We're excited to announce our new website dedicated to the World as a Neural Network group, featuring a comprehensive Frequently Asked Questions section. Explore the interconnected world with us!

https://artificialneuralcomputing.com/wann

We are accepting applications for our 2024 internship program. Extensive knowledge of mathematics, physics and machine learning is required.

Phase I (January 2024 - February 2024): Taking an online class on advanced topics in machine learning (e.g. statistical modeling, thermodynamic description, non-equilibrium dynamics, physics- and bio-inspired algorithms, etc.)

Phase II (March 2024 - April 2024): Working on a research project that involves modeling (biological, physical or machine) learning system using numerical methods (e.g. developing numerical simulations, statistical analysis of learning systems, etc.)

Phase III (May 2024 - June 2024): Working on a research project that involves modeling (biological, physical or machine) learning system using analytical methods (e.g. developing theory of learning, designing machine learning algorithms, etc.) 

If you have no experience, but a lot of knowledge we strongly encourage you to apply. If you have a lot of experience, but no knowledge we strongly discouraged you to apply.
https://artificialneuralcomputing.com/internship

Early Warning System for pandemics? Why not! https://arxiv.org/abs/2401.04444

Why do complex systems tend to exhibit critical behavior? The dataset-learning duality might be the answer. https://arxiv.org/abs/2405.17391

🎉 NeuraPrint 2.0 🎉

Explore your mind like never before with NeuraPrint 2.0 — now available on iOS and Android!

Measure your Free Will and Consciousness with this unique tool:

🎲 Free Will: Determined by the accuracy of predictions of a 'random' sequence generated by the user.
🧠Consciousness: Calculated from the non-linearity of the predictions.

Ready to dive into self-discovery?

📲 Get it on [iOS] (https://apps.apple.com/us/app/neuraprint/id6446124144)

📲 Get it on [Android] (https://play.google.com/store/apps/details?id=org.anc.neuraprintapp.neuraprintapp)

A study of the emergence of classical field theories from the activation and learning dynamics of neural networks—whether artificial, biological, or fundamental. https://arxiv.org/abs/2411.08138

We are now hiring in Artificial Neural Computing.

Extensive knowledge of mathematics, physics and machine learning is required. Responsibilities include developing theory of learning and machine learning algorithms, modeling physical and biological systems, designing computer architecture and hardware components.

If you have no experience, but a lot of knowledge we strongly encourage you to apply. If you have a lot of experience, but no knowledge we strongly discouraged you to apply.

https://artificialneuralcomputing.com/internship

Starting in the new year, we will resume weekly Machine Learning Journal Club meetings. Here are some details:

WHY: The main goal is to discuss theoretical papers and ideas in machine learning. As far as we know, no such journal clubs currently exist, so let's try to keep the emphasis on theory, not numerics.

WHERE: The meetings will be held over Zoom and are open to all researchers and developers. Feel free to invite anyone who might be interested to join our dedicated group on Telegram: @ANCJournalClub.

WHEN: Unless stated otherwise, all meetings will take place on Tuesdays at 11 am ET. Please add this to your calendar, set a reminder, or use whatever method works for you—but please try not to be late.

WHAT: We will only discuss theoretical papers that are either already published or at least posted on arXiv. If you find something interesting, feel free to post it in this group along with your thoughts on why you think it’s relevant and/or interesting.

WHO: The meetings will start on January 7th, and Vitaly Vanchurin will present his paper, "Emergent Field Theories from Neural Networks." However, presenting papers by others (or inviting authors of relevant papers to present) is strongly encouraged. If you’re interested, please let us know, and we will add you to the schedule.

HOW: The meetings should be very informal, and everyone should feel free to actively participate in the discussions, ask relevant questions, make comments, etc.

Happy (upcoming) New Year!

A journey from physics to machine learning and back to physics. Along the way, we unified classical algorithms such as SGD, RMSProp, and Adam, developed new algorithms with superior learning efficiency, and identified a mechanism for the emergence of curved geometry.

“We apply the physics–learning duality to molecular systems ...” coming to arxiv near you this Tuesday https://youtu.be/7cfjS4mHbfQ

“If, in some cataclysm, all of scientific knowledge were to be destroyed, and only one sentence passed on to the next generations of creatures, what statement would contain the most information in the fewest words? I believe it is the atomic hypothesis: that all things are made of atoms.” — Richard Feynman

And here, we show that all atoms are agents, all agents are learning — all things are learning.

https://arxiv.org/abs/2504.10560

Perhaps it's time for physicists to explore machine learning algorithms, biologists to learn differential geometry, and machine learning scientists to study quantum mechanics.
https://arxiv.org/abs/2504.14728

🌟 We Are Now Accepting Applications for the Internship Program 🌟

Do you have extensive knowledge in mathematics, physics, and machine learning? Are you passionate about exploring learning systems at the deepest level—whether they are natural or artificial?

We are looking for motivated interns to join our research team working at the intersection of physics, biology, and machine learning. Our work is driven by the idea that the world itself may function as a neural network, and we aim to uncover the fundamental principles that govern these complex systems.

This internship offers a unique opportunity to contribute to cutting-edge theoretical and computational research, bridging ideas from multiple disciplines to deepen our understanding of the universe through neural physics.

🔗 Apply now: https://artificialneuralcomputing.com/internship
🔬 Discover our research: https://artificialneuralcomputing.com/research

Join us in pushing the boundaries of science, technology, and our understanding of reality.

Our growing product line now features the LADS library, Neurah platform, and NeuraStock — alongside NeuraPilot, NeuraTutor, and NeuraPrint. See https://artificialneuralcomputing.com/products

Join the Neurah Project

📢 We invite scholars from diverse disciplines to submit proposals exploring fundamental questions across the interconnected domains of human knowledge.

🧠 The Neurah project seeks to foster collaboration and catalyze breakthroughs in theory and application through contributions in physics, biology, chemistry, mathematics, computer science, neuroscience, psychology, philosophy, economics, political science, sociology, art, and education.

🤝 By bringing together diverse perspectives, we aim to advance new ideas and frameworks with the potential to transform research and education across fields.

🔗 Learn more and apply at http://Neurah.com

We are pleased to announce the launch of the NeuraStock project, a new research initiative by Artificial Neural Computing (ANC) dedicated to developing and testing theoretical models of financial market behavior. While ANC is not a financial services provider, this project represents a significant step toward building a scientific understanding of financial systems grounded in first principles.

The project began over a year ago with a foundational question: Can financial market behavior be modeled using tools from learning theory and neural physics? Through iterative cycles of theoretical development, numerical experimentation, and model refinement, we have constructed a predictive framework capable of generating consistent forecasts of market behavior. And no, it is not what is written in the books.

To test this framework, we developed the NeuraStock app (available on AppStore and Google Play), which currently implements the most basic version of our model — the local non-linear model. Despite its simplicity, this model predicts market direction correctly in approximately 52% of cases. While seemingly modest, this statistical edge translates into an annualized return of approximately 300%, as estimated through historical backtesting and early real-time trials. The model is available to the public free of charge for independent testing and verification.

Importantly, this is not a heuristic model; it is derived from theoretical principles that enable systematic generalization and extension. One such extension — the non-local non-linear model — is expected to yield annual returns exceeding 1000%. We are currently transitioning this model from simulation to real-time trading for empirical evaluation. If the model performs as expected, it will serve as a real-world confirmation of our theoretical framework. If not, the deviations will highlight important areas for further refinement and discovery. We plan to continue this empirical phase over the coming months.

Should the results align with theoretical predictions, we intend to publish the key findings — not only in the context of financial modeling but also within the broader effort to understand complex social systems. We believe the methods developed in this project may ultimately be applicable beyond finance, including in the study of social networks and political systems.

Stay tuned.