VSE++: Improving Visual-Semantic Embeddings with Hard Negatives

by   Fartash Faghri, et al.

We present a new technique for learning visual-semantic embeddings for cross-modal retrieval. Inspired by the use of hard negatives in structured prediction, and ranking loss functions used in retrieval, we introduce a simple change to common loss functions used to learn multi-modal embeddings. That, combined with fine-tuning and the use of augmented data, yields significant gains in retrieval performance. We showcase our approach, dubbed VSE++, on the MS-COCO and Flickr30K datasets, using ablation studies and comparisons with existing methods. On MS-COCO our approach outperforms state-of-the-art methods by 8.8



page 8


Contrastive Learning of Visual-Semantic Embeddings

Contrastive learning is a powerful technique to learn representations th...

Cross-modal Center Loss

Cross-modal retrieval aims to learn discriminative and modal-invariant f...

Multi-Modal Retrieval using Graph Neural Networks

Most real world applications of image retrieval such as Adobe Stock, whi...

Crisscrossed Captions: Extended Intramodal and Intermodal Semantic Similarity Judgments for MS-COCO

Image captioning datasets have proven useful for multimodal representati...

Learning TFIDF Enhanced Joint Embedding for Recipe-Image Cross-Modal Retrieval Service

It is widely acknowledged that learning joint embeddings of recipes with...

Exploring Uncertainty in Conditional Multi-Modal Retrieval Systems

We cast visual retrieval as a regression problem by posing triplet loss ...

Where Does the Performance Improvement Come From? – A Reproducibility Concern about Image-Text Retrieval

This paper seeks to provide the information retrieval community with som...

Code Repositories


PyTorch Code for the paper "VSE++: Improving Visual-Semantic Embeddings with Hard Negatives"

view repo
This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.