Abstract
In this paper, we study how word-like units are represented and activated in a recurrent neural model of visually grounded speech. The model used in our experiments is trained to project an image and its spoken description in a common representation space. We show that a recurrent model trained on spoken sentences implicitly segments its input into word-like units and reliably maps them to their correct visual referents. We introduce a methodology originating from linguistics to analyse the representation learned by neural networks – the gating paradigm – and show that the correct representation of a word is only activated if the network has access to first phoneme of the target word, suggesting that the network does not rely on a global acoustic pattern. Furthermore, we find out that not all speech frames (MFCC vectors in our case) play an equal role in the final encoded representation of a given word, but that some frames have a crucial effect on it. Finally we suggest that word representation could be activated through a process of lexical competition.- Anthology ID:
- K19-1032
- Volume:
- Proceedings of the 23rd Conference on Computational Natural Language Learning (CoNLL)
- Month:
- November
- Year:
- 2019
- Address:
- Hong Kong, China
- Editors:
- Mohit Bansal, Aline Villavicencio
- Venue:
- CoNLL
- SIG:
- SIGNLL
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 339–348
- Language:
- URL:
- https://aclanthology.org/K19-1032
- DOI:
- 10.18653/v1/K19-1032
- Cite (ACL):
- William N. Havard, Jean-Pierre Chevrot, and Laurent Besacier. 2019. Word Recognition, Competition, and Activation in a Model of Visually Grounded Speech. In Proceedings of the 23rd Conference on Computational Natural Language Learning (CoNLL), pages 339–348, Hong Kong, China. Association for Computational Linguistics.
- Cite (Informal):
- Word Recognition, Competition, and Activation in a Model of Visually Grounded Speech (Havard et al., CoNLL 2019)
- PDF:
- https://preview.aclanthology.org/nschneid-patch-3/K19-1032.pdf
- Data
- MS COCO