Trustworthy Machine Learning under Distribution Shifts

2025-12-29

AI Summary

To address the limited trustworthiness of machine learning systems under distribution shifts, this paper systematically models three shift scenarios—perturbation-, domain-, and modality-induced—and introduces the first unified evaluation and enhancement framework covering robustness, interpretability, and adaptability. We propose a synergistic analytical paradigm integrating invariant risk minimization, causal representation learning, test-time adaptation, and interpretability regularization, augmented by robust optimization and multimodal alignment techniques—yielding theoretically grounded, computationally efficient adaptive trustworthy learning. Evaluated on multiple benchmarks, our approach improves cross-domain generalization accuracy by 5.2–12.7%, reduces prediction uncertainty by 38%, and enables verifiable decision provenance.

Technology Category

• Machine Learning: Adversarial Learning & Robustness
• Natural Language Processing: Safety and Robustness
• Computer Vision: Adversarial Attacks & Robustness

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Security and Privacy: Security and privacy of machine learning and AI applications
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Machine Learning (ML) has been a foundational topic in artificial intelligence (AI), providing both theoretical groundwork and practical tools for its exciting advancements. From ResNet for visual recognition to Transformer for vision-language alignment, the AI models have achieved superior capability to humans. Furthermore, the scaling law has enabled AI to initially develop general intelligence, as demonstrated by Large Language Models (LLMs). To this stage, AI has had an enormous influence on society and yet still keeps shaping the future for humanity. However, distribution shift remains a persistent ``Achilles' heel'', fundamentally limiting the reliability and general usefulness of ML systems. Moreover, generalization under distribution shift would also cause trust issues for AIs. Motivated by these challenges, my research focuses on extit{Trustworthy Machine Learning under Distribution Shifts}, with the goal of expanding AI's robustness, versatility, as well as its responsibility and reliability. We carefully study the three common distribution shifts into: (1) Perturbation Shift, (2) Domain Shift, and (3) Modality Shift. For all scenarios, we also rigorously investigate trustworthiness via three aspects: (1) Robustness, (2) Explainability, and (3) Adaptability. Based on these dimensions, we propose effective solutions and fundamental insights, meanwhile aiming to enhance the critical ML problems, such as efficiency, adaptability, and safety.

Problem

• Enhancing AI robustness against distribution shifts
• Improving reliability and trustworthiness of ML systems
• Addressing generalization issues across different data shifts

Innovation

• Addresses distribution shifts via perturbation, domain, and modality shifts
• Enhances trustworthiness through robustness, explainability, and adaptability
• Proposes solutions for efficiency, adaptability, and safety in ML