# Diagnosing Under-Development of Irreversible Processes in Video Generation

arXiv: https://arxiv.org/abs/2608.00617 · HTML: https://mlsanity.com/paper/2608.00617/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 31 of its 34 references that are in the map.
- First posted on arXiv 1 Aug 2026.

## Builds on (most cited first, 30 of 31)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [High-Resolution Image Synthesis with Latent Diffusion Models](https://mlsanity.com/paper/2112.10752/) (2021)
- [CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer](https://mlsanity.com/paper/2408.06072/) (2024)
- [Scaling Laws for Reward Model Overoptimization](https://mlsanity.com/paper/2210.10760/) (2022)
- [Concept Bottleneck Models](https://mlsanity.com/paper/2007.04612/) (2020)
- [Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representations](https://mlsanity.com/paper/1811.12359/) (2018)
- [Adversarial Diffusion Distillation](https://mlsanity.com/paper/2311.17042/) (2023)
- [Variational Autoencoders and Nonlinear ICA: A Unifying Framework](https://mlsanity.com/paper/1907.04809/) (2019)
- [VideoComposer: Compositional Video Synthesis with Motion Controllability](https://mlsanity.com/paper/2306.02018/) (2023)
- [Latent Video Diffusion Models for High-Fidelity Long Video Generation](https://mlsanity.com/paper/2211.13221/) (2022)
- [VideoPhy: Evaluating Physical Commonsense for Video Generation](https://mlsanity.com/paper/2406.03520/) (2024)
- [Towards World Simulator: Crafting Physical Commonsense-Based Benchmark for Video Generation](https://mlsanity.com/paper/2410.05363/) (2024)
- [Do generative video models understand physical principles?](https://mlsanity.com/paper/2501.09038/) (2025)
- [WorldModelBench: Judging Video Generation Models As World Models](https://mlsanity.com/paper/2502.20694/) (2025)
- [WISA: World Simulator Assistant for Physics-Aware Text-to-Video Generation](https://mlsanity.com/paper/2503.08153/) (2025)
- [Constrained Synthesis with Projected Diffusion Models](https://mlsanity.com/paper/2402.03559/) (2024)
- [ChronoMagic-Bench: A Benchmark for Metamorphic Evaluation of Text-to-Time-lapse Video Generation](https://mlsanity.com/paper/2406.18522/) (2024)
- [PISA Experiments: Exploring Physics Post-Training for Video Diffusion Models by Watching Stuff Drop](https://mlsanity.com/paper/2503.09595/) (2025)
- [Representation Learning via Global Temporal Alignment and Cycle-Consistency](https://mlsanity.com/paper/2105.05217/) (2021)
- [GenHowTo: Learning to Generate Actions and State Transformations from Instructional Videos](https://mlsanity.com/paper/2312.07322/) (2023)
- [Look for the Change: Learning Object States and State-Modifying Actions from Untrimmed Web Videos](https://mlsanity.com/paper/2203.11637/) (2022)
- [Learning Dynamical Systems from Partial Observations](https://mlsanity.com/paper/1902.11136/) (2019)
- [Constrained Monotonic Neural Networks](https://mlsanity.com/paper/2205.11775/) (2022)
- [Seeing the Arrow of Time in Large Multimodal Models](https://mlsanity.com/paper/2506.03340/) (2025)
- [VideoDreamer: Customized Multi-Subject Text-to-Video Generation with Disen-Mix Finetuning on Language-Video Foundation Models](https://mlsanity.com/paper/2311.00990/) (2023)
- [Physical Simulator In-the-Loop Video Generation](https://mlsanity.com/paper/2603.06408/) (2026)
- [Phantom: Physics-Infused Video Generation via Joint Modeling of Visual and Latent Physical Dynamics](https://mlsanity.com/paper/2604.08503/) (2026)
- [Invertible Monotone Operators for Normalizing Flows](https://mlsanity.com/paper/2210.08176/) (2022)
- [AUTM Flow: Atomic Unrestricted Time Machine for Monotonic Normalizing Flows](https://mlsanity.com/paper/2206.02102/) (2022)
- [OSCBench: Benchmarking Object State Change in Text-to-Video Generation](https://mlsanity.com/paper/2603.11698/) (2026)

## Abstract (arXiv)

Many physical attributes are irreversible: ice melts but does not re-freeze, paper chars but does not un-burn. Do video generators respect this? We show the question is hard to measure, and that what can be measured reliably is development rather than reversal. Metrics of local reversal are null-degenerate: a per-clip violation rate scores 0.50 on pure noise, and a variance-normalized reversal residual sits at its noise ceiling. What survives null-testing is a two-part protocol: progress (a directional attribute correlation) and a stasis rate. Under this protocol, generated video separates cleanly from real footage, and the gap is human-validated. Across seven text-to-video models, real reference footage advances (ρ=+0.40, 35% static) while every generator shows near-zero progress and 92--100% stasis; nine annotators rate real footage far above generated (2.75 vs. 0.99 on a 0--4 scale). The reliable finding is under-development: generators barely advance irreversible attributes rather than reversing them. As a complementary mechanism, we show that post-hoc readout guidance is gameable, whereas enforcing monotonicity by construction in a disentangled attribute latent removes the gameable readout, validated in controlled and semi-synthetic settings.