LITE: Intent-based Task Representation Learning Using Weak Supervision

Naoki Otani, Michael Gamon, Sujay Kumar Jauhar, Mei Yang, Sri Raghu Malireddi, Oriana Riva


Abstract
Users write to-dos as personal notes to themselves, about things they need to complete, remember or organize. To-do texts are usually short and under-specified, which poses a challenge for current text representation models. Yet, understanding and representing their meaning is the first step towards providing intelligent assistance for to-do management. We address this problem by proposing a neural multi-task learning framework, LITE, which extracts representations of English to-do tasks with a multi-head attention mechanism on top of a pre-trained text encoder. To adapt representation models to to-do texts, we collect weak-supervision labels from semantically rich external resources (e.g., dynamic commonsense knowledge bases), following the principle that to-do tasks with similar intents have similar labels. We then train the model on multiple generative/predictive training objectives jointly. We evaluate our representation model on four downstream tasks and show that our approach consistently improves performance over baseline models, achieving error reduction of up to 38.7%.
Anthology ID:
2022.naacl-main.172
Volume:
Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies
Month:
July
Year:
2022
Address:
Seattle, United States
Venue:
NAACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
2410–2424
Language:
URL:
https://aclanthology.org/2022.naacl-main.172
DOI:
10.18653/v1/2022.naacl-main.172
Bibkey:
Cite (ACL):
Naoki Otani, Michael Gamon, Sujay Kumar Jauhar, Mei Yang, Sri Raghu Malireddi, and Oriana Riva. 2022. LITE: Intent-based Task Representation Learning Using Weak Supervision. In Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 2410–2424, Seattle, United States. Association for Computational Linguistics.
Cite (Informal):
LITE: Intent-based Task Representation Learning Using Weak Supervision (Otani et al., NAACL 2022)
Copy Citation:
PDF:
https://aclanthology.org/2022.naacl-main.172.pdf
Code
 microsoft/intent-based-task-representation-learning