HiPhy: Hierarchical Alignment for Physically-Plausible Multi-Principle Video Generation

2026-10-01

AI Summary

This study addresses the limitation of existing video generation models in handling complex interactive scenarios involving multiple coexisting physical principles. To this end, we propose HiPhy, a novel framework that introduces a hierarchical physics-alignment mechanism. Specifically, it employs hierarchical reinforcement learning to jointly optimize local temporal dynamics and global semantic consistency, while incorporating physical commonsense constraints to guide the generation process. Furthermore, we construct a prompt dataset comprising 50,000 samples and introduce MultiPhyBench as a dedicated evaluation benchmark. Experimental results demonstrate that the proposed method significantly enhances both the physical plausibility and semantic alignment of generated videos in multi-physics concurrent scenarios, substantially outperforming existing baseline models.

Abstract

Video generation models have achieved remarkable visual fidelity and have strong potential to become general-purpose world simulators. Despite this progress, they still fail to generate videos which adhere to laws of physics. The problem becomes even more apparent in realistic settings where multiple physical principles must work together within the same video; for example, "a balloon floating upward while steam rises from a pot" requires buoyancy and fluid dynamics to unfold coherently and simultaneously. Yet existing methods largely ignore multi-principle interactions, focusing on a single principle per video. We propose HiPhy (Hierarchical Physical Alignment), a reinforcement learning framework that grounds video generation in physical laws through a dual-level objective: locally enforcing the temporal dynamics of individual physical principles, and globally ensuring the physical and semantic coherence of the entire scene. To support multi-principle generation, we construct a 50K-prompt dataset and introduce a prompt benchmark MultiPhyBench, spanning a diverse range of co-occurring physical events. Our experiments show that HiPhy significantly outperforms prior methods and baselines, improving physical commonsense and semantic alignment significantly across various benchmarks, with the largest gains on scenes involving multiple concurrent physical principles where competing methods degrade most sharply.

Problem

• video generation
• physical plausibility
• multi-principle interactions
• physical commonsense

Innovation

• Reinforcement Learning
• Hierarchical Alignment
• Video Generation
• Physical Plausibility
• Multi-Principle