Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2412.19417 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866916543571427328 |
|---|---|
| author | Zhao, Shu Yu, Tan Hao, Xiaoshuai Ma, Wenchao Narayanan, Vijaykrishnan |
| author_facet | Zhao, Shu Yu, Tan Hao, Xiaoshuai Ma, Wenchao Narayanan, Vijaykrishnan |
| contents | Deep hashing has been widely used for large-scale approximate nearest neighbor search due to its storage and search efficiency. However, existing deep hashing methods predominantly rely on abundant training data, leaving the more challenging scenario of low-resource adaptation for deep hashing relatively underexplored. This setting involves adapting pre-trained models to downstream tasks with only an extremely small number of training samples available. Our preliminary benchmarks reveal that current methods suffer significant performance degradation due to the distribution shift caused by limited training samples. To address these challenges, we introduce Class-Calibration LoRA (CLoRA), a novel plug-and-play approach that dynamically constructs low-rank adaptation matrices by leveraging class-level textual knowledge embeddings. CLoRA effectively incorporates prior class knowledge as anchors, enabling parameter-efficient fine-tuning while maintaining the original data distribution. Furthermore, we propose Knowledge-Guided Discrete Optimization (KIDDO), a framework to utilize class knowledge to compensate for the scarcity of visual information and enhance the discriminability of hash codes. Extensive experiments demonstrate that our proposed method, Knowledge- Anchored Low-Resource Adaptation Hashing (KALAHash), significantly boosts retrieval performance and achieves a 4x data efficiency in low-resource scenarios. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2412_19417 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | KALAHash: Knowledge-Anchored Low-Resource Adaptation for Deep Hashing Zhao, Shu Yu, Tan Hao, Xiaoshuai Ma, Wenchao Narayanan, Vijaykrishnan Computer Vision and Pattern Recognition Deep hashing has been widely used for large-scale approximate nearest neighbor search due to its storage and search efficiency. However, existing deep hashing methods predominantly rely on abundant training data, leaving the more challenging scenario of low-resource adaptation for deep hashing relatively underexplored. This setting involves adapting pre-trained models to downstream tasks with only an extremely small number of training samples available. Our preliminary benchmarks reveal that current methods suffer significant performance degradation due to the distribution shift caused by limited training samples. To address these challenges, we introduce Class-Calibration LoRA (CLoRA), a novel plug-and-play approach that dynamically constructs low-rank adaptation matrices by leveraging class-level textual knowledge embeddings. CLoRA effectively incorporates prior class knowledge as anchors, enabling parameter-efficient fine-tuning while maintaining the original data distribution. Furthermore, we propose Knowledge-Guided Discrete Optimization (KIDDO), a framework to utilize class knowledge to compensate for the scarcity of visual information and enhance the discriminability of hash codes. Extensive experiments demonstrate that our proposed method, Knowledge- Anchored Low-Resource Adaptation Hashing (KALAHash), significantly boosts retrieval performance and achieves a 4x data efficiency in low-resource scenarios. |
| title | KALAHash: Knowledge-Anchored Low-Resource Adaptation for Deep Hashing |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2412.19417 |