We present SpanBERT, a pre-training method that is designed to better represent and predict spans of text. Our approach extends BERT by (1) masking contiguous random spans, rather than random tokens, and (2) training the span boundary representations to predict the entire content of the masked span, without relying on the individual token representations within it. SpanBERT consistently outperforms BERT and our better-tuned baselines, with substantial gains on span selection tasks such as question answering and coreference resolution. In particular, with the same training data and model size as BERTlarge, our single model obtains 94.6% and 88.7% F1 on SQuAD 1.1 and 2.0 respectively. We also achieve a new state of the art on the OntoNotes coreference resolution task (79.6% F1), strong performance on the TACRED relation extraction benchmark, and even gains on GLUE. 1

We present TriviaQA, a challenging reading comprehension dataset containing over 650K question-answer-evidence triples. TriviaQA includes 95K questionanswer pairs authored by trivia enthusiasts and independently gathered evidence documents, six per question on average, that provide high quality distant supervision for answering the questions. We show that, in comparison to other recently introduced large-scale datasets, TriviaQA (1) has relatively complex, compositional questions, (2) has considerable syntactic and lexical variability between questions and corresponding answer-evidence sentences, and (3) requires more cross sentence reasoning to find answers. We also present two baseline algorithms: a featurebased classifier and a state-of-the-art neural network, that performs well on SQuAD reading comprehension. Neither approach comes close to human performance (23% and 40% vs. 80%), suggesting that TriviaQA is a challenging testbed that is worth significant future study. 1

The KNOWITALL system aims to automate the tedious process of extracting large collections of facts (e.g., names of scientists or politicians) from the Web in an unsupervised, domain-independent, and scalable manner. The paper presents an overview of KNOW-ITALL's novel architecture and design principles, emphasizing its distinctive ability to extract information without any hand-labeled training examples. In its first major run, KNOW-ITALL extracted over 50,000 facts, but suggested a challenge: How can we improve KNOW-ITALL's recall and extraction rate without sacrificing precision? This paper presents three distinct ways to address this challenge and evaluates their performance. Pattern Learning learns domain-specific extraction rules, which enable additional extractions. Subclass Extraction automatically identifies sub-classes in order to boost recall. List Extraction locates lists of class instances, learns a "wrapper" for each list, and extracts elements of each list. Since each method bootstraps from KNOWITALL's domainindependent methods, the methods also obviate hand-labeled training examples. The paper reports on experiments, focused on named-entity extraction, that measure the relative efficacy of each method and demonstrate their synergy. In concert, our methods gave KNOW-ITALL a 4-fold to 8-fold increase in recall, while maintaining high precision, and discovered over 10,000 cities missing from the Tipster Gazetteer.

Manually querying search engines in order to accumulate a large body of factual information is a tedious, error-prone process of piecemeal search. Search engines retrieve and rank potentially relevant documents for human perusal, but do not extract facts, assess confidence, or fuse information from multiple documents. This paper introduces KNOWITALL, a system that aims to automate the tedious process of extracting large collections of facts from the web in an autonomous, domain-independent, and scalable manner.The paper describes preliminary experiments in which an instance of KNOWITALL, running for four days on a single machine, was able to automatically extract 54,753 facts. KNOWITALL associates a probability with each fact enabling it to trade off precision and recall. The paper analyzes KNOWITALL's architecture and reports on lessons learned for the design of large-scale information extraction systems.

Traditionally, Information Extraction (IE) has focused on satisfying precise, narrow, pre-specified requests from small homogeneous corpora (e.g., extract the location and time of seminars from a set of announcements). Shifting to a new domain requires the user to name the target relations and to manually create new extraction rules or hand-tag new training examples. This manual labor scales linearly with the number of target relations. This paper introduces Open IE (OIE), a new extraction paradigm where the system makes a single data-driven pass over its corpus and extracts a large set of relational tuples without requiring any human input. The paper also introduces TEXTRUNNER, a fully implemented, highly scalable OIE system where the tuples are assigned a probability and indexed to support efficient extraction and exploration via user queries. We report on experiments over a 9,000,000 Web page corpus that compare TEXTRUNNER with KNOWITALL, a state-of-the-art Web IE system. TEXTRUNNER achieves an error reduction of 33% on a comparable set of extractions. Furthermore, in the amount of time it takes KNOWITALL to perform extraction for a handful of pre-specified relations, TEXTRUNNER extracts a far broader set of facts reflecting orders of magnitude more relations, discovered on the fly. We report statistics on TEXTRUNNER's 11,000,000 highest probability tuples, and show that they contain over 1,000,000 concrete facts and over 6,500,000 more abstract assertions.

We apply BERT to coreference resolution, achieving strong improvements on the OntoNotes (+3.9 F1) and GAP (+11.5 F1) benchmarks. A qualitative analysis of model predictions indicates that, compared to ELMo and BERT-base, BERT-large is particularly better at distinguishing between related but distinct entities (e.g., President and CEO). However, there is still room for improvement in modeling document-level context, conversations, and mention paraphrasing. Our code and models are publicly available 1 .

Today's computer-human interfaces are typically designed with the assumption that they are going to be used by an able-bodied person, who is using a typical set of input and output devices, who has typical perceptual and cognitive abilities, and who is sitting in a stable, warm environment. Any deviation from these assumptions may drastically hamper the person's effectiveness-not because of any inherent barrier to interaction, but because of a mismatch between the person's effective abilities and the assumptions underlying the interface design. We argue that automatic personalized interface generation is a feasible and scalable solution to this challenge. We present our Supple system, which can automatically generate interfaces adapted to a person's devices, tasks, preferences, and abilities. In this paper we formally define interface generation as an optimization problem and demonstrate that, despite a large solution space (of up to 10 17 possible interfaces), the problem is computationally feasible. In fact, for a particular class of cost functions, Supple produces exact solutions in under a second for most cases, and in a little over a minute in the worst case encountered, thus enabling run-time generation of user interfaces. We further show how several different design criteria can be expressed in the cost function, enabling different kinds of personalization. We also demonstrate how this approach enables extensive user-and system-initiated run-time adaptations to the interfaces after they have been generated. Supple is not intended to replace human user interface designers-instead, it offers alternative user interfaces for those people whose devices, tasks, preferences, and abilities are not sufficiently addressed by the hand-crafted designs. Indeed, the results of our study show that, compared to manufacturers' defaults, interfaces automatically generated by Supple significantly improve speed, accuracy and satisfaction of people with motor impairments.

No abstract