ReBaPL: Repulsive Bayesian Prompt Learning

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Bendou, Yassir, Ezzahir, Omar, Montesuma, Eduardo Fernandes, Mahuas, Gabriel, Shevchenko, Victoria, Gartrell, Mike
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866911547696087040
author Bendou, Yassir
Ezzahir, Omar
Montesuma, Eduardo Fernandes
Mahuas, Gabriel
Shevchenko, Victoria
Gartrell, Mike
author_facet Bendou, Yassir
Ezzahir, Omar
Montesuma, Eduardo Fernandes
Mahuas, Gabriel
Shevchenko, Victoria
Gartrell, Mike
contents Prompt learning has emerged as an effective technique for fine-tuning large-scale foundation models for downstream tasks. However, conventional prompt learning methods are prone to overfitting and can struggle with out-of-distribution generalization. To address these limitations, Bayesian prompt learning has been proposed, which frames prompt optimization as a Bayesian inference problem to enhance robustness. This paper introduces Repulsive Bayesian Prompt Learning (ReBaPL), a novel method for Bayesian prompt learning, designed to efficiently explore the complex and often multimodal posterior landscape of prompts. Our method integrates a cyclical step-size schedule with a stochastic gradient Hamiltonian Monte Carlo (SGHMC) algorithm, enabling alternating phases of exploration to discover new modes, and exploitation to refine existing modes. Furthermore, we introduce a repulsive force derived from a potential function over probability metrics (including Maximum Mean Discrepancy and Wasserstein distance) computed on the distributions of representations produced by different prompts. This representation-space repulsion diversifies exploration and prevents premature collapse to a single mode. Our approach allows for a more comprehensive characterization of the prompt posterior distribution, leading to improved generalization. In contrast to prior Bayesian prompt learning methods, our method provides a modular plug-and-play Bayesian extension of any existing prompt learning method based on maximum likelihood estimation. We demonstrate the efficacy of ReBaPL on several benchmark datasets, showing superior performance over state-of-the-art prompt learning methods.
format Preprint
id arxiv_https___arxiv_org_abs_2511_17339
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle ReBaPL: Repulsive Bayesian Prompt Learning
Bendou, Yassir
Ezzahir, Omar
Montesuma, Eduardo Fernandes
Mahuas, Gabriel
Shevchenko, Victoria
Gartrell, Mike
Machine Learning
Prompt learning has emerged as an effective technique for fine-tuning large-scale foundation models for downstream tasks. However, conventional prompt learning methods are prone to overfitting and can struggle with out-of-distribution generalization. To address these limitations, Bayesian prompt learning has been proposed, which frames prompt optimization as a Bayesian inference problem to enhance robustness. This paper introduces Repulsive Bayesian Prompt Learning (ReBaPL), a novel method for Bayesian prompt learning, designed to efficiently explore the complex and often multimodal posterior landscape of prompts. Our method integrates a cyclical step-size schedule with a stochastic gradient Hamiltonian Monte Carlo (SGHMC) algorithm, enabling alternating phases of exploration to discover new modes, and exploitation to refine existing modes. Furthermore, we introduce a repulsive force derived from a potential function over probability metrics (including Maximum Mean Discrepancy and Wasserstein distance) computed on the distributions of representations produced by different prompts. This representation-space repulsion diversifies exploration and prevents premature collapse to a single mode. Our approach allows for a more comprehensive characterization of the prompt posterior distribution, leading to improved generalization. In contrast to prior Bayesian prompt learning methods, our method provides a modular plug-and-play Bayesian extension of any existing prompt learning method based on maximum likelihood estimation. We demonstrate the efficacy of ReBaPL on several benchmark datasets, showing superior performance over state-of-the-art prompt learning methods.
title ReBaPL: Repulsive Bayesian Prompt Learning
topic Machine Learning
url https://arxiv.org/abs/2511.17339