From Kepler to Newton: Inductive Biases Guide Learned World Models in Transformers

• URL: http://arxiv.org/abs/2602.06923v1
• Date: Fri, 06 Feb 2026 18:17:37 GMT
• Title: From Kepler to Newton: Inductive Biases Guide Learned World Models in Transformers
• Abstract summary: General-purpose AI architectures can go beyond prediction to discover the physical laws governing the universe.<n>We show that ensuring spatial smoothness enables generic Transformers to surpass prior failures and learn a coherent Keplerian world model.<n>We force the model to abandon curve-fitting and discover Newtonian force representations.
• Score: 31.952312085850668
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Can general-purpose AI architectures go beyond prediction to discover the physical laws governing the universe? True intelligence relies on "world models" -- causal abstractions that allow an agent to not only predict future states but understand the underlying governing dynamics. While previous "AI Physicist" approaches have successfully recovered such laws, they typically rely on strong, domain-specific priors that effectively "bake in" the physics. Conversely, Vafa et al. recently showed that generic Transformers fail to acquire these world models, achieving high predictive accuracy without capturing the underlying physical laws. We bridge this gap by systematically introducing three minimal inductive biases. We show that ensuring spatial smoothness (by formulating prediction as continuous regression) and stability (by training with noisy contexts to mitigate error accumulation) enables generic Transformers to surpass prior failures and learn a coherent Keplerian world model, successfully fitting ellipses to planetary trajectories. However, true physical insight requires a third bias: temporal locality. By restricting the attention window to the immediate past -- imposing the simple assumption that future states depend only on the local state rather than a complex history -- we force the model to abandon curve-fitting and discover Newtonian force representations. Our results demonstrate that simple architectural choices determine whether an AI becomes a curve-fitter or a physicist, marking a critical step toward automated scientific discovery.

Related papers

• Can AI Follow In Einstein's Footsteps? [4.43316916502814]
  We review the current landscape of AI for physics discovery.<n>We highlight a critical missing skill: the ability to pose the right questions or invent the right principles.
  arXiv  Detail & Related papers  (2026-07-30T07:27:07Z)
• PhysiFormer: Learning to Simulate Mechanics in World Space [56.16644631495]
  We present PhysiFormer, a diffusion transformer for physically-plausible 3D object motion.<n>It represents objects as 3D meshes expressed in world coordinates.<n>It substantially outperforms autoregressive baselines in trajectory accuracy, rigidity preservation, and momentum-based physical consistency.
  arXiv  Detail & Related papers  (2026-06-25T17:59:00Z)
• Sensorimotor World Models: Perception for Action via Inverse Dynamics [59.121736411156384]
  We introduce a sensorimotor world model trained end-to-end with inverse dynamics regularization.<n>It prevents representation collapse and induces action-aligned representations.<n>It learns compact, interpretable latent spaces and enables competitive planning performance across simple 2D and 3D control tasks.
  arXiv  Detail & Related papers  (2026-06-18T11:25:16Z)
• Agentic World Modeling: Foundations, Capabilities, Laws, and Beyond [209.35045331678043]
  We introduce a "levels x laws" taxonomy organized along two axes.<n>The first defines three capability levels: L1 Predictor, which learns one-step local transition operators; L2 Simulator, which composes them into multi-step, action-conditioned rollouts that respect domain laws; and L3 Evolver, which autonomously revises its own model when predictions fail against new evidence.<n>We synthesize over 400 works and summarize more than 100 representative systems spanning model-based reinforcement learning, video generation, web and GUI agents, multi-agent social simulation, and AI-driven scientific discovery.
  arXiv  Detail & Related papers  (2026-04-24T17:48:47Z)
• Flow Learners for PDEs: Toward a Physics-to-Physics Paradigm for Scientific Computing [21.18315999355946]
  Partial differential equations govern nearly every physical process in science and engineering, yet solving them at scale remains prohibitively expensive.<n>We argue that the core issue is the abstraction used to train learned solvers. Many models are asked to predict states, while many scientific settings require modeling how uncertainty moves through constrained dynamics.<n>This motivates emphflow learners: models that parameterize transport vector fields and generate trajectories through integration, echoing the continuous dynamics that define PDE evolution.
  arXiv  Detail & Related papers  (2026-04-02T02:56:29Z)
• Walk through Paintings: Egocentric World Models from Internet Priors [65.30611174953958]
  We present the Egocentric World Model (EgoWM), a simple, architecture-agnostic method that transforms any pretrained video diffusion model into an action-conditioned world model.<n>Rather than training from scratch, we repurpose the rich world priors of Internet-scale video models and inject motor commands through lightweight conditioning layers.<n>Our approach scales naturally across embodiments and action spaces, ranging from 3-DoF mobile robots to 25-DoF humanoids.
  arXiv  Detail & Related papers  (2026-01-21T18:59:32Z)
• What Happens Next? Anticipating Future Motion by Generating Point Trajectories [76.16266402727643]
  We consider the problem of forecasting motion from a single image, predicting how objects in the world are likely to move.<n>We formulate this task as conditional generation of dense trajectory grids with a model that closely follows the architecture of modern video generators.<n>This approach captures scene-wide dynamics and uncertainty, yielding more accurate and diverse predictions than prior regressors and generators.
  arXiv  Detail & Related papers  (2025-09-25T21:03:56Z)
• GWM: Towards Scalable Gaussian World Models for Robotic Manipulation [53.51622803589185]
  We propose a novel branch of world model named Gaussian World Model (GWM) for robotic manipulation.<n>At its core is a latent Diffusion Transformer (DiT) combined with a 3D variational autoencoder, enabling fine-grained scene-level future state reconstruction.<n>Both simulated and real-world experiments depict that GWM can precisely predict future scenes conditioned on diverse robot actions.
  arXiv  Detail & Related papers  (2025-08-25T02:01:09Z)
• Mimicking the Physicist's Eye:A VLM-centric Approach for Physics Formula Discovery [98.58830663687911]
  VIPERR-aq1 is a multimodal model that performs Visual Induction for Equation Reasoning.<n>It integrates visual perception, trajectory data, and symbolic reasoning to emulate the scientific discovery process.<n>It consistently outperforms state-of-the-art VLM baselines in accuracy and interpretability.
  arXiv  Detail & Related papers  (2025-08-24T14:34:21Z)
• Universal Physics Simulation: A Foundational Diffusion Approach [0.0]
  We present the first foundational AI model for universal physics simulation that learns physical laws directly from boundary-condition data.<n>Our sketch-guided diffusion transformer approach reimagines computational physics by treating simulation as a conditional generation problem.<n>Unlike sequential time-stepping methods that accumulate errors over iterations, our approach bypasses temporal integration entirely.
  arXiv  Detail & Related papers  (2025-07-13T18:12:34Z)
• What Has a Foundation Model Found? Using Inductive Bias to Probe for World Models [3.394160022376002]
  We develop a technique for evaluating foundation models that examines how they adapt to synthetic datasets.<n>We find that foundation models can excel at their training tasks yet fail to develop inductive biases towards the underlying world model when adapted to new tasks.
  arXiv  Detail & Related papers  (2025-07-09T15:36:15Z)
• Advanced long-term earth system forecasting by learning the small-scale nature [74.19833913539053]
  We present Triton, an AI framework designed to address this fundamental challenge.<n>Inspired by increasing grids to explicitly resolve small scales in numerical models, Triton employs a hierarchical architecture processing information across multiple resolutions to mitigate spectral bias.<n>We demonstrate Triton's superior performance on challenging forecast tasks, achieving stable year-long global temperature forecasts, skillful Kuroshio eddy predictions till 120 days, and high-fidelity turbulence simulations.
  arXiv  Detail & Related papers  (2025-05-26T02:49:00Z)
• SPOTR: Spatio-temporal Pose Transformers for Human Motion Prediction [12.248428883804763]
  3D human motion prediction is a research area computation of high significance and a challenge in computer vision. Traditionally, autogregressive models have been used to predict human motion. We present a non-autoregressive model for human motion prediction.
  arXiv  Detail & Related papers  (2023-03-11T01:44:29Z)
• Predictive World Models from Real-World Partial Observations [66.80340484148931]
  We present a framework for learning a probabilistic predictive world model for real-world road environments. While prior methods require complete states as ground truth for learning, we present a novel sequential training method to allow HVAEs to learn to predict complete states from partially observed states only.
  arXiv  Detail & Related papers  (2023-01-12T02:07:26Z)
• A Spatio-temporal Transformer for 3D Human Motion Prediction [39.31212055504893]
  We propose a Transformer-based architecture for the task of generative modelling of 3D human motion. We empirically show that this effectively learns the underlying motion dynamics and reduces error accumulation over time observed in auto-gressive models.
  arXiv  Detail & Related papers  (2020-04-18T19:49:28Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.