The paper identifies a fundamental limitation in current deep learning, describing an "ergodic ceiling" that forces models to converge to statistical averages of existing human knowledge. It proposes a discrete geometric physics sandbox—built on a Holographic E8 Projection Engine and a lattice of Planck‑scale spheres—to inject non‑ergodic, path‑dependent insight and enforce thermodynamic constraints. By modeling spacetime as nested FCC lattices and using topological constraints, the sandbox aims to prevent model collapse, recover quantum mechanics via renormalisation‑group flows, and potentially collapse NP‑to‑P complexity by restricting algorithmic search to physically conserved trajectories.
arXiv:2607. 07127v1 Announce Type: cross Abstract: Lattice field theory is the workhorse of non-perturbative physics, used to simulate phenomena from the strong nuclear force to critical phenomena in materials.
arXiv:2510. 25306v3 Announce Type: replace Abstract: Partial physical knowledge--governing structures known, constitutive relations or their combinations not--pervades spatiotemporal systems.
arXiv:2606. 30512v1 Announce Type: cross Abstract: Why overparameterised deep networks generalise so remarkably well remains one of the most stubborn open questions in machine learning theory.
arXiv:2608. 12624v1 Announce Type: new Abstract: Structure-preserving machine learning embeds physical structure directly into model architectures, yet uncertainty quantification (UQ) for such hard-constrained models remains limited because standard UQ methods may violate the encoded admissibility conditions, require architectural modifications, or impose substantial computational costs.
arXiv:2607. 05436v1 Announce Type: new Abstract: The rapid scaling of over-parameterized machine learning architectures, particularly LLMs, raises a profound crisis: do these systems exhibit genuine intelligence, or are they merely sophisticated statistical pattern matchers?
arXiv:2607. 05127v1 Announce Type: cross Abstract: Understanding the physics of many-body complex dynamical systems is typically non-trivial.
The paper introduces NS‑Flows, a flow‑based nested sampling method that replaces Markov‑chain updates with a conditional normalizing flow trained on live sets. By applying this technique to a Lennard‑Jones particle system, the authors achieve over two orders of magnitude fewer energy evaluations and a roughly one‑third reduction in wall‑clock time compared to traditional nested sampling. The study also shows that the flow’s generation efficiency varies non‑monotonically along the annealing trajectory, providing a diagnostic of the system’s internal mode complexity and identifying liquid‑like ensembles as the most challenging for current flow architectures.
arXiv:2607. 03329v1 Announce Type: new Abstract: Conventional uniform convergence bounds and empirical risk minimization break down in massive over-parameterized models, such as large language transformers and biological sequence networks.
arXiv:2610.28487v1 Announce Type: new Abstract: We introduce a two-stage data-driven neural framework for discovering dynamically admissible extreme scenarios. Rather than assuming prior knowledge of...
arXiv:2608. 08485v1 Announce Type: new Abstract: Current LLM safety guardrails face a fundamental tension: fine-tuning distorts pre-trained representations while generative judges incur prohibitive inference costs.
arXiv:2606. 01179v1 Announce Type: cross Abstract: Entropy production governs irreversibility and uncertainty in both physical and information-theoretic systems.
HoloAegis is a minimally parametric topological inference framework that uses frozen representations to map text onto the unit sphere and makes decisions via Gibbs‑Boltzmann free‑energy differences over pre‑computed anchor centroids. On a frozen three‑benchmark protocol, it matches WildGuard‑7B on toxicity, outperforms it on harmful behaviors, but underperforms on oversafety detection, while ShieldGemma‑2B fails on indirect harms. The study demonstrates that geometric guardrails can substitute for LLM judges in some cases and must defer to them in others, with anchor banks reducing score variance and boundary displacement.