Adversarial hubs can be generated to be retrieved as top-1 for over 84% of test queries in text-to-image retrieval, far exceeding natural hubs.
Improving zero-shot learning by mitigating the hubness problem
5 Pith papers cite this work. Polarity classification is still indexing.
5
Pith papers citing it
abstract
The zero-shot paradigm exploits vector-based word representations extracted from text corpora with unsupervised methods to learn general mapping functions from other feature spaces onto word space, where the words associated to the nearest neighbours of the mapped vectors are used as their linguistic labels. We show that the neighbourhoods of the mapped elements are strongly polluted by hubs, vectors that tend to be near a high proportion of items, pushing their correct labels down the neighbour list. After illustrating the problem empirically, we propose a simple method to correct it by taking the proximity distribution of potential neighbours across many mapped vectors into account. We show that this correction leads to consistent improvements in realistic zero-shot experiments in the cross-lingual, image labeling and image retrieval domains.
citation-role summary
background 1
citation-polarity summary
roles
background 1polarities
background 1representative citing papers
Hubness dominates as the causal driver of cross-lingual retrieval asymmetry in multilingual embeddings, with CSLS correction closing most of the reciprocity gap across five models and a 6518-item parallel corpus.
SATTC improves top-k accuracy in cross-subject EEG-to-image retrieval by fusing geometric whitening and structural nearest-neighbor experts on the similarity matrix without labels.
UAGA aligns two graph embedding spaces via adversarial training in a fully unsupervised setting, with an incremental extension iUAGA that uses discovered pseudo-anchors to refine both embeddings and alignments.
A single hub text can unreasonably match many images in CLIP-based similarity, exposing vulnerabilities in cross-modal encoders for caption evaluation and retrieval.
citing papers explorer
-
Adversarial Hubness in Multi-Modal Retrieval
Adversarial hubs can be generated to be retrieved as top-1 for over 84% of test queries in text-to-image retrieval, far exceeding natural hubs.
-
Hubness, Not Anisotropy, Drives Cross-Lingual Retrieval Asymmetry in Multilingual Embedding Models
Hubness dominates as the causal driver of cross-lingual retrieval asymmetry in multilingual embeddings, with CSLS correction closing most of the reciprocity gap across five models and a 6518-item parallel corpus.
-
SATTC: Structure-Aware Label-Free Test-Time Calibration for Cross-Subject EEG-to-Image Retrieval
SATTC improves top-k accuracy in cross-subject EEG-to-image retrieval by fusing geometric whitening and structural nearest-neighbor experts on the similarity matrix without labels.
-
Unsupervised Adversarial Graph Alignment with Graph Embedding
UAGA aligns two graph embedding spaces via adversarial training in a fully unsupervised setting, with an incremental extension iUAGA that uses discovered pseudo-anchors to refine both embeddings and alignments.
-
One Single Hub Text Breaks CLIP: Identifying Vulnerabilities in Cross-Modal Encoders via Hubness
A single hub text can unreasonably match many images in CLIP-based similarity, exposing vulnerabilities in cross-modal encoders for caption evaluation and retrieval.