Searcharxiv⌕ Search

arXiv subjects

Yue Liu

Publications and source records attributed to Yue Liu.

At least 37 records · Page 2Linked to original sources
arXiv2026

TitanCA: Lessons from Orchestrating LLM Agents to Discover 100+ CVEs

Software vulnerabilities remain one of the most persistent threats to modern digital infrastructure. While static application security testing (SAST) tools have long served as the first line of defense, they suffer from high false-positive rates. This article presents TitanCA, a collaborative project between Singapore Management University and GovTech Singapore that orchestrates multiple large language model (LLM)-powered agents into a unified vulnerability discovery pipeline. Applied in open-source software, TitanCA has discovered 203 confirmed zero-day vulnerabilities and yielded 118 CVEs. We describe the four-module architecture, i.e., matching, filtering, inspection, and adaptation, and share key lessons from building and deploying an LLM-based vulnerability discovery solution in practice.

cs.CR↗
arXiv2026

AsmEvo: Agentic Assembly-Level Optimization of AMD GPU Kernels with Functional Equivalence Verification

High-performance ML systems increasingly rely on GPU kernels whose editable source is unavailable, generated, or too distant from final machine code to expose remaining optimizations. Existing LLM kernel optimizers and autotuners mainly operate on CUDA, Triton, HIP, or tensor-program source and validate against reference implementations. We study a stricter setting: optimizing an already compiled AMDGPU code object, where the deployed binary is the only behavioral oracle. We present AsmEvo, an agentic assembly-level optimizer for AMD GPU kernels. Given an AMDGPU code object K0, AsmEvo reconstructs a reassemblable representation, proposes low-level edits with a long-horizon agent, rebuilds an ABI-preserving optimized object, and accepts candidates only after differential verification against K0 under identical launches. AsmEvo combines code-object recovery, metadata-aware rebuilding, profiling-guided hot-window editing, correctness-gated timing, and conservative in-place patch fallback. We conduct extensive experiments with AsmEvo on various AMD GPU kernels. On MI308X, AsmEvo improves 29 of 30 selected KernelBench kernels, reaching 1.35x geometric-mean and 3.88x maximum speedup. On MI300X production workloads, it improves all evaluated AITer binaries and vLLM/SGLang Triton assembly kernels, reaching 1.09x/1.31x and 1.18x/1.34x geometric-mean/maximum speedups, respectively, while preserving functional equivalence.

cs.CL↗
arXiv2026

SCRIBES: Web-Scale Script-Based Semi-Structured Data Extraction with Reinforcement Learning

Semi-structured content in HTML tables, lists, and infoboxes accounts for a substantial share of factual data on the web, yet the formatting complicates usage, and reliably extracting structured information from them remains challenging. Existing methods either lack generalization or are resource-intensive due to per-page LLM inference. In this paper, we introduce SCRIBES (SCRIpt-Based Semi-Structured Content Extraction at Web-Scale), a novel reinforcement learning framework that leverages layout similarity across webpages within the same site as a reward signal. Instead of processing each page individually, SCRIBES generates reusable extraction scripts that can be applied to groups of structurally similar webpages. Our approach further improves by iteratively training on synthetic annotations from in-the-wild CommonCrawl data. Experiments show that our approach outperforms strong baselines by over 13% in script quality and boosts downstream question answering accuracy by more than 4% for GPT-4o, enabling scalable and resource-efficient web information extraction.

cs.CL↗
arXiv2026

A Network-driven Framework for Public Event Forecasting via Dynamic Interaction Network Evolution

Effective public event forecasting is essential for intelligent service systems, enabling proactive risk management, adaptive resource allocation, and timely decision-making. In many real-world scenarios, the evolution of public events is driven by dynamic interactions among participants. Motivated by this observation, this paper proposes auto-ibDLM, a network-driven deep learning framework that represents events as dynamic interaction networks and predicts public event evolution through participant growth forecasting. The proposed framework adopts a hybrid representation learning strategy that first represents network evolution using network science-informed structural metrics and subsequently transforms the resulting structural feature vectors into compact and robust latent representations through an auto-learning layer. A GRU-based temporal forecasting module is then employed to capture temporal dependencies and predict future participant growth. Extensive experiments on 13 real-world public event datasets and two publicly available dynamic network datasets demonstrate that auto-ibDLM consistently outperforms representative state-of-the-art methods in both forecasting accuracy and generalization capability, achieving over 97% accuracy in public event forecasting. Comprehensive experimental analyses further validate the effectiveness of the proposed hybrid representation learning strategy and demonstrate its representation-level interpretability. These results indicate that auto-ibDLM provides an effective and practical solution for intelligent public event forecasting.

cs.AI↗
arXiv2026

SWE-Bench ProMax: Benchmarking Agents on Large-Scale Multilingual Code Refactoring

As AI coding agents take on increasingly complex, long-horizon software engineering tasks, existing benchmarks are rapidly saturating and their evaluation quality has come under serious scrutiny: a recent audit found that nearly 60% of unsolved SWE-bench Verified instances contain flawed tests -- either overly narrow tests that reject correct solutions or overly broad tests that check unstated requirements -- and that frontier models can verbatim reproduce gold patches from training data. Code refactoring, which requires coordinated, behavior-preserving changes across many files, offers a substantially harder and more realistic test of agent capability, yet remains underserved by current benchmarks. We introduce SWE-Bench ProMax, an expert-curated, multilingual code refactoring benchmark of 170 instances drawn from real commits across seven programming languages (Python, Java, TypeScript, Go, C, C++, and Rust). Every instance undergoes rigorous, multi-stage curation that directly addresses the quality problems identified in prior benchmarks: issue descriptions are rewritten from scratch to provide precise, unambiguous specifications, and test suites are manually reviewed to remove overly narrow and overly broad tests. Tasks with insufficient complexity or limited cross-file scope are filtered out, yielding a benchmark of challenging, large-scale refactoring tasks that average 11.4 modified files and 261.6 lines of code per instance, substantially exceeding the scale of existing benchmarks. Experiments with frontier models under two agent scaffolds show that the best model achieves only 41.2% resolve rate, confirming that SWE-Bench ProMax presents a meaningful and unsaturated challenge for current AI coding agents. Our benchmark is available at https://huggingface.co/datasets/swe-bench-promax/SWE-Bench-ProMax.

cs.CL↗
arXiv2026

Entanglement Mpemba Effect

Generating entanglement rapidly and reliably is essential for quantum information processing, communication, and metrology. Dissipative preparation is attractive because engineered reservoirs robustly drive a system toward an entangled target, yet relaxation can carry a substantial time cost. Here we formulate the entanglement Mpemba effect, whereby an initially less entangled state overtakes a more entangled state under the same open-system dynamics. This effect turns initial-state engineering into a route for faster preparation without altering the dissipative protocol. We derive a general criterion for the reversal from the relaxation spectrum, applicable even when entanglement evolves nonmonotonically. A reversal of deterministic local operations and classical communication (LOCC)-reachability preorder provides a measure-independent certificate of reversed entanglement order. Exactly solvable models show that initial-state selection can substantially shorten the time required to reach high entanglement. We further propose an experimentally relevant trapped-ion protocol that can realize the entanglement Mpemba effect.

cond-mat.stat-mech↗
arXiv2026

Efficient Modeling of Surrogates to Improve Multi-source High-dimensional Integrative Regression

Surrogate variables play an important role in various fields due to the scarcity or absence of gold standard labels. We develop a novel approach named SASH for Surrogate-Assisted and data-Shielding High-dimensional integrative regression. It is a semi-supervised approach that efficiently leverages sizable unlabeled samples with error-prone surrogate outcomes from multiple local sites, to improve the learning accuracy of the small gold-labeled data. To facilitate stable and efficient knowledge extraction from the surrogates, our method first obtains a preliminary supervised estimator, and then uses it to assist training a regularized single index model (SIM) for the surrogates. Interestingly, through a chain of convex and properly penalized sparse regressions that approximate the SIM loss with bias-correction, our method avoids the local minima issue of the SIM, and fully eliminates the impact of the preliminary estimator's excessive error. In addition, it protects individual-level information through the aggregation of summary statistics from local sites, leveraging a similar idea of bias-corrected approximation. Through simulation studies, we demonstrate that our method outperforms existing approaches. Finally, we apply our method to develop a genetic risk model for type II diabetes using large-scale data sets from UK and Mass General Brigham biobanks, where only a small fraction of subjects in one site are labeled through chart reviewing.

stat.ME↗
arXiv2026

What Do Chinese-Language Generative Search Engines Cite and Surface? A Large-Scale Empirical Study

Generative AI question-answering systems increasingly mediate information access, shifting content visibility from ranked search results to retrieval, citation, and presentation in generated answers. We conduct a large-scale empirical study of Chinese-language generative search across the Web and App interfaces of four mainstream platforms. The controlled design covers eight platform interfaces, 614 queries, and three replications per query-platform-interface combination. From 214,119 raw records, we construct a cleaned citation-level dataset of 160,860 records and analyze citation behavior, source attribution, entity exposure, and cross-interface consistency. Five findings emerge. First, brands in the citation pool were selectively surfaced in answers: the overall brand-selection rate was 8.3%, and 12.4% of retrieved sources containing contact information contributed contact information to answers. Second, content fit, cross-source occurrence count, and semantic role were relatively important in predictive models, whereas the 5118-Baidu Composite Quality Score was not the leading predictor for any examined outcome. Third, among cited pages with publication dates, fitted half-lives were approximately 39 days for high-timeliness queries and 68 days for low-timeliness queries. Fourth, approximately 13% of brand exposures could not be matched to the contemporaneous citation pool, and approximately 71% of contact-information exposures could not be matched to the crawled body text. Fifth, source sets differed systematically between the App and Web interfaces of the same platform. These results characterize how Chinese-language generative search systems select, attribute, and surface information and show that interface type is an important dimension of analysis.

cs.IR↗
arXiv2026

The Mpemba effect likes to hit a wall

The original Mpemba effect refers to the idea that a macroscopic hot system cools faster than an initially colder one. In its microscopic and classical versions, the system is modeled as an overdamped particle in an external potential, and the corresponding Mpemba effect has been observed experimentally and explored theoretically. We establish that the existence of the one-dimensional Mpemba effect for a particle in an asymmetric polynomial double-well potential is driven solely by the presence of a hard wall, irrespective of the potential's double-well shape and metastability. The Mpemba effect disappears if we consider an infinitely large system. We then show that the underlying mechanism of the Mpemba effect is governed by the fine structure of the initial statistical population as it probes the tails of the potential, which also explains the Mpemba effect in single-well and symmetric double-well potentials.

cond-mat.stat-mech↗
arXiv2026

Predicting the conditions for observing the Mpemba effect

The Mpemba effect, a counterintuitive phenomenon where a hotter system relaxes faster than a colder one, has been widely observed in various nonequilibrium systems. Despite this progress, the fundamental structural features of the energy landscape required for its emergence remain a subject of debate. In this study, we investigate the conditions for the Mpemba effect within one-dimensional overdamped Langevin dynamics. We classify the potential landscapes based on the presence of single or double wells, their symmetry properties, and the existence of walls. We establish that the existence of the effect is primarily driven by the presence of boundaries, either hard or soft, rather than the specific internal structure of the potential landscape, such as metastability or the number of minima. By employing a spectral decomposition of the Fokker-Planck operator, we analyze the behavior of the first nontrivial eigenmode and demonstrate that its derivative acts as a Dirac delta peak in the low-temperature regime. This helps us elucidate the mechanism underlying the Mpemba effect: it appears as the interplay between this behavior and the initial population dynamics in a non-trivial way induced by the presence of the wall. Our analysis provides a unified classification across single- and double-well potentials, highlighting the crucial role of boundary conditions and asymmetry. Furthermore, we demonstrate that this framework allows for the engineering of potential landscapes capable of producing multistage Mpemba transitions.

cond-mat.stat-mech↗
arXiv2026

Nonreciprocal Relaxation Acceleration

Driven by recent discoveries regarding the quantum Mpemba effect, the anomalous relaxation dynamics of open quantum systems have garnered significant attention. While expediting thermalization to equilibrium has been extensively studied, dynamically accelerating the convergence toward nonequilibrium steady states remains a formidable challenge. In this article, we find a transient engineered nonreciprocal dissipative channel can provide a shortcut that accelerates convergence to the target reciprocal nonequilibrium steady state for the considered two-mode model and initial states. Using interacting bosonic modes, we demonstrate that the temporal activation of a nonreciprocal channel efficiently suppresses prolonged inter-mode energy oscillations, enforcing a rapid, unidirectional thermal dump into the environment. Counterintuitively, we find that this relaxation speedup is robust and independent of the direction of the nonreciprocity. Our results provide a powerful thermodynamic technique for rapid state preparation and cooling in continuous-variable quantum systems, particularly critical for low-temperature quantum information processing.

quant-ph↗
arXiv2026

Ai2-Kit: Streamlining AI-Accelerated Ab Initio Workflows for Complex Chemical Systems

Molecular simulations of complex chemical systems, such as catalysis, electrochemistry, and energy storage, often need to capture the interplay of effects such as electronic structure, finite-temperature fluctuations, and electric-field response. Such complexity is difficult to address with traditional ab initio calculations, which are limited by the time and length scales they can reach. AI-accelerated ab initio (AI2) methods use machine learning potentials trained on first-principles data to replace expensive electronic-structure calculations, extending ab initio accuracy to these regimes, but their routine application requires reliable workflows that connect first-principles calculations, model training, molecular dynamics, enhanced sampling, trajectory analysis, and HPC orchestration. Here we present ai2-kit, a software toolkit for developing accessible, reproducible, and extensible AI2 workflows. ai2-kit provides high-semantic-density command-line interfaces and Python APIs for structure and dataset conversion, batch task generation, active-learning screening, job orchestration, and workflow recovery. We demonstrate ai2-kit in four representative applications: active-learning-based machine learning potential construction, free-energy perturbation for redox and acid-base processes, electrochemical machine learning potentials for electrified interfaces, and spectroscopies from machine learning molecular dynamics. ai2-kit also provides AI-agent skills that help users adapt these use cases into customized workflows for their own chemical systems and computational software stacks. Together, ai2-kit helps turn AI2 methods from bespoke computational protocols into reusable and extensible workflows for complex chemical systems, from model construction to property prediction.

physics.chem-ph↗
arXiv2026

Guess Where You Go: Generative Next Point-of-Interest Recommendation in Amap

Generative retrieval enables recommender systems to retrieve items by generating compact item identifiers, but scaling it to industrial scenarios remains challenging due to redundant or colliding token assignments and insufficient integration of heterogeneous item signals. These challenges are particularly critical for next Point-of-Interest (POI) recommendation, where models must represent structured spatial entities, capture sequential mobility patterns, and produce predictions consistent with real user behavior. We propose Gwhere, an end-to-end industrial framework that integrates semantic identifier (SID) generation with LLM-based generative next POI recommendation. Gwhere first learns discriminative POI SIDs through a contrastive residual-quantization tokenizer that aligns textual, visual, spatial, and collaborative signals. Based on these SIDs, Gwhere adapts LLMs to mobility scenarios via continued pretraining on enriched spatio-temporal corpora, supervised fine-tuning, and Exposure-Aware Kahneman-Tversky Optimization (EAKTO), a reinforcement learning objective for behavioral preference alignment. Experiments on public datasets and Amap's large-scale industrial dataset demonstrate the effectiveness of Gwhere. The system has been deployed in Amap's homepage service under high-concurrency and low-latency constraints. Long-term online A/B tests show improvements of 5.83% in P-CTR and 6.20% in U-CTR over the production baseline. The implementation is publicly available at https://github.com/alibaba/SimCIT.

cs.IR↗
arXiv2026

Resolving Primitive-Sharing Ambiguity in Long-Tailed TLS-Based Industrial MEP Point Cloud Segmentation via Spatial Context Constraints

In terrestrial laser scanning (TLS)-based mechanical, electrical, and plumbing (MEP) point cloud segmentation, safety-critical components such as reducers and valves are persistently misclassifed, blocking reliable engineering knowledge extraction. This stems from a dual crisis--extreme class imbalance (215:1) compounded by geometric ambiguity, since most tail classes share cylindrical primitives with dominant head classes--that existing frequencybased re-weighting methods cannot resolve. We propose spatial context constraints that exploit neighborhood prediction consistency to disambiguate locally similar structures. Our approach extends Class-Balanced (CB) Loss with two architecture-agnostic mechanisms: Boundary-CB, an entropy-based constraint that emphasizes ambiguous boundaries and encodes an MEP assemblytopology prior, and Density-CB, a density-based constraint that compensates for scan-dependent variations and encodes TLS sensor-physics knowledge. Both operate at the loss level and integrate into existing pipelines without backbone modifcations. On the Industrial3D dataset (612.7M labelled points from water treatment facilities), our method achieves 55.74% mIoU, exceeding the strongest of three representative fully supervised backbone baselines (39.83-52.48% mIoU), with a 21.7% relative improvement on tail-class performance (29.59% vs. 24.32%) while preserving head-class accuracy (88.14%). Components with primitive-sharing ambiguity show strong gains: reducer improves from 0% to 21.12% IoU, and valve improves by 24.3% relative. These results show that spatial context constraints reduce primitive-sharing errors in the target industrial MEP setting and support more reliable identifcation of safety-critical components for Digital Twin and Scan-to-BIM applications. Code: https://github.com/PointCloudYC/LongTail3D.git.

cs.CV↗
arXiv2026

Hall Coefficient Sign Reversal Driven by Orbital-Selective Oxygen-Vacancy Scattering in Nickelate Films

Hall measurements in superconducting bilayer nickelate films show sign reversals that cannot be explained by rigid-band electron doping alone. We combine a DFT+CDMFT-derived correlated multi-orbital quasiparticle model with a $T$-matrix treatment of oxygen-vacancy scattering in a semiclassical Boltzmann transport framework. We find that multiband compensation is insufficient by itself: in-plane vacancies selectively suppress the transport channel dominated by the $d_{x^2-y^2}$ orbital and drive $R_H$ through zero, whereas inner-apical vacancies make $R_H$ more negative. These results identify pocket-resolved and orbital-selective oxygen-vacancy scattering as the microscopic origin of the Hall coefficient sign reversal and provide a framework for oxygen-stoichiometry-dependent transport in nickelate films.

cond-mat.supr-con↗
arXiv2026

Autonomous discovery of traffic laws with AI traffic scientists

Universal traffic laws describe recurrent patterns in congestion, mobility and driving behavior across cities, providing a scientific basis for transportation planning, management and control. Their discovery, however, remains expert-driven, requiring candidate regularities to be identified from heterogeneous observational evidence or validated through intervention experiments. Although autonomous artificial intelligence (AI) systems have advanced scientific discovery in controlled laboratory settings, extending them to complex transportation domains remains a challenge. Here we present TrafficSci, an agentic AI system that formulates traffic-law discovery as an iterative, auditable workflow integrating evidence scoping, critic-judge hypothesis induction, and observational-interventional validation. Across four case studies spanning population, network, control and trajectory scales, TrafficSci autonomously rediscovers three established traffic laws and identifies an unreported intrinsic temporal memory scale in urban driving behavior, statistically consistent across eight cities and two trajectory datasets. TrafficSci provides a route for extending AI-driven scientific discovery from controlled domains to complex urban systems.

cs.AI↗
arXiv2026

Sustainable Hybrid Document-Routed Retrieval for Financial RAG: Resolving the Robustness-Precision Trade-off

Retrieval-Augmented Generation (RAG) systems for financial document QA typically follow a chunk-based paradigm: documents are split into fragments, embedded, and retrieved by similarity. In structurally homogeneous corpora such as regulatory filings, this suffers from cross-document chunk confusion. Semantic File Routing (SFR), which uses LLM structured output to route queries to whole documents, reduces catastrophic failures but sacrifices targeted-chunk precision. We identify this robustness-precision trade-off on the FinDER benchmark (1,500 queries across five groups): SFR achieves higher average scores (6.45 vs. 6.02) and fewer failures (10.3% vs. 22.5%), while chunk-based retrieval (CBR) yields more perfect answers (13.8% vs. 8.5%). To resolve it, we propose Hybrid Document-Routed Retrieval (HDRR), a two-stage architecture that uses SFR as a document filter followed by chunk retrieval scoped to the identified document(s), eliminating cross-document confusion while preserving chunk precision. HDRR achieves the best performance on every metric: an average score of 7.54 (25.2% above CBR, 16.9% above SFR), a 6.4% failure rate, 67.7% correctness (+18.7 pp over CBR), and a 20.1% perfect-answer rate (+6.3 pp over CBR, +11.6 pp over SFR), simultaneously attaining the lowest failure rate and highest precision across all five groups. Beyond accuracy, HDRR is also the most efficient of the high-quality systems: it preserves CBR's compact per-query token budget (~5K-15K, an order of magnitude below SFR's ~50K-200K), incurs no indexing-time LLM spend (versus the one-time ~$100 cost of contextual indexing), and uses fewer per-query LLM calls than self-correcting agentic baselines, translating directly to lower API spend and inference-time energy at deployment scale.

cs.CL↗
arXiv2026

Position: Reasoning After Perception Means Reasoning Without Vision

A common belief in multimodal research is that the perceptual weaknesses of vision--language models can be compensated by stronger language reasoning (e.g., chain-of-thought, in-context learning, or external tools). We challenge this assumption. We argue that for a broad class of visual tasks hard to specify in language, failures stem from a structural fatality where the temporal decision of \textit{when} to reason strictly dictates the spatial constraint of \textit{where} reasoning takes place. When visual reasoning is deferred to language generation, current architectures do not merely delay computation; they displace it from the continuous visual representation to a discrete textual space. Consequently, the sequential ``Perception-then-Reasoning'' paradigm degenerates perception into a passive, one-off feature encoding process, rendering it functionally equivalent to ``Reasoning-in-Text-Space'', where task-critical spatial signals are collapsed before reasoning begins. We substantiate this claim with the Turing Eye Test (TET): tasks that must be resolved in \emph{visual space} and are hard to verbalize; results show text-only reasoning cannot remedy these perceptual failures. Our findings suggest rethinking the architectural divide: shifting from reasoning \textit{about} perception to reasoning \textit{within} perception. This facilitates actively reasoning-driven perception that operates directly on pixel-level visual representations, rather than within a collapsed textual space.

cs.CV↗