Details
| Original language | English |
|---|---|
| Title of host publication | WSDM 2021 - Proceedings of the 14th ACM International Conference on Web Search and Data Mining |
| Pages | 418-426 |
| Number of pages | 9 |
| ISBN (electronic) | 9781450382977 |
| Publication status | Published - 8 Mar 2021 |
| Event | 14th ACM International Conference on Web Search and Data Mining, WSDM 2021 - online, Virtual, Online, Israel Duration: 8 Mar 2021 → 12 Mar 2021 |
Abstract
A desirable property of learning systems is to be both effective and interpretable. Towards this goal, recent models have been proposed that first generate an extractive explanation from the input text and then generate a prediction on just the explanation called explain-then-predict models. These models primarily consider the task input as a supervision signal in learning an extractive explanation and do not effectively integrate rationales data as an additional inductive bias to improve task performance. We propose a novel yet simple approach ExPred, which uses multi-task learning in the explanation generation phase effectively trading-off explanation and prediction losses. Next, we use another prediction network on just the extracted explanations for optimizing the task performance. We conduct an extensive evaluation of our approach on three diverse language datasets - sentiment classification, fact-checking, and question answering - and find that we substantially outperform existing approaches.
Keywords
- explanation, interpretable by design, learning with rationales, machine learning interpretation, multitask learning
ASJC Scopus subject areas
- Computer Science(all)
- Computer Networks and Communications
- Computer Science(all)
- Computer Science Applications
- Computer Science(all)
- Software
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
WSDM 2021 - Proceedings of the 14th ACM International Conference on Web Search and Data Mining. 2021. p. 418-426.
Research output: Chapter in book/report/conference proceeding › Conference contribution › Research › peer review
}
TY - GEN
T1 - Explain and Predict, and then Predict Again
AU - Zhang, Zijian
AU - Rudra, Koustav
AU - Anand, Avishek
N1 - Funding Information: Funding for this project was in part provided by the European Union's Horizon 2020 research and innovation program under grant agreement No 832921, and No 871042.
PY - 2021/3/8
Y1 - 2021/3/8
N2 - A desirable property of learning systems is to be both effective and interpretable. Towards this goal, recent models have been proposed that first generate an extractive explanation from the input text and then generate a prediction on just the explanation called explain-then-predict models. These models primarily consider the task input as a supervision signal in learning an extractive explanation and do not effectively integrate rationales data as an additional inductive bias to improve task performance. We propose a novel yet simple approach ExPred, which uses multi-task learning in the explanation generation phase effectively trading-off explanation and prediction losses. Next, we use another prediction network on just the extracted explanations for optimizing the task performance. We conduct an extensive evaluation of our approach on three diverse language datasets - sentiment classification, fact-checking, and question answering - and find that we substantially outperform existing approaches.
AB - A desirable property of learning systems is to be both effective and interpretable. Towards this goal, recent models have been proposed that first generate an extractive explanation from the input text and then generate a prediction on just the explanation called explain-then-predict models. These models primarily consider the task input as a supervision signal in learning an extractive explanation and do not effectively integrate rationales data as an additional inductive bias to improve task performance. We propose a novel yet simple approach ExPred, which uses multi-task learning in the explanation generation phase effectively trading-off explanation and prediction losses. Next, we use another prediction network on just the extracted explanations for optimizing the task performance. We conduct an extensive evaluation of our approach on three diverse language datasets - sentiment classification, fact-checking, and question answering - and find that we substantially outperform existing approaches.
KW - explanation
KW - interpretable by design
KW - learning with rationales
KW - machine learning interpretation
KW - multitask learning
UR - http://www.scopus.com/inward/record.url?scp=85103007189&partnerID=8YFLogxK
U2 - 10.1145/3437963.3441758
DO - 10.1145/3437963.3441758
M3 - Conference contribution
AN - SCOPUS:85103007189
SP - 418
EP - 426
BT - WSDM 2021 - Proceedings of the 14th ACM International Conference on Web Search and Data Mining
T2 - 14th ACM International Conference on Web Search and Data Mining, WSDM 2021
Y2 - 8 March 2021 through 12 March 2021
ER -