Beyond Any-Shot Adaptation: Predicting Optimization Outcome for Robustness Gains without Extra Pay

2025-01-19

AI Summary

In high-cost settings, selecting training data for foundation models is challenging, and task adaptation struggles to balance robustness and efficiency. Method: This paper proposes the Model-Predictive Task Sampling (MPTS) framework, which introduces a novel risk-prediction mechanism integrating generative task representations with Bayesian posterior inference—requiring no additional annotations, evaluations, or computational overhead—to actively identify high-risk tasks and optimize sampling strategies. MPTS unifies zero-shot, few-shot, and many-shot adaptation via risk-aware optimization and meta-adaptation synergy. Contribution/Results: Extensive evaluation across multiple benchmarks demonstrates that MPTS significantly enhances cross-task robustness while preserving original learning efficiency, exhibiting strong generalizability and practical utility.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Search and Optimization: Sampling/Simulation-based Search
• Natural Language Processing: Safety and Robustness

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

The foundation model enables fast problem-solving without learning from scratch, and such a desirable adaptation property benefits from its adopted cross-task generalization paradigms, e.g., pretraining, meta-training, or finetuning. Recent trends have focused on the curation of task datasets during optimization, which includes task selection as an indispensable consideration for either adaptation robustness or sampling efficiency purposes. Despite some progress, selecting crucial task batches to optimize over iteration mostly exhausts massive task queries and requires intensive evaluation and computations to secure robust adaptation. This work underscores the criticality of both robustness and learning efficiency, especially in scenarios where tasks are risky to collect or costly to evaluate. To this end, we present Model Predictive Task Sampling (MPTS), a novel active task sampling framework to establish connections between the task space and adaptation risk landscape achieve robust adaptation. Technically, MPTS characterizes the task episodic information with a generative model and predicts optimization outcome after adaptation from posterior inference, i.e., forecasting task-specific adaptation risk values. The resulting risk learner amortizes expensive annotation, evaluation, or computation operations in task robust adaptation learning paradigms. Extensive experimental results show that MPTS can be seamlessly integrated into zero-shot, few-shot, and many-shot learning paradigms, increases adaptation robustness, and retains learning efficiency without affording extra cost. The code will be available at the project site https://github.com/thu-rllab/MPTS.

Problem

• Efficient Data Selection
• Training Effectiveness
• High Data Collection Cost

Innovation

• MPTS
• Optimized Training Sample Selection
• Cost-Efficient Learning