A Unified Framework for Heterogeneous Semi-supervised Learning

2025-03-01

AI Summary

This paper addresses heterogeneous semi-supervised learning (HSSL), a novel setting where labeled and unlabeled domains differ in both label and feature distributions, and test samples originate from multiple source domains. To overcome limitations of conventional semi-supervised learning (SSL) and unsupervised domain adaptation (UDA), we propose the first unified framework jointly modeling SSL and UDA, explicitly capturing inter-domain heterogeneity and intra-class semantic relationships. Our adaptive fine-grained classification method integrates consistency regularization, domain-invariant feature alignment, semantic class-relation distillation, and dynamic weight assignment, enabling end-to-end joint optimization. Extensive experiments on multiple heterogeneous semi-supervised benchmarks demonstrate significant improvements over state-of-the-art methods, with average accuracy gains of 3.2–7.8 percentage points, validating the effectiveness and robustness of cross-domain knowledge transfer.

Technology Category

• Machine Learning: Semi-Supervised Learning
• Application Domains: Humanities & Computational Social Science
• Search and Optimization: Learning to Search

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

In this work, we introduce a novel problem setup termed as Heterogeneous Semi-Supervised Learning (HSSL), which presents unique challenges by bridging the semi-supervised learning (SSL) task and the unsupervised domain adaptation (UDA) task, and expanding standard semi-supervised learning to cope with heterogeneous training data. At its core, HSSL aims to learn a prediction model using a combination of labeled and unlabeled training data drawn separately from heterogeneous domains that share a common set of semantic categories; this model is intended to differentiate the semantic categories of test instances sampled from both the labeled and unlabeled domains. In particular, the labeled and unlabeled domains have dissimilar label distributions and class feature distributions. This heterogeneity, coupled with the assorted sources of the test data, introduces significant challenges to standard SSL and UDA methods. Therefore, we propose a novel method, Unified Framework for Heterogeneous Semi-supervised Learning (Uni-HSSL), to address HSSL by directly learning a fine-grained classifier from the heterogeneous data, which adaptively handles the inter-domain heterogeneity while leveraging both the unlabeled data and the inter-domain semantic class relationships for cross-domain knowledge transfer and adaptation. We conduct comprehensive experiments and the experimental results validate the efficacy and superior performance of the proposed Uni-HSSL over state-of-the-art semi-supervised learning and unsupervised domain adaptation methods.

Problem

• Bridges SSL and UDA tasks with heterogeneous data.
• Handles dissimilar label and feature distributions across domains.
• Proposes Uni-HSSL for cross-domain knowledge transfer and adaptation.

Innovation

• Unified Framework for Heterogeneous Semi-supervised Learning
• Adaptive inter-domain heterogeneity handling
• Cross-domain knowledge transfer and adaptation