Trustworthy Transfer Learning: A Survey

Authors: Jun Wu, Jingrui He

JAIR 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical In this paper, we understand transfer learning from the perspectives of knowledge transferability and trustworthiness. This involves two research questions: How is knowledge transferability quantitatively measured and enhanced across domains? Can we trust the transferred knowledge in the transfer learning process? To answer these questions, this paper provides a comprehensive review of trustworthy transfer learning from various aspects, including problem definitions, theoretical analysis, empirical algorithms, and real-world applications. Specifically, we summarize recent theories and algorithms for understanding knowledge transferability under (within-domain) IID and non-IID assumptions. In addition to knowledge transferability, we review the impact of trustworthiness on transfer learning, e.g., whether the transferred knowledge is adversarially robust or algorithmically fair, how to transfer the knowledge under privacy-preserving constraints, etc. Beyond discussing the current advancements, we highlight the open questions and future directions for understanding transfer learning in a reliable and trustworthy manner.
Researcher Affiliation Academia JUN WU , Michigan State University, USA JINGRUI HE, University of Illinois Urbana-Champaign, USA Authors Contact Information: Jun Wu, orcid: 0000-0002-1512-524X, EMAIL, Michigan State University, East Lansing, MI, USA; Jingrui He, orcid: 0000-0002-6429-6272, EMAIL, University of Illinois Urbana-Champaign, Urbana, IL, USA.
Pseudocode No The paper describes methods and concepts through prose, mathematical formulas, and figures. No explicit pseudocode or algorithm blocks were found.
Open Source Code No The paper discusses concepts and reviews existing work, including mentions of 'open-source model repositories such as Hugging Face (Wolf et al. 2020) and PyTorch Hub (Paszke et al. 2019)'. However, it does not provide concrete access to source code for the methodology or analysis presented in this survey paper itself.
Open Datasets No This is a survey paper and does not present new experimental results using specific datasets created or maintained by the authors. It refers to datasets (e.g., 'Image Net data set (J. Deng et al. 2009)') in the context of reviewed literature, but does not provide access information for data used in its own methodology or analysis.
Dataset Splits No This is a survey paper and does not describe experimental methodology that would involve dataset splits for its own work.
Hardware Specification No No specific hardware details for running the authors' own experiments were found, as this is a survey paper that reviews existing literature rather than conducting new experiments.
Software Dependencies No No specific ancillary software details with version numbers were found, as this is a survey paper and does not describe a new methodology requiring such dependencies.
Experiment Setup No No specific experimental setup details, hyperparameters, or training configurations were found, as this is a survey paper and does not present new experimental results.