Episodic Multi-Task Learning with Heterogeneous Neural Processes

TL;DR

This paper introduces Heterogeneous Neural Processes (HNPs), a novel approach for episodic multi-task learning that leverages heterogeneous task information and meta-knowledge to improve performance with limited data.

Contribution

HNPs integrate hierarchical Bayes and transformer modules to effectively utilize prior experiences and task relatedness in episodic multi-task learning, addressing data-insufficiency.

Findings

HNPs outperform baseline models in experiments.

Transformer inference modules enhance efficiency.

Ablation studies confirm module effectiveness.

Abstract

This paper focuses on the data-insufficiency problem in multi-task learning within an episodic training setup. Specifically, we explore the potential of heterogeneous information across tasks and meta-knowledge among episodes to effectively tackle each task with limited data. Existing meta-learning methods often fail to take advantage of crucial heterogeneous information in a single episode, while multi-task learning models neglect reusing experience from earlier episodes. To address the problem of insufficient data, we develop Heterogeneous Neural Processes (HNPs) for the episodic multi-task setup. Within the framework of hierarchical Bayes, HNPs effectively capitalize on prior experiences as meta-knowledge and capture task-relatedness among heterogeneous tasks, mitigating data-insufficiency. Meanwhile, transformer-structured inference modules are designed to enable efficient inferences toward meta-knowledge and task-relatedness. In this way, HNPs can learn more powerful functional priors for adapting to novel heterogeneous tasks in each meta-test episode. Experimental results show the superior performance of the proposed HNPs over typical baselines, and ablation studies verify the effectiveness of the designed inference modules.