Revisit Visual Prompt Tuning: The Expressiveness of Prompt Experts

Minh Le, Anh Nguyen, Huy Nguyen, Chau Nguyen, Anh Tran, Nhat Ho · 2025

Read Revisit Visual Prompt Tuning: The Expressiveness of Prompt Experts by Minh Le, Anh Nguyen, Huy Nguyen, Chau Nguyen, Anh Tran, Nhat Ho (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-18936.txt.

Description

Minh Le and five co-authors deepen the theory of visual prompt tuning in twenty twenty-five. Prompt tuning adapts a pretrained vision model by adding learnable prompt tokens. It works well in practice, but why it works is still being worked out. A recent link ties prompt methods to mixture-of-experts models, where each attention head acts like several experts.

Seen that way, prompt tuning adds new prompt experts into those structures. The authors find a flaw: those experts are static, so their expressiveness is limited. Their method, VAPT, gives prompt experts more flexibility while keeping the parameter count low. Readers get both a clearer theory of a popular trick and a stronger version of it.

Catalog record

CREATORMinh Le, Anh Nguyen, Huy Nguyen, Chau Nguyen, Anh Tran, Nhat Ho
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE168 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-18936

Browse the full library · About the Tower of Babel