Abstract
We demonstrate that a continuous relaxation of the argmax operation can be used to create a differentiable approximation to greedy decoding in sequence-to-sequence (seq2seq) models. By incorporating this approximation into the scheduled sampling training procedure–a well-known technique for correcting exposure bias–we introduce a new training objective that is continuous and differentiable everywhere and can provide informative gradients near points where previous decoding decisions change their value. By using a related approximation, we also demonstrate a similar approach to sampled-based training. We show that our approach outperforms both standard cross-entropy training and scheduled sampling procedures in two sequence prediction tasks: named entity recognition and machine translation.- Anthology ID:
- P17-2058
- Volume:
- Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)
- Month:
- July
- Year:
- 2017
- Address:
- Vancouver, Canada
- Editors:
- Regina Barzilay, Min-Yen Kan
- Venue:
- ACL
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 366–371
- Language:
- URL:
- https://aclanthology.org/P17-2058
- DOI:
- 10.18653/v1/P17-2058
- Cite (ACL):
- Kartik Goyal, Chris Dyer, and Taylor Berg-Kirkpatrick. 2017. Differentiable Scheduled Sampling for Credit Assignment. In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 366–371, Vancouver, Canada. Association for Computational Linguistics.
- Cite (Informal):
- Differentiable Scheduled Sampling for Credit Assignment (Goyal et al., ACL 2017)
- PDF:
- https://preview.aclanthology.org/nschneid-patch-4/P17-2058.pdf