Beyond Interpretability: Exploring the Comprehensibility of Adaptive Video Streaming through Large Language Models

TL;DR

This paper introduces ComTree, a framework that uses large language models to evaluate and select adaptive video streaming algorithms based on their comprehensibility for developers, balancing performance with human understanding.

Contribution

The paper presents the first framework integrating large language models to assess and optimize the comprehensibility of adaptive streaming algorithms.

Findings

ComTree improves algorithm comprehensibility significantly.

It maintains competitive performance while enhancing understandability.

The approach is validated through experimental results.

Abstract

Over the past decade, adaptive video streaming technology has witnessed significant advancements, particularly driven by the rapid evolution of deep learning techniques. However, the black-box nature of deep learning algorithms presents challenges for developers in understanding decision-making processes and optimizing for specific application scenarios. Although existing research has enhanced algorithm interpretability through decision tree conversion, interpretability does not directly equate to developers' subjective comprehensibility. To address this challenge, we introduce \texttt{ComTree}, the first bitrate adaptation algorithm generation framework that considers comprehensibility. The framework initially generates the complete set of decision trees that meet performance requirements, then leverages large language models to evaluate these trees for developer comprehensibility, ultimately selecting solutions that best facilitate human understanding and enhancement. Experimental results demonstrate that \texttt{ComTree} significantly improves comprehensibility while maintaining competitive performance, showing potential for further advancement. The source code is available at https://github.com/thu-media/ComTree.