arXiv subjects

Beichen Huang

Publications and source records attributed to Beichen Huang.

12 recordsLinked to original sources

arXiv2026

Semantics-Guided Automatic Tensorization for Multiobjective Evolutionary Algorithms: A Multi-Agent Framework

Multiobjective evolutionary algorithms (MOEAs) naturally expose population-level parallelism, but many mature implementations encode their computation in sequential program structures designed for central processing units. Exploiting modern tensor computing platforms therefore requires more than direct code translation: the implementation must be restructured without changing the defining optimization mechanism of the underlying MOEA. We formulate automatic tensorization for MOEAs as semantics-guided computational restructuring and develop Evolutionary Code Conversion (EvoCoCo), a multi-agent framework that realizes this formulation. EvoCoCo reconstructs algorithm-specific states, dependencies, operators, and update logic into a structured semantic representation and organizes them through a shared tensorization blueprint. Specialized transformation branches then explore alternative tensor realizations, while execution feedback guides validation, repair, and candidate selection. Experiments on a benchmark of 48 MOEAs evaluate migration reliability, optimization fidelity, and computational scalability. Under matched large language model backends, EvoCoCo attains higher migration reliability than direct one-shot translation. Across the benchmark suites, 88.2% of valid comparisons satisfy the predefined optimization-fidelity criterion. The tensorized implementations also exhibit increasing acceleration on graphics processing units as population size or decision dimension grows, with median measured speedups ranging from $22.6\times$ under population scaling to $80.2\times$ under decision-dimension scaling. External-source and ablation studies further assess transfer beyond the main benchmark and the roles of the major framework components.

cs.NE↗

arXiv2026

Persistent Recursive Worlds Enable Autonomous Software Evolution

Complex software systems develop over timescales that exceed the lifespan of any individual coding agent. Most agentic software systems preserve continuity through persistent sessions, memories, managers or shared context. We introduce EvoX Genesis (hereafter, Genesis), which instead makes the software project persistent while allowing local agents to remain finite-lived. Genesis represents software as a persistent recursive world: each local world is situated by an accepted version and a repository path, finite-lived agents propose local changes, recursive delegation moves work across paths, and only accepted consequences advance the persistent version history. We evaluate this organization across formation, continuation and redevelopment. Starting from a repository with no compiler implementation, Genesis used DeepSeek V4 Flash to build a Rust-based C compiler with about 250k tracked lines; the run lasted over 120 hours, archived over 1,000 agent episodes and incurred only US\$44 in model-token charges. The compiler passed the complete c-testsuite and most LLVM and Csmith tests. In a separate compiler world generated with GLM 5.2, development continued after repeated agent replacement while retaining full test performance. Genesis also reimplemented 13 MESA modules with over 100k Fortran lines as a Rust workspace with nearly 90k Rust lines; across six numerical workloads, it achieved median speedups of 1.55--6.87x. These results show that long-horizon software development can be organized around a persistent project rather than a persistent agent. Project Website: https://genesis.evox.group/

cs.SE↗

arXiv2026

SearchAuditor: Auditing and Attributing Failures in Long-Horizon Search Agents

Deep search agents tackle challenging questions through long-horizon web interactions, a process that is both complex and fragile: small reasoning errors may propagate through long, noisy trajectories into fluent but incorrect answers. Diagnosing such failures is difficult, requiring the manual inspection of extremely long execution traces, which could be beyond human capacity. We therefore introduce SearchAuditBench, a benchmark that evaluates whether LLM auditors can localize, attribute, and repair these failures, thereby reducing the human burden. SearchAuditBench comprises 1,243 failed trajectories, averaging 73.1 messages and 65.1K tokens, collected from eight open-weight models on five deep-search benchmarks, each expert-annotated with the critical error step, a search-specific root cause, and a reference repair with grading rubrics. We further propose SearchAuditor, a multi-perspective auditing framework that effectively localizes, attributes, and repairs search-agent failures through evidence-grounded adjudication. Experimental results show that even the strongest baseline, when powered by a frontier model like GPT-5.5, attains only a 26.6% end-to-end pass rate. In contrast, our SearchAuditor consistently outperforms all baselines across different frontier models, achieving an end-to-end pass rate of 32.3%, and resuming failed runs with its repairs enables agents to better recover from errors.

cs.AI↗

arXiv2026

Quantization Inflates Reasoning: Token Inflation as a Hidden Cost of Low-Bit Reasoning Models

Quantization is widely used to reduce the inference cost of large language models, but its effect on reasoning models is not fully captured by final-answer accuracy or per-token latency. We show that low-bit post-training quantization can introduce a hidden test-time compute cost: quantized reasoning models often generate longer chains of thought even when they still answer correctly. Across mathematical reasoning, code generation, scientific question answering, and agentic tool-use benchmarks, we find that INT4/INT3 quantization can preserve accuracy but increase reasoning-token usage, offsetting the expected per-token speedup. To measure this effect, we introduce the CoT Token Inflation Ratio, which compares reasoning length between quantized and full-precision models averaged across all evaluation benchmarks. We further show that token inflation is accompanied by behavioral changes in the reasoning trace, including more intermediate steps and greater semantic repetition. These changes translate into measurable end-to-end real-world serving penalties. Finally, we evaluate mitigation strategies and find that prompting and decoding-time sampling offer inconsistent accuracy-length trade-offs, while quantization-aware training shows more promise in reducing both accuracy degradation and token inflation. Our results suggest that reasoning-token usage should be reported alongside accuracy when evaluating quantized reasoning models.

cs.AI↗

arXiv2026

A nuclear clock based on $^{229}$Th

Atomic clocks have made time and frequency the most precisely measured quantities in physics, progressing from microwave standards that realize the SI second to optical clocks that now reach unprecedented levels of precision. A nuclear clock would shift the frequency reference from an electronic transition to the uniquely low-lying, laser-accessible isomeric transition in the $^{229}$Th nucleus, offering a route to compact, robust timekeeping and sensitive tests of fundamental physics. However, turning recent advances in spectroscopy of the $^{229}$Th nuclear resonance into clock operation requires the nuclear transition to serve as a stable discriminator for steering a traceable oscillator. Here we demonstrate the operation of a $^{229}$Th nuclear clock by stabilizing a continuous-wave narrow-linewidth 148.4 nm vacuum-ultraviolet (VUV) laser to a resolved nuclear transition in a solid-state host. This clock operation is enabled by fast frequency discrimination based on phototube photocurrent readout of the transmitted VUV power. The 10 $\mu$W VUV laser, generated by four-wave mixing in cadmium vapour, provides a high-signal-to-noise absorption signal from a home-grown $^{229}$Th:CaF$_2$ crystal, allowing the laser to be locked to a weakly temperature-sensitive nuclear transition. The clock reaches a fractional frequency instability of $2\times10^{-12}/\sqrt{\tau/s} $, where $\tau$ is the averaging time. Remarkably, nuclear-clock frequencies measured with two distinct crystals agree at the $10^{-13}$ level, demonstrating the reproducibility of solid-state nuclear frequency references. By making a laser-addressed atomic nucleus an operational clock reference, this work extends quantum metrology from electronic to nuclear transitions, and opens a new platform for compact clocks, solid-state nuclear quantum sensors and precision tests of fundamental physics.

physics.atom-ph↗

arXiv2026

Hidden States as Early Signals: Step-level Trace Evaluation and Pruning for Efficient Test-Time Scaling

Large Language Models (LLMs) can enhance reasoning capabilities through test-time scaling by generating multiple traces. However, the combination of lengthy reasoning traces with multiple sampling introduces substantial computation and high end-to-end latency. Prior work on accelerating this process has relied on similarity-based or confidence-based pruning, but these signals do not reliably indicate trace quality. To address these limitations, we propose STEP: Step-level Trace Evaluation and Pruning, a novel pruning framework that evaluates reasoning steps using hidden states and dynamically prunes unpromising traces during generation. We train a lightweight step scorer to estimate trace quality, and design a GPU memory-aware pruning strategy that triggers pruning as the GPU memory is saturated by KV cache to reduce end-to-end latency. Experiments across challenging reasoning benchmarks demonstrate that STEP reduces end-to-end inference latency by 45%-70% on average compared to self-consistency while also improving reasoning accuracy. Our code is released at: https://github.com/Supercomputing-System-AI-Lab/STEP

cs.LG↗

arXiv2025

A continuous-wave vacuum ultraviolet laser for the nuclear clock

The exceptionally low-energy isomeric transition in $^{229}$Th at around 148.4 nm offers a unique opportunity for coherent nuclear control and the realisation of a nuclear clock. Recent advances, most notably the incorporation of large ensembles of $^{229}$Th nuclei in transparent crystals and the development of pulsed vacuum-ultraviolet (VUV) lasers, have enabled initial laser spectroscopy of this transition. However, the lack of an intense, narrow-linewidth VUV laser has precluded coherent nuclear manipulation. Here we introduce and demonstrate the first continuous-wave laser at 148.4 nm, generated via four-wave mixing (FWM) in cadmium vapor. The source delivers 100 nW of power with a linewidth well below 100 Hz and supports broad wavelength tunability. This represents a five-orders-of-magnitude improvement in linewidth over all previous single-frequency lasers below 190 nm, marking a major advance in laser technology. We develop a spatially resolved homodyne technique to place a stringent upper bound on the phase noise induced by the FWM process and demonstrate sub-hertz linewidth capability. These results eliminate the final technical hurdle to a $^{229}$Th-based nuclear clock, opening new directions in quantum metrology, nuclear quantum optics and precision tests of the Standard Model. More broadly, they establish a widely tunable, ultranarrow-linewidth laser platform for applications across quantum information science, condensed matter physics, and high-resolution VUV spectroscopy.

physics.atom-ph↗

arXiv2025

EvoGit: Decentralized Code Evolution via Git-Based Multi-Agent Collaboration

We introduce EvoGit, a decentralized multi-agent framework for collaborative software development driven by autonomous code evolution. EvoGit deploys a population of independent coding agents, each proposing edits to a shared codebase without centralized coordination, explicit message passing, or shared memory. Instead, all coordination emerges through a Git-based phylogenetic graph that tracks the full version lineage and enables agents to asynchronously read from and write to the evolving code repository. This graph-based structure supports fine-grained branching, implicit concurrency, and scalable agent interaction while preserving a consistent historical record. Human involvement is minimal but strategic: users define high-level goals, periodically review the graph, and provide lightweight feedback to promote promising directions or prune unproductive ones. Experiments demonstrate EvoGit's ability to autonomously produce functional and modular software artifacts across two real-world tasks: (1) building a web application from scratch using modern frameworks, and (2) constructing a meta-level system that evolves its own language-model-guided solver for the bin-packing optimization problem. Our results underscore EvoGit's potential to establish a new paradigm for decentralized, automated, and continual software development. EvoGit is open-sourced at https://github.com/BillHuang2001/evogit.

cs.DC↗

arXiv2025

MiLo: Efficient Quantized MoE Inference with Mixture of Low-Rank Compensators

A critical approach for efficiently deploying Mixture-of-Experts (MoE) models with massive parameters is quantization. However, state-of-the-art MoE models suffer from non-negligible accuracy loss with extreme quantization, such as under 4 bits. To address this, we introduce MiLo, a novel method that augments highly quantized MoEs with a mixture of low-rank compensators. These compensators consume only a small amount of additional memory but significantly recover accuracy loss from extreme quantization. MiLo also identifies that MoEmodels exhibit distinctive characteristics across weights due to their hybrid dense-sparse architectures, and employs adaptive rank selection policies along with iterative optimizations to close the accuracy gap. MiLo does not rely on calibration data, allowing it to generalize to different MoE models and datasets without overfitting to a calibration set. To avoid the hardware inefficiencies of extreme quantization, such as 3-bit, MiLo develops Tensor Core-friendly 3-bit kernels, enabling measured latency speedups on 3-bit quantized MoE models. Our evaluation shows that MiLo outperforms existing methods on SoTA MoE models across various tasks.

cs.LG↗

arXiv2024

Proposal for the generation of continuous-wave vacuum ultraviolet laser light for Th-229 isomer precision spectroscopy

We propose to generate continuous-wave vacuum ultraviolet (VUV) laser light at 148.4 nm using four-wave mixing in cadmium vapor for precision spectroscopy of the Th-229 isomer transition. Due to the large transition matrix elements of cadmium, the readily accessible wavelengths for the incident laser beams, and the high coherence of the four-wave mixing process, over 30 $\mu$W of VUV power can be generated with a narrow linewidth. This development paves the way for coherently driving the Th-229 isomer transition and developing the nuclear optical clock.

physics.atom-ph↗

arXiv2024

Exploring the True Potential: Evaluating the Black-box Optimization Capability of Large Language Models

Large language models (LLMs) have demonstrated exceptional performance not only in natural language processing tasks but also in a great variety of non-linguistic domains. In diverse optimization scenarios, there is also a rising trend of applying LLMs. However, whether the application of LLMs in the black-box optimization problems is genuinely beneficial remains unexplored. This paper endeavors to offer deep insights into the potential of LLMs in optimization through a comprehensive investigation, which covers both discrete and continuous optimization problems to assess the efficacy and distinctive characteristics that LLMs bring to this field. Our findings reveal both the limitations and advantages of LLMs in optimization. Specifically, on the one hand, despite the significant power consumed for running the models, LLMs exhibit subpar performance in pure numerical tasks, primarily due to a mismatch between the problem domain and their processing capabilities; on the other hand, although LLMs may not be ideal for traditional numerical optimization, their potential in broader optimization contexts remains promising, where LLMs exhibit the ability to solve problems in non-numerical domains and can leverage heuristics from the prompt to enhance their performance. To the best of our knowledge, this work presents the first systematic evaluation of LLMs for numerical optimization. Our findings pave the way for a deeper understanding of LLMs' role in optimization and guide future application of LLMs in a wide range of scenarios.

cs.NE↗

arXiv2023

EvoX: A Distributed GPU-accelerated Framework for Scalable Evolutionary Computation

Inspired by natural evolutionary processes, Evolutionary Computation (EC) has established itself as a cornerstone of Artificial Intelligence. Recently, with the surge in data-intensive applications and large-scale complex systems, the demand for scalable EC solutions has grown significantly. However, most existing EC infrastructures fall short of catering to the heightened demands of large-scale problem solving. While the advent of some pioneering GPU-accelerated EC libraries is a step forward, they also grapple with some limitations, particularly in terms of flexibility and architectural robustness. In response, we introduce EvoX: a computing framework tailored for automated, distributed, and heterogeneous execution of EC algorithms. At the core of EvoX lies a unique programming model to streamline the development of parallelizable EC algorithms, complemented by a computation model specifically optimized for distributed GPU acceleration. Building upon this foundation, we have crafted an extensive library comprising a wide spectrum of 50+ EC algorithms for both single- and multi-objective optimization. Furthermore, the library offers comprehensive support for a diverse set of benchmark problems, ranging from dozens of numerical test functions to hundreds of reinforcement learning tasks. Through extensive experiments across a range of problem scenarios and hardware configurations, EvoX demonstrates robust system and model performances. EvoX is open-source and accessible at: https://github.com/EMI-Group/EvoX.

cs.NE↗