End-to-end Image Captioning Exploits Multimodal Distributional Similarity

September 11, 2018 ยท Declared Dead ยท ๐Ÿ› British Machine Vision Conference

๐Ÿ‘ป CAUSE OF DEATH: Ghosted
No code link whatsoever

"No code URL or promise found in abstract"

Evidence collected by the PWNC Scanner

Authors Pranava Madhyastha, Josiah Wang, Lucia Specia arXiv ID 1809.04144 Category cs.CV: Computer Vision Citations 7 Venue British Machine Vision Conference Last Checked 3 months ago
Abstract
We hypothesize that end-to-end neural image captioning systems work seemingly well because they exploit and learn `distributional similarity' in a multimodal feature space by mapping a test image to similar training images in this space and generating a caption from the same space. To validate our hypothesis, we focus on the `image' side of image captioning, and vary the input image representation but keep the RNN text generation component of a CNN-RNN model constant. Our analysis indicates that image captioning models (i) are capable of separating structure from noisy input representations; (ii) suffer virtually no significant performance loss when a high dimensional representation is compressed to a lower dimensional space; (iii) cluster images with similar visual and linguistic information together. Our findings indicate that our distributional similarity hypothesis holds. We conclude that regardless of the image representation used image captioning systems seem to match images and generate captions in a learned joint image-text semantic subspace.
Community shame:
Not yet rated
Community Contributions

Found the code? Know the venue? Think something is wrong? Let us know!

๐Ÿ“œ Similar Papers

In the same crypt โ€” Computer Vision

Died the same way โ€” ๐Ÿ‘ป Ghosted