TSEML: A task-specific embedding-based method for few-shot classification of cancer molecular subtypes

2024-12-03

IEEE International Conference on Bioinformatics and Biomedicine

AI Summary

To address the few-shot classification challenge arising from scarce molecular subtype annotations in cancer, this paper proposes Task-Specific Embedded Meta-Learning (TSEML), the first framework integrating Model-Agnostic Meta-Learning (MAML) and Prototypical Networks (ProtoNet) with a novel task-specific embedding mechanism. TSEML jointly models molecular subtypes (primary task) and cancer types (auxiliary task) to enable cross-task knowledge transfer. We further introduce multi-task embedding alignment, heterogeneous data augmentation, and feature disentanglement. Additionally, we construct TCGA Few-Shot—the first standardized few-shot benchmark for cancer molecular subtyping. Extensive experiments demonstrate that TSEML achieves an average 8.2–14.7% improvement in subtype classification accuracy over state-of-the-art few-shot methods on TCGA Few-Shot, empirically validating the efficacy of cross-task knowledge transfer under sparse annotation regimes.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Computer Vision: Multi-modal Vision
• Multiagent Systems: Multiagent Learning

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Molecular subtyping of cancer is recognized as a critical and challenging upstream task for personalized therapy. Existing deep learning methods have achieved significant performance in this domain when abundant data samples are available. However, the acquisition of densely labeled samples for cancer molecular subtypes remains a significant challenge for conventional data-intensive deep learning approaches. In this work, we focus on the few-shot molecular subtype prediction problem in heterogeneous and small cancer datasets, aiming to enhance precise diagnosis and personalized treatment. We first construct a new few-shot dataset for cancer molecular subtype classification and auxiliary cancer classification, named TCGA Few-Shot, from existing publicly available datasets. To effectively leverage the relevant knowledge from both tasks, we introduce a task-specific embedding-based meta-learning framework (TSEML). TSEML leverages the synergistic strengths of a model-agnostic meta-learning (MAML) approach and a prototypical network (ProtoNet) to capture diverse and fine-grained features. Comparative experiments conducted on the TCGA FewShot dataset demonstrate that our TSEML framework achieves superior performance in addressing the problem of few-shot molecular subtype classification.

Problem

• Cancer molecular subtype prediction
• Limited sample size
• Complex data

Innovation

• TSEML
• Few-Shot Learning
• Cancer Subtype Classification