Improving Personalized Search with Regularized Low-Rank Parameter Updates

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21730%2F25%3A00388538" target="_blank" >RIV/68407700:21730/25:00388538 - isvavai.cz</a>

Výsledek na webu

<a href="https://doi.org/10.1109/CVPR52734.2025.01839" target="_blank" >https://doi.org/10.1109/CVPR52734.2025.01839</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1109/CVPR52734.2025.01839" target="_blank" >10.1109/CVPR52734.2025.01839</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Improving Personalized Search with Regularized Low-Rank Parameter Updates

Popis výsledku v původním jazyce

Personalized vision-language retrieval seeks to recognize new concepts (e.g., “my dog Fido”) from only a few examples. This task is challenging because it requires not only learning a new concept from a few images, but also integrating the personal and general knowledge together to recognize the concept in different contexts. In this paper, we show how to effectively adapt the internal representation of a vision-language dual encoder model for personalized vision-language retrieval. We find that regularized low-rank adaption of a small set of parameters in the language encoder’s final layer serves as a highly effective alternative to textual inversion for recognizing the personal concept while preserving general knowledge. Additionally, we explore strategies for combining parameters of multiple learned personal concepts, finding that parameter addition is effective. To evaluate how well general knowledge is preserved in a finetuned representation, we introduce a metric that measures image retrieval accuracy based on captions generated by a vision language model (VLM). Our approach achieves state-of-the-art accuracy on two benchmarks for personalized image retrieval with natural language queries – DeepFashion2 and ConCon-Chi – outperforming the prior art by 4% - 22% on personal retrievals.

Název v anglickém jazyce

Improving Personalized Search with Regularized Low-Rank Parameter Updates

Popis výsledku anglicky

Personalized vision-language retrieval seeks to recognize new concepts (e.g., “my dog Fido”) from only a few examples. This task is challenging because it requires not only learning a new concept from a few images, but also integrating the personal and general knowledge together to recognize the concept in different contexts. In this paper, we show how to effectively adapt the internal representation of a vision-language dual encoder model for personalized vision-language retrieval. We find that regularized low-rank adaption of a small set of parameters in the language encoder’s final layer serves as a highly effective alternative to textual inversion for recognizing the personal concept while preserving general knowledge. Additionally, we explore strategies for combining parameters of multiple learned personal concepts, finding that parameter addition is effective. To evaluate how well general knowledge is preserved in a finetuned representation, we introduce a metric that measures image retrieval accuracy based on captions generated by a vision language model (VLM). Our approach achieves state-of-the-art accuracy on two benchmarks for personalized image retrieval with natural language queries – DeepFashion2 and ConCon-Chi – outperforming the prior art by 4% - 22% on personal retrievals.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

—

Návaznosti

N - Vyzkumna aktivita podporovana z neverejnych zdroju

Rok uplatnění

2025

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

2025 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

ISBN

979-8-3315-4364-8

ISSN

1063-6919

e-ISSN

2575-7075

Počet stran výsledku

10

Strana od-do

19748-19757

Název nakladatele

IEEE Computer Society

Místo vydání

Los Alamitos

Místo konání akce

Nashville

Datum konání akce

11. 6. 2025

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

001601158200161