Feature-Based vs. GAN-Based Learning from Demonstrations: When and Why

TL;DR

This paper compares feature-based and GAN-based learning from demonstrations, analyzing their strengths, weaknesses, and suitable application scenarios, emphasizing the importance of structured motion representations and task-specific considerations.

Contribution

It provides a comprehensive framework for understanding the trade-offs between feature-based and GAN-based methods in learning from demonstrations, guiding method selection based on task needs.

Findings

Feature-based methods excel at high-fidelity motion imitation.

GAN-based methods offer scalability and flexibility.

Structured motion representations improve both approaches.

Abstract

This survey provides a comparative analysis of feature-based and GAN-based approaches to learning from demonstrations, with a focus on the structure of reward functions and their implications for policy learning. Feature-based methods offer dense, interpretable rewards that excel at high-fidelity motion imitation, yet often require sophisticated representations of references and struggle with generalization in unstructured settings. GAN-based methods, in contrast, use implicit, distributional supervision that enables scalability and adaptation flexibility, but are prone to training instability and coarse reward signals. Recent advancements in both paradigms converge on the importance of structured motion representations, which enable smoother transitions, controllable synthesis, and improved task integration. We argue that the dichotomy between feature-based and GAN-based methods is increasingly nuanced: rather than one paradigm dominating the other, the choice should be guided by task-specific priorities such as fidelity, diversity, interpretability, and adaptability. This work outlines the algorithmic trade-offs and design considerations that underlie method selection, offering a framework for principled decision-making in learning from demonstrations.