Smaller is Better: Generative Models Can Power Short Video Preloading

2026-02-10

AI Summary

This work addresses the inherent trade-off in short video preloading between bandwidth conservation and playback stuttering. The authors propose PromptPream, the first framework to integrate generative models into video preloading by transmitting semantic prompts instead of raw pixel data and reconstructing high-quality video frames locally on the client using on-device computation. This approach circumvents the traditional bandwidth–stuttering dilemma. PromptPream synergistically optimizes network and computational resources through gradient-based prompt inversion, computation-aware scheduling, and a scalable search algorithm. Experimental results demonstrate that, compared to conventional preloading strategies, PromptPream reduces both stuttering rate and bandwidth waste by over 31% while significantly improving Quality of Experience (QoE) by 45%.

Technology Category

• Natural Language Processing: Prompt Engineering / Prompting
• Computer Vision: Diffusion Models for Vision
• Machine Learning: Learning on the Edge & Model Compression

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Search and Retrieval-Augmented AI: Personalized, context-aware and across-device search

Abstract

Preloading is widely used in short video platforms to minimize playback stalls by downloading future content in advance. However, existing strategies face a tradeoff. Aggressive preloading reduces stalls but wastes bandwidth, while conservative strategies save data but increase the risk of playback stalls. This paper presents PromptPream, a computation powered preloading paradigm that breaks this tradeoff by using local computation to reduce bandwidth demand. Instead of transmitting pixel level video chunks, PromptPream sends compact semantic prompts that are decoded into high quality frames using generative models such as Stable Diffusion. We propose three core techniques to enable this paradigm: (1) a gradient based prompt inversion method that compresses frames into small sets of compact token embeddings; (2) a computation aware scheduling strategy that jointly optimizes network and compute resource usage; and (3) a scalable searching algorithm that addresses the enlarged scheduling space introduced by scheduler. Evaluations show that PromptStream reduces both stalls and bandwidth waste by over 31%, and improves Quality of Experience (QoE) by 45%, compared to traditional strategies.

Problem

• preloading
• bandwidth waste
• playback stalls
• short video
• quality of experience

Innovation

• generative models
• semantic prompts
• computation-aware scheduling
• prompt inversion
• video preloading