Diagnosing Under-Development of Irreversible Processes in Video Generation
Many physical attributes are \emph{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 \emph{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 \emph{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.
Code (0)
등록된 구현이 없습니다.
Tasks
Video GenerationSimilar Papers 제목 키워드 기반
AdaSpark: Adaptive Sparsity for Efficient Long-Video Understanding
Processing long-form videos with Video Large Language Models (Video-LLMs) is computationally prohibitive. Current efficiency methods often compromise fine-grained perception through irreversible information disposal or i…
Smartphone-Based Test and Predictive Models for Rapid, Non-Invasive, and Point-of-Care Monitoring of Ocular and Cardiovascular Complications Related to Diabetes
Among the most impactful diabetic complications are diabetic retinopathy, the leading cause of blindness among working class adults, and cardiovascular disease, the leading cause of death worldwide. This study describes …
image-classificationImage ClassificationTransfer LearningRepresenting, reasoning and answering questions about biological pathways - various applications
Biological organisms are composed of numerous interconnected biochemical processes. Diseases occur when normal functionality of these processes is disrupted. Thus, understanding these biochemical processes and their inte…
Which Way Does Time Flow? A Psychophysics-Grounded Evaluation for Vision-Language Models
Modern vision-language models (VLMs) excel at many multimodal tasks, yet their grasp of temporal information in video remains weak and has not been adequately evaluated. We probe this gap with a deceptively simple but re…
Modeling the Effect of Blunt Impact on Mitochondrial Dysfunction in Cartilage
Mounting evidence for the role of oxidative stress in the degeneration of articular cartilage after an injurious impact requires our modeling & simulation efforts to temporarily shift from just describing the effect of m…