# Emergence of Physical Intelligence via Controllable Information Production

arXiv: https://arxiv.org/abs/2601.22449 · HTML: https://mlsanity.com/paper/2601.22449/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Deep RL algorithms and representations, one of 15 directions in Deep reinforcement learning methods.
- Builds on 22 of its 47 references that are in the map.
- First posted on arXiv 30 Jan 2026.

## Placed in

- Deep reinforcement learning methods / [Deep RL algorithms and representations](https://mlsanity.com/deeprl/deep-rl-algorithms-and-representations/)

## Builds on (most cited first, 22 of 22)

- [Convex optimization](https://mlsanity.com/paper/2106.01946/) (2021)
- [Mastering Diverse Domains through World Models](https://mlsanity.com/paper/2301.04104/) (2023)
- [Exploration by Random Network Distillation](https://mlsanity.com/paper/1810.12894/) (2018)
- [Diversity is All You Need: Learning Skills without a Reward Function](https://mlsanity.com/paper/1802.06070/) (2018)
- [TD-MPC2: Scalable, Robust World Models for Continuous Control](https://mlsanity.com/paper/2310.16828/) (2023)
- [MimicGen: A Data Generation System for Scalable Robot Learning using Human Demonstrations](https://mlsanity.com/paper/2310.17596/) (2023)
- [Dynamics-Aware Unsupervised Discovery of Skills](https://mlsanity.com/paper/1907.01657/) (2019)
- [DiffTaichi: Differentiable Programming for Physical Simulation](https://mlsanity.com/paper/1910.00935/) (2019)
- [URLB: Unsupervised Reinforcement Learning Benchmark](https://mlsanity.com/paper/2110.15191/) (2021)
- [Training Agents Inside of Scalable World Models](https://mlsanity.com/paper/2509.24527/) (2025)
- [End-to-End Robotic Reinforcement Learning without Reward Engineering](https://mlsanity.com/paper/1904.07854/) (2019)
- [Sample-efficient Cross-Entropy Method for Real-time Planning](https://mlsanity.com/paper/2008.06389/) (2020)
- [GPU-Accelerated Robotic Simulation for Distributed Reinforcement Learning](https://mlsanity.com/paper/1810.05762/) (2018)
- [MaxInfoRL: Boosting exploration in reinforcement learning through information gain maximization](https://mlsanity.com/paper/2412.12098/) (2024)
- [Dojo: A Differentiable Physics Engine for Robotics](https://mlsanity.com/paper/2203.00806/) (2022)
- [Unsupervised Real-Time Control through Variational Empowerment](https://mlsanity.com/paper/1710.05101/) (2017)
- [Intrinsic Motivation in Dynamical Control Systems](https://mlsanity.com/paper/2301.00005/) (2022)
- [Automatic Environment Shaping is the Next Frontier in RL](https://mlsanity.com/paper/2407.16186/) (2024)
- [Whole-Body Model-Predictive Control of Legged Robots with MuJoCo](https://mlsanity.com/paper/2503.04613/) (2025)
- [Machine-learning optimized measurements of chaotic dynamical systems via the information bottleneck](https://mlsanity.com/paper/2311.04896/) (2023)
- [Decentralized Traffic Flow Optimization Through Intrinsic Motivation](https://mlsanity.com/paper/2505.11520/) (2025)
- [ARMimic: Learning Robotic Manipulation from Passive Human Demonstrations in Augmented Reality](https://mlsanity.com/paper/2509.22914/) (2025)

## Built on by (most cited first, 1 of 1)

- [Bellman Meets Lyapunov: Unsupervised Reinforcement Learning via Mastering Chaos](https://mlsanity.com/paper/2610.02012/) (2026)

## Abstract (arXiv)

Intrinsic Motivation (IM) aims to train agents without external rewards, enabling useful behavior to emerge from the agent's interaction with its environment alone. However, the dominant IM approaches rely on information-theoretic quantities with designer-chosen variables, introducing bias and lacking a principled connection to dynamics or optimal control (OC). We introduce Controllable Information Production (CIP), a novel IM method explicitly grounded in dynamical systems and OC. CIP measures the rate at which an agent produces information, capturing controllable complexity without external knowledge or bias. CIP unifies IM and OC into a single framework, formalizing physical intelligence as the control of information production. It further reveals connections between the structure of the value function and Kolmogorov-Sinai entropy. CIP consistently outperforms prior IM methods on standard benchmarks in robot learning and solves tasks they fail on, including humanoid self-righting. These results support an organizing principle: that physical intelligence emerges from driving systems toward the edge of controllable chaos.