Towards Making a Dependency Parser See

09/03/2019
by   Michalina Strzyz, et al.
0

We explore whether it is possible to leverage eye-tracking data in an RNN dependency parser (for English) when such information is only available during training, i.e., no aggregated or token-level gaze features are used at inference time. To do so, we train a multitask learning model that parses sentences as sequence labeling and leverages gaze features as auxiliary tasks. Our method also learns to train from disjoint datasets, i.e. it can be used to test whether already collected gaze features are useful to improve the performance on new non-gazed annotated treebanks. Accuracy gains are modest but positive, showing the feasibility of the approach. It can serve as a first step towards architectures that can better leverage eye-tracking data or other complementary information available only for training sentences, possibly leading to improvements in syntactic parsing.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
02/26/2019

Entity Recognition at First Sight: Improving NER with Eye Movement Information

Previous research shows that eye-tracking data contains information abou...
research
01/18/2021

LNSMM: Eye Gaze Estimation With Local Network Share Multiview Multitask

Eye gaze estimation has become increasingly significant in computer visi...
research
04/30/2020

Mutlitask Learning for Cross-Lingual Transfer of Semantic Dependencies

We describe a method for developing broad-coverage semantic dependency p...
research
02/04/2016

Many Languages, One Parser

We train one multilingual model for dependency parsing and use it to par...
research
06/05/2020

RIT-Eyes: Rendering of near-eye images for eye-tracking applications

Deep neural networks for video-based eye tracking have demonstrated resi...
research
03/23/2015

Yara Parser: A Fast and Accurate Dependency Parser

Dependency parsers are among the most crucial tools in natural language ...

Please sign up or login with your details

Forgot password? Click here to reset