Ming-xing Luo

Liming Zhao, Naixu Guo, Ming-Xing Luo et al. · berkeley

The complete learning of an $n$-qubit quantum state requires samples exponentially in $n$. Several works consider subclasses of quantum states that can be learned in polynomial sample complexity such as stabilizer states or high-temperature Gibbs states. Other works consider a weaker sense of learning, such as PAC learning and shadow tomography. In this work, we consider learning states that are close to neural network quantum states, which can efficiently be represented by a graphical model called restricted Boltzmann machines (RBMs). To this end, we exhibit robustness results for efficient provable two-hop neighborhood learning algorithms for ferromagnetic and locally consistent RBMs. We consider the $L_p$-norm as a measure of closeness, including both total variation distance and max-norm distance in the limit. Our results allow certain quantum states to be learned with a sample complexity \textit{exponentially} better than naive tomography. We hence provide new classes of efficiently learnable quantum states and apply new strategies to learn them.

Ming-Xing Luo · berkeley

Shannon information was defined for characterizing the uncertainty information of classical probabilistic distributions. As an uncertainty measure it is generally believed to be positive. This holds for any information quantity from two random variables because of the polymatroidal axioms. However, it is unknown why there is negative information for more than two random variables on finite dimensional spaces. We first show the negative tripartite Shannon mutual information implies specific Bayesian network representations of its joint distribution. We then show that the negative Shannon information is obtained from general tripartite Bayesian networks with quantum realizations. This provides a device-independent witness of negative Shannon information. We finally extend the result for general networks. The present result shows new insights in the network compatibility from non-Shannon information inequalities.

Shi Qiu, Junyi Deng, Yiwei Deng et al.

AI agents powered by large language models exhibit strong reasoning and problem-solving capabilities, enabling them to assist scientific research tasks such as formula derivation and code generation. However, whether these agents can reliably perform end-to-end reproduction from real scientific papers remains an open question. We introduce PRBench, a benchmark of 30 expert-curated tasks spanning 11 subfields of physics. Each task requires an agent to comprehend the methodology of a published paper, implement the corresponding algorithms from scratch, and produce quantitative results matching the original publication. Agents are provided only with the task instruction and paper content, and operate in a sandboxed execution environment. All tasks are contributed by domain experts from over 20 research groups at the School of Physics, Peking University, each grounded in a real published paper and validated through end-to-end reproduction with verified ground-truth results and detailed scoring rubrics. Using an agentified assessment pipeline, we evaluate a set of coding agents on PRBench and analyze their capabilities across key dimensions of scientific reasoning and execution. The best-performing agent, OpenAI Codex powered by GPT-5.3-Codex, achieves a mean overall score of 34%. All agents exhibit a zero end-to-end callback success rate, with particularly poor performance in data accuracy and code correctness. We further identify systematic failure modes, including errors in formula implementation, inability to debug numerical simulations, and fabrication of output data. Overall, PRBench provides a rigorous benchmark for evaluating progress toward autonomous scientific research.

Ming-Xing Luo

We prove that the $L^2$ CVP distance from a random short ring element to the log-unit lattice of $\Q(ζ_{2^k})$ converges to $\fracπ{2\sqrt{6}}\sqrt{n}$ as $n=2^{k-1}\to\infty$. We then show that this target lies inside the Voronoi cell of the origin for $k\ge 4$. For the $L^\infty$ norm, the maximum over $n$ sub-Gaussian coordinates yields $O(\sqrt{\log n})$ which translates into a sub-polynomial approximation factor for the Short Generator Problem. We show a Coarse Lattice Theorem that Babai's algorithm returns zero for all structured targets, yet exactly recovers unit perturbations of arbitrary size. For module determinant ideals, we further prove the Trigamma Theorem that proves an intrinsic imbalance $σ_{g_0}=O(1)$ independent of the modulus $q$. Finally, combined with Parts I and II, we reduce the CDPR factor for ML-KEM from $\exp(\tO(\sqrt{n}))$ to a sub-polynomial value.

Ming-Xing Luo

We present a quantum attack on ML-KEM and related 2-power cyclotomic lattice schemes. Combining with Parts I-III, we provide an algorithm and verify the resulting approximation factor satisfies $γ\le 21 < q/2=1665$ for ML-KEM-1024, with a success probability $\ge 0.99$. We apply a tower decomposition of the Principal Ideal Problem (PIP) through the chain $\Q \subset \Q(ζ_8) \subset \cdots \subset \Q(ζ_{2^k})$ which yields a polynomial-time quantum algorithm costing $O(n^3 \log^2 n)$ gates, $O(n^2 \log n)$ qubits, and $\mathrm{poly}(n)$ classical bit operations. We extend the analysis to Falcon, Hawk, and NTRU over 2-power cyclotomic rings. This means that ML-KEM, Falcon, Hawk, NTRU-HPS, and NTRU-HRSS with all standardized parameter sets are broken under quantum attack.

Shi Qiu, Zeyu Cai, Jiashen Wei et al.

We present, to our knowledge, the first language-driven agent system capable of executing end-to-end collider phenomenology tasks, instantiated within a decoupled, domain-agnostic architecture for autonomous High-Energy Physics phenomenology. Guided only by natural-language prompts supplemented with standard physics notation, ColliderAgent carries out workflows from a theoretical Lagrangian to final phenomenological outputs without relying on package-specific code. In this framework, a hierarchical multi-agent reasoning layer is coupled to Magnus, a unified execution backend for phenomenological calculations and simulation toolchains. We validate the system on representative literature reproductions spanning leptoquark and axion-like-particle scenarios, higher-dimensional effective operators, parton-level and detector-level analyses, and large-scale parameter scans leading to exclusion limits. These results point to a route toward more automated, scalable, and reproducible research in collider physics, cosmology, and physics more broadly.

Zhuo-Yang Song, Qing-Hong Cao, Ming-xing Luo et al.

Large language model (LLM)-driven agents are emerging as a powerful new paradigm for solving complex problems. Despite the empirical success of these practices, a theoretical framework to understand and unify their macroscopic dynamics remains lacking. This Letter proposes a method based on the least action principle to estimate the underlying generative directionality of LLMs embedded within agents. By experimentally measuring the transition probabilities between LLM-generated states, we statistically discover a detailed balance in LLM-generated transitions, indicating that LLM generation may not be achieved by generally learning rule sets and strategies, but rather by implicitly learning a class of underlying potential functions that may transcend different LLM architectures and prompt templates. To our knowledge, this is the first discovery of a macroscopic physical law in LLM generative dynamics that does not depend on specific model details. This work is an attempt to establish a macroscopic dynamics theory of complex AI systems, aiming to elevate the study of AI agents from a collection of engineering practices to a science built on effective measurements that are predictable and quantifiable.

Shi Qiu, Shaoyang Guo, Zhuo-Yang Song et al.

Current benchmarks for evaluating the reasoning capabilities of Large Language Models (LLMs) face significant limitations: task oversimplification, data contamination, and flawed evaluation items. These deficiencies necessitate more rigorous assessment methods. To address these limitations, we introduce PHYBench, a benchmark of 500 original physics problems ranging from high school to Physics Olympiad difficulty. PHYBench addresses data contamination through original content and employs a systematic curation pipeline to eliminate flawed items. Evaluations show that PHYBench activates more tokens and provides stronger differentiation between reasoning models compared to other baselines like AIME 2024, OlympiadBench and GPQA. Even the best-performing model, Gemini 2.5 Pro, achieves only 36.9% accuracy compared to human experts' 61.9%. To further enhance evaluation precision, we introduce the Expression Edit Distance (EED) Score for mathematical expression assessment, which improves sample efficiency by 204% over binary scoring. Moreover, PHYBench effectively elicits multi-step and multi-condition reasoning, providing a platform for examining models' reasoning robustness, preferences, and deficiencies. The benchmark results and dataset are publicly available at https://www.phybench.cn/.

Ming-Xing Luo

We extend the CDPR lattice reduction algorithm from ideal to module lattices, leveraging the trace orthogonality of the power basis to decompose the module into rank-1 submodules and applying CDPR independently to each. This base module reduction achieves a Hermite factor $\exp(\tilde{O}(\sqrt{n}))$ matching the ideal case, with a module reduction factor $O(1)$ independent of the rank, under a balance hypothesis automatically satisfied for MLWE-distributed bases. To control precision, we introduce CRT-scaled rounding at totally split primes, reducing the Gram-Schmidt rounding error and yielding a bounded-precision implementation. We further reformulate the CDPR sign-selection subproblem as a mixed-integer linear program, determining the optimal balanced discrepancy to be a universal constant $δ^*\approx 0.4407$. All results build on the class number one condition $h_k^+=1$ established in Part I of this series.

Ming-Xing Luo

Weber's conjecture (1886) governs three aspects of lattice-based cryptography: the solvability of the Principal Ideal Problem, the freeness of modules over rings of integers, and the tightness of worst-case-to-average-case reductions in Ring-LWE (R-LWE) and Module-LWE (MLWE). Existing verifications for $k \ge 9$ rely on Generalized Riemann Hypothesis (GRH). In this paper, we present the first unconditional proof for $k \le 12$. Our method combines the Fukuda-Komatsu computational sieve, inductive structure of the cyclotomic $\mathbb{Z}_2$-tower, and Herbrand's theorem.

Zhuo-Yang Song, Zeyu Li, Qing-Hong Cao et al.

The geometric evolution of token representations in large language models (LLMs) presents a fundamental paradox: while human language inherently organizes semantic information in low-dimensional spaces ($\sim 10^1$ dimensions), modern LLMs employ high-dimensional embeddings ($\sim 10^3$ dimensions) processed through Transformer architectures. To resolve this paradox, this work bridges this conceptual gap by developing a geometric framework that tracks token dynamics across Transformers layers. Through layer-wise analysis of intrinsic dimensions across multiple architectures, we reveal an expansion-contraction pattern where tokens diffuse to a "working space" and then progressively project onto lower-dimensional submanifolds. Our finding implies a negative correlation between the working space dimension and parameter-sensitive performance of the LLMs, and indicates that effective models tend to compress tokens into approximately 10-dimensional submanifolds, closely resembling human semantic spaces. This work not only advances LLM interpretability by reframing Transformers layers as projectors that mediate between high-dimensional computation and low-dimensional semantics, but also provides practical tools for model diagnostics that do not rely on task-specific evaluations.

Zhuo-Yang Song, Zeyu Cai, Shutao Zhang et al.

Symbolic regression (SR), the automated discovery of mathematical expressions from data, is a cornerstone of scientific inquiry. However, it is often hindered by the combinatorial explosion of the search space and a tendency to overfit. Popular methods, rooted in genetic programming, explore this space syntactically, often yielding overly complex, uninterpretable models. This paper introduces IdeaSearchFitter, a framework that employs Large Language Models (LLMs) as semantic operators within an evolutionary search. By generating candidate expressions guided by natural-language rationales, our method biases discovery towards models that are not only accurate but also conceptually coherent and interpretable. We demonstrate IdeaSearchFitter's efficacy across diverse challenges: it achieves competitive, noise-robust performance on the Feynman Symbolic Regression Database (FSReD), outperforming several strong baselines; discovers mechanistically aligned models with good accuracy-complexity trade-offs on real-world data; and derives compact, physically-motivated parametrizations for Parton Distribution Functions in a frontier high-energy physics application. IdeaSearchFitter is a specialized module within our broader iterated agent framework, IdeaSearch, which is publicly available at https://www.ideasearch.cn/.