• Publications
  • Influence
Improving Text-to-SQL Evaluation Methodology
TLDR
It is shown that the current division of data into training and test sets measures robustness to variations in the way questions are asked, but only partially tests how well systems generalize to new queries, and proposes a complementary dataset split for evaluation of future work. Expand
Reasoning about Goals, Steps, and Temporal Ordering with WikiHow
TLDR
This work proposes a suite of reasoning tasks on two types of relations between procedural events: goal-step relations and step-step temporal relations, and introduces a dataset targeting these two relations based on wikiHow, a website of instructional how-to articles. Expand
Intent Detection with WikiHow
TLDR
A suite of pretrained intent detection models which can predict a broad range of intended goals from many actions because they are trained on wikiHow, a comprehensive instructional website are presented. Expand
Goal-Oriented Script Construction
TLDR
This work proposes the GoalOriented Script Construction task, where a model produces a sequence of steps to accomplish a given goal, and considers both a generationbased approach using a language model and a retrieval-based approach. Expand
Visual Goal-Step Inference using wikiHow
TLDR
This work proposes the Visual Goal-Step Inference (VGSI) task where a model is given a textual goal and must choose a plausible step towards that goal from among four candidate images, and introduces the visual analogue. Expand
Direct Network Transfer: Transfer Learning of Sentence Embeddings for Semantic Similarity
TLDR
It is shown that applying direct network transfer to existing encoders can lead to state-of-the-art performance, and that the choice of transfer learning setting greatly affects the performance in many cases, and differs by encoder and dataset. Expand
Small but Mighty: New Benchmarks for Split and Rephrase
TLDR
It is found that the widely used benchmark dataset universally contains easily exploitable syntactic cues caused by its automatic generation process, and it is shown that even a simple rule-based model can perform on par with the state-of-the-art model. Expand
Multi-Label Transfer Learning for Multi-Relational Semantic Similarity
TLDR
A multi-label transfer learning approach based on LSTM to make predictions for several relations simultaneously and aggregate the losses to update the parameters and achieves state-of-the-art performance on all but one relation of the Human Activity Phrase dataset. Expand
Entity and Event Extraction from Scratch Using Minimal Training Data
TLDR
This work is building the first step of the overall system, which involves translating all the raw documents, as well as transcribing and translating audio and video data, and building a graph from all the entities, events, and relations. Expand