Papers
See AI2's Award Winning Papers
Learn more about AI2's Lasting Impact Award
Viewing 411-420 of 991 papers
What's in your Head? Emergent Behaviour in Multi-Task Transformer Models
Mor Geva, Uri Katz, Aviv Ben-Arie, Jonathan BerantEMNLP • 2021 The primary paradigm for multi-task training in natural language processing is to represent the input with a shared pre-trained language model, and add a small, thin network (head) per task. Given an input, a target head is the head that is selected for…DIALKI: Knowledge Identification in Conversational Systems through Dialogue-Document Contextualization
Zeqiu Wu, Bo-Ru Lu, Hannaneh Hajishirzi, Mari OstendorfEMNLP • 2021 Identifying relevant knowledge to be used in conversational systems that are grounded in long documents is critical to effective response generation. We introduce a knowledge identification model that leverages the document structure to provide dialogue…Finding needles in a haystack: Sampling Structurally-diverse Training Sets from Synthetic Data for Compositional Generalization
Inbar Oren, Jonathan Herzig, Jonathan BerantEMNLP • 2021 Modern semantic parsers suffer from two principal limitations. First, training requires expensive collection of utterance-program pairs. Second, semantic parsers fail to generalize at test time to new compositions/structures that have not been observed during…Container: Context Aggregation Network
Peng Gao, Jiasen Lu, Hongsheng Li, R. Mottaghi, Aniruddha KembhaviarXiv • 2021 Convolutional neural networks (CNNs) are ubiquitous in computer vision, with a myriad of effective and efficient variations. Recently, Transformers – originally introduced in natural language processing – have been increasingly adopted in computer vision…SciA11y: Converting Scientific Papers to Accessible HTML
Lucy Lu Wang, Isabel Cachola, Jonathan Bragg, Evie (Yu-Yen) Cheng, Chelsea Hess Haupt, Matt Latzke, Bailey Kuehl, Madeleine van Zuylen, Linda M. Wagner, Daniel S. WeldASSETS • 2021We present SciA11y, a system that renders inaccessible scientific paper PDFs into HTML. SciA11y uses machine learning models to extract and understand the content of scientific PDFs, and reorganizes the resulting paper components into a form that better…Best Artifact AwardCan Machines Learn Morality? The Delphi Experiment
Liwei Jiang, Chandra Bhagavatula, Jenny Liang, Jesse Dodge, Keisuke Sakaguchi, Maxwell Forbes, Jon Borchardt, Saadia Gabriel, Yulia Tsvetkov, Regina A. Rini, Yejin ChoiarXiv • 2021 As AI systems become increasingly powerful and pervasive, there are growing concerns about machines’ morality or a lack thereof. Yet, teaching morality to machines is a formidable task, as morality remains among the most intensely debated questions in…Delphi: Towards Machine Ethics and Norms
Liwei Jiang, Jena D. Hwang, Chandrasekhar Bhagavatula, Ronan Le Bras, Maxwell Forbes, Jon Borchardt, Jenny Liang, Oren Etzioni, Maarten Sap, Yejin ChoiarXiv • 2021 Failing to account for moral norms could notably hinder AI systems’ ability to interact with people. AI systems empirically require social, cultural, and ethical norms to make moral judgments. However, open-world situations with different groundings may shift…Reflective Decoding: Beyond Unidirectional Generation with Off-the-Shelf Language Models
Peter West, Ximing Lu, Ari Holtzman, Chandra Bhagavatula, Jena D. Hwang, Yejin ChoiACL • 2021 Publicly available, large pretrained Language Models (LMs) generate text with remarkable quality, but only sequentially from left to right. As a result, they are not immediately applicable to generation tasks that break the unidirectional assumption, such as…SciCo: Hierarchical Cross-Document Coreference for Scientific Concepts
Arie Cattan, Sophie Johnson, Daniel S. Weld, Ido Dagan, Iz Beltagy, Doug Downey, Tom HopeAKBC • 2021 Determining coreference of concept mentions across multiple documents is fundamental for natural language understanding. Work on cross-document coreference resolution (CDCR) typically considers mentions of events in the news, which do not often involve…Scientific Language Models for Biomedical Knowledge Base Completion: An Empirical Study
Rahul Nadkarni, David Wadden, Iz Beltagy, Noah A. Smith, Hannaneh Hajishirzi, Tom HopeAKBC • 2021 Biomedical knowledge graphs (KGs) hold rich information on entities such as diseases, drugs, and genes. Predicting missing links in these graphs can boost many important applications, such as drug design and repurposing. Recent work has shown that general…