Probing Cross-Modal Representations in Multi-Step Relational Reasoning
Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review
Documents
- Fulltext
Final published version, 674 KB, PDF document
We investigate the representations learned by vision and language models in tasks that require relational reasoning. Focusing on the problem of assessing the relative size of objects in abstract visual contexts, we analyse both one-step and two-step reasoning. For the latter, we construct a new dataset of three-image scenes and define a task that requires reasoning at the level of the individual images and across images in a scene. We probe the learned model representations using diagnostic classifiers. Our experiments show that pretrained multimodal transformer-based architectures can perform higher-level relational reasoning, and are able to learn representations for novel tasks and data that are very different from what was seen in pretraining.
Original language | English |
---|---|
Title of host publication | Proceedings of the 6th Workshop on Representation Learning for NLP (RepL4NLP-2021) |
Publisher | Association for Computational Linguistics |
Publication date | 2021 |
Pages | 152-162 |
DOIs | |
Publication status | Published - 2021 |
Event | 6th Workshop on Representation Learning for NLP (RepL4NLP-2021) - Online, Online Duration: 1 Aug 2021 → 1 Aug 2021 |
Conference
Conference | 6th Workshop on Representation Learning for NLP (RepL4NLP-2021) |
---|---|
Location | Online |
By | Online |
Periode | 01/08/2021 → 01/08/2021 |
ID: 299038005