<!-- mobian-agent-page publisher="dailydev" canonical="https://daily.dev/posts/new-approach-enables-physically-consistent-ai-models-of-chaotic-systems-uwkl81ap4" -->

---
title: New approach enables physically consistent AI models of...
description: MIT researchers built a machine learning method that guarantees bounded, physically consistent predictions for chaotic systems like weather and turbulent...
canonical: https://daily.dev/posts/new-approach-enables-physically-consistent-ai-models-of-chaotic-systems-uwkl81ap4
twitter:card: summary_large_image
twitter:site: @dailydotdev
og:type: website
og:site_name: daily.dev
og:title: New approach enables physically consistent AI models of chaotic systems | daily.dev
og:description: MIT researchers built a machine learning method that guarantees bounded, physically consistent predictions for chaotic systems like weather and turbulent...
og:url: https://daily.dev/posts/new-approach-enables-physically-consistent-ai-models-of-chaotic-systems-uwkl81ap4
og:image: https://api.daily.dev/og/posts/UWkL81AP4.png
og:image:alt: New approach enables physically consistent AI models of chaotic systems
og:image:width: 1200
og:image:height: 630
og:locale: en
---

> ## Documentation Index
> Fetch the complete documentation index at: https://daily.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# New approach enables physically consistent AI models of chaotic systems

**[MIT News](https://daily.dev/sources/mit)** · 6 min read · 0 upvotes · 0 comments

## Summary

MIT researchers built a machine learning method that guarantees bounded, physically consistent predictions for chaotic systems like weather and turbulent flows. The approach, from MIT's LIDS, adds a projection layer that enforces energy dissipation as a hard constraint rather than a training-time penalty. Outside a learned boundary, the energy-like quantity must decrease, pulling the system back toward realistic behavior. Tested on systems from simple textbook cases to complex multidimensional ones, it matched true outcomes while conventional models drifted into impossible predictions. The work, published in PNAS by Sunbochen Tang, Navid Azizan and Themistoklis Sapsis, is model-agnostic and the code is open source. A follow-up paper by members of the team won the Learning for Dynamics and Control conference's Best Paper Award.

## Full article

daily.dev links to this article rather than hosting it. Read it at the original source: <https://news.mit.edu/2026/new-approach-enables-physically-consistent-ai-models-chaotic-systems-1106>

## Questions this post answers

### How can you prevent a machine learning model of a chaotic system like weather from producing physically impossible predictions over long forecast windows?

MIT researchers add a projection layer that enforces energy dissipation as a hard constraint rather than a training-time penalty. The model learns an energy-like function and a boundary; outside that boundary the energy-like quantity must decrease, pulling trajectories back toward the bounded region. This gives theoretical guarantees that predictions stay bounded over arbitrarily long forecast windows, unlike soft penalty-based approaches used previously.

_Anyone building reliable forecasting models can follow this kind of physics-constrained research on daily.dev._

### What is the difference between hard constraining and soft constraining in physics-informed neural networks?

Soft constraining imposes penalties or incentives during training to encourage a model to respect physical limits, but offers no guarantee the model obeys them on new inputs at test time, so it can still produce unbounded predictions. Hard constraining, used in MIT's hard-constrained neural network approach, adds a projection layer that enforces the constraint structurally, ensuring outputs stay physically consistent even on unseen cases.

_Developers weighing modeling approaches can track this kind of constraint-design research on daily.dev._

## Similar posts on daily.dev

- [New research could lead to faster, trustworthy AI models for climate prediction](https://daily.dev/posts/new-research-could-lead-to-faster-trustworthy-ai-models-for-climate-prediction-lquoutmiv) · MIT News · 0 upvotes · 0 comments
- [The Next Frontier of AI in Production Is Chaos Engineering](https://daily.dev/posts/the-next-frontier-of-ai-in-production-is-chaos-engineering-3q6sgux7c) · Towards Data Science · 0 upvotes · 0 comments
- [New method enables AI for safety-critical situations](https://daily.dev/posts/new-method-enables-ai-for-safety-critical-situations-sjxxovxig) · MIT News · 0 upvotes · 0 comments
- [Why Model Collapse In LLMs Is Inevitable With Self-Learning](https://daily.dev/posts/why-model-collapse-in-llms-is-inevitable-with-self-learning-xmyaxtadg) · Hackaday · 2 upvotes · 0 comments
- [How to Build AI Systems That Know When They Don't Know: A Practical Guide](https://daily.dev/posts/how-to-build-ai-systems-that-know-when-they-don-t-know-a-practical-guide-tej37tpcc) · freeCodeCamp · 0 upvotes · 0 comments

---

Tags: [#neural-networks](https://daily.dev/tags/neural-networks), [#mit](https://daily.dev/tags/mit)

[View this post on daily.dev](https://daily.dev/posts/new-approach-enables-physically-consistent-ai-models-of-chaotic-systems-uwkl81ap4)

```json
{"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]}
{"@context":"https://schema.org","@type":"TechArticle","headline":"New approach enables physically consistent AI models of chaotic systems","url":"https://daily.dev/posts/new-approach-enables-physically-consistent-ai-models-of-chaotic-systems-uwkl81ap4","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/new-approach-enables-physically-consistent-ai-models-of-chaotic-systems-uwkl81ap4"},"datePublished":"2026-11-20T21:20:43.821Z","dateModified":"2026-11-20T21:21:08.953Z","description":"MIT researchers built a machine learning method that guarantees bounded, physically consistent predictions for chaotic systems like weather and turbulent...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/14ee865504dc91dd26cfab1ab2c0474e?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/14ee865504dc91dd26cfab1ab2c0474e?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"MIT News","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"MIT News","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/881e20a40aa44b12ba4bd27f5c14289d","url":"https://daily.dev/sources/mit"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/new-approach-enables-physically-consistent-ai-models-of-chaotic-systems-uwkl81ap4","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"neural-networks,mit","timeRequired":"PT6M"}
{"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"MIT News","item":"https://daily.dev/sources/mit"},{"@type":"ListItem","position":3,"name":"New approach enables physically consistent AI models of chaotic systems"}]}
{"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/new-approach-enables-physically-consistent-ai-models-of-chaotic-systems-uwkl81ap4#faq","mainEntity":[{"@type":"Question","name":"How can you prevent a machine learning model of a chaotic system like weather from producing physically impossible predictions over long forecast windows?","acceptedAnswer":{"@type":"Answer","text":"MIT researchers add a projection layer that enforces energy dissipation as a hard constraint rather than a training-time penalty. The model learns an energy-like function and a boundary; outside that boundary the energy-like quantity must decrease, pulling trajectories back toward the bounded region. This gives theoretical guarantees that predictions stay bounded over arbitrarily long forecast windows, unlike soft penalty-based approaches used previously. Anyone building reliable forecasting models can follow this kind of physics-constrained research on daily.dev."}},{"@type":"Question","name":"What is the difference between hard constraining and soft constraining in physics-informed neural networks?","acceptedAnswer":{"@type":"Answer","text":"Soft constraining imposes penalties or incentives during training to encourage a model to respect physical limits, but offers no guarantee the model obeys them on new inputs at test time, so it can still produce unbounded predictions. Hard constraining, used in MIT's hard-constrained neural network approach, adds a projection layer that enforces the constraint structurally, ensuring outputs stay physically consistent even on unseen cases. Developers weighing modeling approaches can track this kind of constraint-design research on daily.dev."}}]}
```

