Search Results for author: Vicky Zayats

Found 11 papers, 1 papers with code

MultiTurnCleanup: A Benchmark for Multi-Turn Spoken Conversational Transcript Cleanup

1 code implementation19 May 2023 Hua Shen, Vicky Zayats, Johann C. Rocholl, Daniel D. Walker, Dirk Padfield

Current disfluency detection models focus on individual utterances each from a single speaker.

Teaching BERT to Wait: Balancing Accuracy and Latency for Streaming Disfluency Detection

no code implementations NAACL 2022 Angelica Chen, Vicky Zayats, Daniel D. Walker, Dirk Padfield

In modern interactive speech-based systems, speech is consumed and transcribed incrementally prior to having disfluencies removed.

Machine Translation

Representations for Question Answering from Documents with Tables and Text

no code implementations EACL 2021 Vicky Zayats, Kristina Toutanova, Mari Ostendorf

Tables in Web documents are pervasive and can be directly used to answer many of the queries searched on the Web, motivating their integration in question answering.

Natural Questions Question Answering

Disfluencies and Human Speech Transcription Errors

no code implementations8 Apr 2019 Vicky Zayats, Trang Tran, Richard Wright, Courtney Mansfield, Mari Ostendorf

This paper explores contexts associated with errors in transcrip-tion of spontaneous speech, shedding light on human perceptionof disfluencies and other conversational speech phenomena.

Robust cross-domain disfluency detection with pattern match networks

no code implementations17 Nov 2018 Vicky Zayats, Mari Ostendorf

In this paper we introduce a novel pattern match neural network architecture that uses neighbor similarity scores as features, eliminating the need for feature engineering in a disfluency detection task.

Feature Engineering

Conversation Modeling on Reddit using a Graph-Structured LSTM

no code implementations TACL 2018 Vicky Zayats, Mari Ostendorf

This paper presents a novel approach for modeling threaded discussions on social media using a graph-structured bidirectional LSTM which represents both hierarchical and temporal conversation structure.

Disfluency Detection using a Bidirectional LSTM

no code implementations12 Apr 2016 Vicky Zayats, Mari Ostendorf, Hannaneh Hajishirzi

We introduce a new approach for disfluency detection using a Bidirectional Long-Short Term Memory neural network (BLSTM).

Cannot find the paper you are looking for? You can Submit a new open access paper.