[2507.09264] Overtone: Cyclic Patch Modulation for Clean, Efficient, and Flexible Physics Emulators

[2507.09264] Overtone: Cyclic Patch Modulation for Clean, Efficient, and Flexible Physics Emulators

arXiv - Machine Learning 4 min read

About this article

Abstract page for arXiv paper 2507.09264: Overtone: Cyclic Patch Modulation for Clean, Efficient, and Flexible Physics Emulators

Computer Science > Machine Learning arXiv:2507.09264 (cs) [Submitted on 12 Jul 2025 (v1), last revised 5 Mar 2026 (this version, v2)] Title:Overtone: Cyclic Patch Modulation for Clean, Efficient, and Flexible Physics Emulators Authors:Payel Mukhopadhyay, Michael McCabe, Ruben Ohana, Miles Cranmer View a PDF of the paper titled Overtone: Cyclic Patch Modulation for Clean, Efficient, and Flexible Physics Emulators, by Payel Mukhopadhyay and 3 other authors View PDF HTML (experimental) Abstract:Transformer-based PDE surrogates achieve remarkable performance but face two key challenges: fixed patch sizes cause systematic error accumulation at harmonic frequencies, and computational costs remain inflexible regardless of problem complexity or available resources. We introduce Overtone, a unified solution through dynamic patch size control at inference. Overtone's key insight is that cyclically modulating patch sizes during autoregressive rollouts distributes errors across the frequency spectrum, mitigating the systematic harmonic artifact accumulation that plague fixed-patch models. We implement this through two architecture-agnostic modules--CSM (using dynamic stride modulation) and CKM (using dynamic kernel resizing)--that together provide both harmonic mitigation and compute-adaptive deployment. This flexible tokenization lets users trade accuracy for speed dynamically based on computational constraints, and the cyclic rollout strategy yields up to 40% lower long rollout erro...

Originally published on March 06, 2026. Curated by AI News.

Related Articles

Llms

World models will be the next big thing, bye-bye LLMs

Was at Nvidia's GTC conference recently and honestly, it was one of the most eye-opening events I've attended in a while. There was a lot...

Reddit - Artificial Intelligence · 1 min ·
Machine Learning

[D] Got my first offer after months of searching — below posted range, contract-to-hire, and worried it may pause my search. Do I take it?

I could really use some outside perspective. I’m a senior ML/CV engineer in Canada with about 5–6 years across research and industry. Mas...

Reddit - Machine Learning · 1 min ·
Machine Learning

[Research] AI training is bad, so I started an research

Hello, I started researching about AI training Q:Why? R: Because AI training is bad right now. Q: What do you mean its bad? R: Like when ...

Reddit - Machine Learning · 1 min ·
Machine Learning

[P] Unix philosophy for ML pipelines: modular, swappable stages with typed contracts

We built an open-source prototype that applies Unix philosophy to retrieval pipelines. Each stage (PII redaction, chunking, dedup, embedd...

Reddit - Machine Learning · 1 min ·
More in Machine Learning: This Week Guide Trending

No comments

No comments yet. Be the first to comment!

Stay updated with AI News

Get the latest news, tools, and insights delivered to your inbox.

Daily or weekly digest • Unsubscribe anytime