analiza aiqbench
This arXiv paper probes frozen audio and vision encoders for how well their embeddings rank impacts by damping (quality factor Q), and reports that audio embeddings (Spearman ρ=0.73) far outperform a single CLIP video frame (ρ=0.28). It also shows that distilling an audio teacher into CLIP narrows that gap under a leave-one-material-out protocol, and that a four-frame input recovers about half the improvement. For ML readers, the useful takeaway is that much of the missing physical signal in single frames is temporal, though the results are limited to a specific benchmark and workshop-stage evidence.
How Much Physics Can Vision Learn from Audio? Probing and Distilling Damping Structure
arXiv:2610.22221v1 Announce Type: cross Abstract: Physical understanding is crucial for decision making, yet vision models struggle to infer properties such as damping that govern how materials behave. Audio encodes such properties directly, because an object's acoustic response to impact reveals its damping through measurable decay. We probe frozen encoders for the quality factor $Q$, a dimensionless measure of how many oscillations an object completes before its ring dies away, and report how well each embedding orders impacts by $Q$ using the Spearman rank correlation $\rho$, for which $\rho=1$ is a perfect ordering and $\rho=0$ an unrelated one. Audio embeddings reach $\rho=0.73$ and retain most of that advantage on materials withheld from training, while a single video frame taken at the moment of impact, encoded by CLIP, supports only $\rho=0.28$. Having established where the signal lives, we ask how much of that gap vision can recover. We distill a frozen audio teacher into a frozen CLIP backbone through a trained projection head and find that this recovers a useful fraction of the audio advantage. Because pooled correlations on this task are dominated in practice by category identity, we also evaluate under a leave-one-material-out protocol in which the held-out material is withheld from the projection head as well as the probe. Distillation with a four-frame input improves the correlation there from 0.234 to 0.338 (+0.053 over the matched four-frame baseline, 14 of 17 materials), a gain that no category-level strategy can produce. Widening the visual input to four frames spanning the impact supplies half of that improvement at no acoustic cost, which indicates that part of what a single frame misses is visible dynamics.
Abstract:Physical understanding is crucial for decision making, yet vision models struggle to infer properties such as damping that govern how materials behave. Audio encodes such properties directly, because an object's acoustic response to impact reveals its damping through measurable decay. We probe frozen encoders for the quality factor $Q$, a dimensionless measure of how many oscillations an object completes before its ring dies away, and report how well each embedding orders impacts by $Q$ using the Spearman rank correlation $\rho$, for which $\rho=1$ is a perfect ordering and $\rho=0$ an unrelated one. Audio embeddings reach $\rho=0.73$ and retain most of that advantage on materials withheld from training, while a single video frame taken at the moment of impact, encoded by CLIP, supports only $\rho=0.28$. Having established where the signal lives, we ask how much of that gap vision can recover. We distill a frozen audio teacher into a frozen CLIP backbone through a trained projection head and find that this recovers a useful fraction of the audio advantage. Because pooled correlations on this task are dominated in practice by category identity, we also evaluate under a leave-one-material-out protocol in which the held-out material is withheld from the projection head as well as the probe. Distillation with a four-frame input improves the correlation there from 0.234 to 0.338 (+0.053 over the matched four-frame baseline, 14 of 17 materials), a gain that no category-level strategy can produce. Widening the visual input to four frames spanning the impact supplies half of that improvement at no acoustic cost, which indicates that part of what a single frame misses is visible dynamics.
11 pages, 2 figures, 7 tables. Accepted at two NeurIPS 2026 workshops: Representations for the Physical Sciences (RPS) and Physical Understanding for Decision-Making (PhysUnderstand)
Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Sound (cs.SD)
arXiv:2610.22221 [cs.CV]
(or arXiv:2610.22221v1 [cs.CV] for this version)
https://doi.org/10.48550/arXiv.2610.22221 arXiv-issued DOI via DataCite (pending registration)
Text integral sub licența deschisă a publicației · 317 cuvinte
Subiecte conexe
Runway’s WorldPrompt and the Engineering of Real-Time Worlds
Latent Space · 25 sept. 2026
Dyna-2: General Purpose Physical Agents with Hour-Long Autonomy — Jason Ma, Dyna Robotics
Latent Space · 13 oct. 2026
[AINews] Opus 5.5 is good at explainer videos
Latent Space · 29 sept. 2026
[AINews] The Future of Latent Space
Latent Space · 25 sept. 2026