic Text SummariesNasrin Mostafazadeh, Seyed Abolghasem Mirroshandel, Gholamreza Ghassem-Sani, Omid Bakhshandeh}In Proceedings of Canadian Conference on Artificial Intelligence (AI'2011), 2011.[Bibtex]


Organization


NAACL 2019

I will be co-chairing the system demonstration track at NAACL 2019. The Call for Papers to be announced soon.

LSDSem 2017

I co-organized Linking Models of Lexical, Sentential and Discourse-level Semantics (LSDSem) workshop, with the ***Story Cloze Test*** as the shared task. Visit the Shared Task pagefor more details.

QA TempEval 2015

I co-organized Question Answering Temporal Evaluation (QA TempEval) as the task 5 at SemEval-2015, continuing the TempEval series. Visit the page for more details.

25 views0 comments