Videos
See AI2's full collection of videos on our YouTube channel.Viewing 41-50 of 252 videos
Environment-oriented modeling for grounded NLP and collective intelligence systems
February 21, 2023 | Ronen TamariAbstract: “Ask not what's inside your head, but what's your head inside of” — W.M. Mace, 1977 In contrast to classical cognitive science which studied brains in isolation, ecological cognitive science accounts focus on the role of the body and external environment in shaping cognition. In this talk I will…
Understanding and Improving Compositional Generalization
February 10, 2023 | Ben BoginPre-trained language models perform well on a large variety of question-answering tasks, but they still often fail in the compositional generalization setup, where models are tested on unseen compositions of reasoning skills. In this talk, I will go over three research directions that address this challenge. In…
No Language Left Behind Unlocking Text Data for Under Resourced
February 9, 2023 | Shruti RijhwaniNLP systems are limited by the availability of text data, and because machine-readable text exists only in a few hundred languages, most of the world’s languages are under-represented in modern language technologies. Text data exists in many more languages! However, it is locked away in printed books and…
Artificial Intelligence and Ethics: Towards a Robust Normative Framework
February 8, 2023 | S. Matthew LiaoAbstract: As AI technologies continue to advance, questions about the ethics of AI become more pressing than ever. In this talk, I aim to outline some of the key issues in the study of AI ethics, identify some of the core claims that have been made, and propose an ethical framework that can help us guide these…
What Do NLP Researchers Believe? Results of the NLP Community Metasurvey
January 12, 2023 | Julian Michael/Research Scientist, New York UniversityAbstract: I will present the results of the NLP Community Metasurvey (nlpsurvey.net). This was a questionnaire that we ran from May to June 2022 which elicited the opinions of NLP researchers on controversial issues, including industry influence in the field, concerns about AGI, and ethics. Our results put…
Global storm-resolving climate simulation in Python
December 14, 2022 | Oliver FuhrerThis video shows an animation of a climate simulation done with a Python implementation of NOAA's climate model FV3GFS on the Piz Daint supercomputer. Even thought the code is implemented in Python, it runs 3.9x faster than the original implementation in Fortran. This is due to the help of a domain-specific…
Just-DREAM-about-it: Figurative Language Understanding with DREAM-FLUTE
November 18, 2022 | Yuling GuFigurative language (e.g., "he flew like the wind") is challenging to understand, as it is hard to tell what implicit information is being conveyed from the surface form alone. We hypothesize that to perform this task well, the reader needs to mentally elaborate the scene being described to identify a sensible…
Towards Teachable Reasoning Systems: Using a Dynamic Memory of User Feedback for Continual System Improvement
November 16, 2022 | Bhavana Dalvi MishraEMNLP'22 talk for paper: https://www.semanticscholar.org/paper/Towards-Teachable-Reasoning-Systems%3A-Using-a-Memory-Dalvi-Tafjord/e7d75b80e0fa3ae190ff91676dbf18a006d3a311 Our goal is a teachable reasoning system for question-answering (QA), where a user can interact with faithful answer explanations, and…
Entailer: Answering Questions with Faithful and Truthful Chains of Reasoning
November 16, 2022 | Bhavana DalviEMNLP '22 Talk for paper: https://www.semanticscholar.org/paper/Explaining-Answers-with-Entailment-Trees-Dalvi-Jansen/4a56f72b9c529810ba4ecfe9eac522d87f6db81d Our goal is a question-answering (QA) system that can show how its answers are implied by its own internal beliefs via a systematic chain of reasoning…
Biomedical AI for Precision Health
August 16, 2022 | Hoifung PoonAbstract: The advent of big data promises to revolutionize medicine by making it more personalized and effective, but big data also presents a grand challenge of information overload. For example, tumor sequencing has become routine in cancer treatment, yet interpreting the genomic data requires painstakingly…