Contrastive Learning for Weakly Supervised Phrase Grounding
Autor: | Gupta, Tanmay, Vahdat, Arash, Chechik, Gal, Yang, Xiaodong, Kautz, Jan, Hoiem, Derek |
---|---|
Rok vydání: | 2020 |
Předmět: | |
Druh dokumentu: | Working Paper |
Popis: | Phrase grounding, the problem of associating image regions to caption words, is a crucial component of vision-language tasks. We show that phrase grounding can be learned by optimizing word-region attention to maximize a lower bound on mutual information between images and caption words. Given pairs of images and captions, we maximize compatibility of the attention-weighted regions and the words in the corresponding caption, compared to non-corresponding pairs of images and captions. A key idea is to construct effective negative captions for learning through language model guided word substitutions. Training with our negatives yields a $\sim10\%$ absolute gain in accuracy over randomly-sampled negatives from the training data. Our weakly supervised phrase grounding model trained on COCO-Captions shows a healthy gain of $5.7\%$ to achieve $76.7\%$ accuracy on Flickr30K Entities benchmark. Comment: ECCV 2020 (spotlight paper), Project page: http://tanmaygupta.info/info-ground |
Databáze: | arXiv |
Externí odkaz: |