arXiv subjects

Siyuan Chen

Publications and source records attributed to Siyuan Chen.

At least 19 recordsLinked to original sources

arXiv2026

Poisson bialgebras by deformations-to-quasiclassical limits

Poisson algebras are the quasiclassical limits of associative algebra deformations of commutative associative algebras. This paper extends this process to the level of bialgebras. We derive Poisson bialgebras as the quasiclassical limits of antisymmetric infinitesimal bialgebra deformations of commutative and cocommutative antisymmetric infinitesimal bialgebras. It might be regarded as the ``infinitesimal" version of the quantization process of Lie bialgebras in terms of Hopf algebras. Such deformations-quasiclassical limits process for a Poisson bialgebra is equivalently characterized in terms of the introduced notions of deformations of a matched pair of associative algebras as well as a standard Manin triple of associative algebras, whose corresponding quasiclassical limits are a matched pair of Poisson algebras and a standard Manin triple of Poisson algebras, respectively. We illustrate these equivalent deformations-quasiclassical limits processes via coherent derivations.

math.QA↗

arXiv2026

Task-Specific Prompt with Global Context for Multi-Task Graph Pre-Training

Graph prompt learning is an effective paradigm to adapt pre-trained graph models to downstream tasks in low-resource scenarios. However, existing multi-task graph pre-training frameworks generally use randomly initialized prompts, leading to poor alignment between the prompt space, pretext objectives and graph structural characteristics. This greatly weakens the task relevance, structural awareness and transferability of prompt representations. To address this challenge, we propose TPGC, a dual-prior prompt initialization solution that explicitly models the synergy between task prior and structural prior. Specifically, the Task-Prior Injection Module first conducts a short homologous multi-task pre-training on an auxiliary graph, enabling prompt initialization to inherit optimization preferences associated with multiple pretext tasks. Built on the task-aware representations, the Structure-Prior Injection Module further extracts transferable global structural context from the auxiliary graph, converting it into layer-wise prompt vectors by aggregating structurally informative node embeddings. Extensive experiments on 6 mainstream benchmarks covering node and graph classification show that TPGC achieves consistently better performance under few-shot settings than state-of-the-art baselines, with fewer downstream tunable parameters and lower runtime. The code is available at https://github.com/Virgilqiu/TPGC

cs.LG↗

arXiv2026

Beyond Scaling: Self-Evolving LLM Agents for Hardware Kernel Optimization via an Experience-Driven Workflow and Experience Graph Memory

Hardware kernel optimization requires repeated compilation, correctness testing, profiling, and revision. LLM agents can automate parts of this process, and stronger foundation models, longer context windows, and longer execution horizons have improved optimization within individual tasks. These advances alone do not enable an agent to learn from completed optimization runs. Existing kernel-optimization agents seldom preserve a decision, its observed execution feedback, and the later decisions that use that evidence. Retaining every prior trajectory is also impractical because an expanding history competes with the current task for context. We present KOPE, an experience-driven framework for hardware kernel optimization. KOPE records optimization trajectories with correctness and performance feedback in Experience Graph Memory, then uses Active Context Management and Injection to retrieve relevant experience under a fixed token budget. The graph retains decision order, observed outcomes, and alternative branches, allowing evidence collected on the target hardware to inform later optimization steps and tasks. Under the same GLM-5.2 setting, the geometric mean of KOPE's per-operator speedups is $1.54\times$ that of CANNBot, the strongest competing baseline. In a complete 53-operator ablation, Active Context Management and Injection raises pass rate from 60.0\% to 84.6\%, increases the evaluator-reported positive-field geometric mean from 0.0382 to 0.0661, and reduces optimization token consumption from 15.9B to 1.113B tokens relative to passive agent-led context construction. Enabling Experience Graph Memory raises full-suite pass rate from 55.2\% to 84.6\% and yields a $1.43\times$ geometric-mean speedup on valid timing comparisons. These results support continual optimization through external experience while the foundation model remains fixed.

cs.LG↗

arXiv2026

TRACE: An Evidence-Grounded Benchmark for Safety Evaluation of Large Reasoning Models

Large Reasoning Models (LRMs) generate intermediate reasoning traces that may contain unsafe content, even when their final responses appear safe. Guardrail models are designed to detect and block unsafe content, yet existing benchmarks for unsafe content detection focus primarily on prompts and final responses, leaving reasoning traces largely unexamined. Moreover, these benchmarks typically provide only binary safety labels, without evidence annotations that justify the judgments. To address these limitations, we introduce TRACE, an evidence-grounded safety evaluation benchmark that covers the entire LRM inference pipeline: prompts, reasoning traces, and final responses. TRACE includes prompts in two languages spanning nine risk categories and ten attack strategies. For each prompt, four LRMs generate reasoning traces and final responses, and we annotate the safety of each component and extract supporting evidence from the corresponding source text. Evaluating 18 guardrail models on TRACE reveals that safety judgment for reasoning traces is substantially more challenging than for prompts or final responses, and that current models struggle to accurately extract supporting evidence. These findings highlight the need for guardrail models that can reliably detect and precisely localize unsafe content across the LRM inference pipeline.

cs.AI↗

arXiv2026

Retrieval-grounded robot program generation and simulation-based correction via Model Context Protocol

Flexible manufacturing requires industrial robots to be reprogrammed rapidly as product variants change. This paper presents a language-model-based workflow that generates, validates, and iteratively corrects ABB RAPID robot programs from natural language task descriptions. A dual-stream retrieval-augmented generation (RAG) pipeline grounds code generation in verified technical documentation and production templates, reducing domain-specific errors produced by ungrounded language models. A custom Model Context Protocol (MCP) server connects the language-model client directly to ABB RobotStudio for automated code upload, simulation execution, and diagnostic feedback. The evaluation combines a 30-query retrieval benchmark, scoped code-generation checks, and RobotStudio case studies in a simulated pickand- place manufacturing cell. The simulation loop exposes execution failures that static and semantic checks alone cannot catch, including suction release-height errors, unreachable placement targets, and configuration-dependent recovery motions. The results show how RAG and MCP can connect grounded code generation with executable feedback from industrial robot simulation software, while reducing but not eliminating expert setup and final supervision.

cs.AI↗

arXiv2026

Nitrogen Vacancy Centers in Hexagonal Diamond Exhibit Long Coherence Times

We show that negatively charged nitrogen-vacancy (NV) centers in the hexagonal diamond polymorph lonsdaleite offer a route to spin qubits with enhanced coherence relative to their cubic-diamond counterparts. Using first-principles calculations, we examine two distinct defect configurations, AA, with the same symmetry as in cubic diamond and AB, with reduced symmetry. We find that the AB configuration of the NV center exhibits a finite transverse zero-field splitting, giving rise to an approximate fourfold enhancement of the Hahn-echo coherence time $T_2$ at zero magnetic field. The AA configuration, by contrast, closely reproduces the electronic structure and coherence properties of the cubic NV center. We further characterize the many-body electronic structure, vertical excitation energies, and photoluminescence spectra of both configurations, providing spectral fingerprints for their experimental identification. Our results establish symmetry-broken NV centers in lonsdaleite as promising candidates for quantum sensing and information science applications.

cond-mat.mtrl-sci↗

arXiv2026

Don't Regenerate, Debug: A Domain-Specific Agent for Repairing Near-Miss Hardware Operators

Kernel generation for hardware accelerators such as GPUs and NPUs has become a proving ground for large language models (LLMs), and state-of-the-art systems raise correctness through pipelines that couple LLMs with agentic reinforcement learning and evolutionary search. Such pipelines generate, compile, and execute large numbers of candidate kernels, discarding most of them and forgoing the opportunity to distill failures into reusable knowledge. Many discarded candidates are near-miss operators that compile and run but fail numerical validation; each embodies genuine domain knowledge and a nontrivial investment in LLM inference, cross-compilation, and hardware execution. We argue for a paradigm shift: rather than regenerate, debug. Debugging is far more constrained than generating from scratch: the search space is small and feedback is dense. We present a domain-specific debug agent that addresses three core challenges in autonomous repair: mitigating knowledge scarcity through retrieved patterns and diagnostic instrumentation, ensuring integrity through anti-cheat detection and full-coverage evaluation, and controlling cost via convergence guards and bounded iteration. Debugging serves two complementary roles: it extends the capability frontier by recovering operators that repeated regeneration fails to produce, and it lowers cost per deliverable operator. Debug Pass@1 achieves 66.7% versus Regenerate Avg Pass@1's 25.9% and Regenerate Pass@3's 40.7%, while consuming 92.8% fewer tokens per success than three-trial regeneration. Component ablations show that the knowledge base drives recovery, while integrity gates reject 12.5-33.3% of the successes the workflow itself accepted.

cs.SE↗

arXiv2026

The WEST code for large-scale excited-state materials simulations

We present WEST, an open-source plane-wave pseudopotential code for large-scale excited-state materials simulations, and describe its theoretical foundations, software architecture, and capabilities. WEST implements full-frequency GW, quantum defect embedding theory, the Bethe-Salpeter equation, and time-dependent density functional theory within a common algorithmic framework that avoids the explicit computation of virtual electronic states. By combining density functional and density matrix perturbation theory, low-rank representations of the dielectric screening and exact exchange, and localization techniques, WEST achieves favorable computational scaling with system size. The code supports the calculation of quasi-particle and neutral excitation energies, optical and photoluminescence spectra, excited-state forces, and non-adiabatic couplings, with interoperable workflows connecting to quantum chemistry, vibronic coupling, and quantum computing packages. A hierarchical parallelization strategy and GPU acceleration deliver near-ideal strong scaling to thousands of GPUs, enabling accurate excited-state simulations of systems with more than a thousand atoms. Representative applications, spanning the full optical cycle of solid-state spin defects, self-trapped excitons in metal-halide perovskites, and the optical response of liquid water and ice, demonstrate the accuracy and versatility of the code across diverse material classes. The capabilities implemented in WEST establish the code as a scalable platform for predictive excited-state simulations, high-throughput materials discovery, and the generation of high-fidelity datasets for machine learning in computational materials science.

cond-mat.mtrl-sci↗

arXiv2026

The SKAO Pulsar Timing Array

Pulsar timing arrays (PTAs) are ensembles of millisecond pulsars observed for years to decades. The primary goal of PTAs is to study gravitational-wave astronomy at nanohertz frequencies, with secondary goals of undertaking other fundamental tests of physics and astronomy. Recently, compelling evidence has emerged in established PTA experiments for the presence of a gravitational-wave background. To accelerate a confident detection of such a signal and then study gravitational-wave emitting sources, it is necessary to observe a larger number of millisecond pulsars to greater timing precision. The SKAO telescopes, which will be a factor of three to four greater in sensitivity compared to any other southern hemisphere facility, are poised to make such an impact. In this chapter, we motivate an SKAO pulsar timing array (SKAO PTA) experiment. We discuss the classes of gravitational waves present in PTA observations and how an SKAO PTA can detect and study them. We then describe the sources that can produce these signals. We discuss the astrophysical noise sources that must be mitigated to undertake the most sensitive searches. We then describe a realistic PTA experiment implemented with the SKA and place it in context alongside other PTA experiments likely ongoing in the 2030s. We describe the techniques necessary to search for gravitational waves in the SKAO PTA and motivate how very long baseline interferometry can improve the sensitivity of an SKAO PTA. The SKAO PTA will provide a view of the Universe complementary to those of the other large facilities of the 2030s.

astro-ph.IM↗

arXiv2026

HC-RAG: Evidence-Centric Retrieval-Augmented Generation over Heterogeneous Financial Filings

Financial question answering over annual reports requires more than retrieving semantically similar passages. It often involves identifying relevant companies and fiscal years, locating standardized filing sections, collecting textual and tabular evidence, and checking answers against the original documents. Existing RAG systems, however, usually flatten long filings into unordered chunks, pay limited attention to the typed structure of financial reports, and use fixed text-table fusion strategies without considering query intent. To address these limitations, we propose \textbf{HC-RAG}, a hierarchical cross-modal retrieval-augmented generation framework for evidence-centric financial QA. HC-RAG organizes filings into a typed financial evidence graph with documents, sections, text units, table units, and metadata nodes. It retrieves evidence through document-section-unit paths, aligns textual and tabular evidence in a shared retrieval space, and routes evidence according to four semantic intents: calculation, trend, fact, and comparison. We further introduce \textbf{Multi-Doc-2025}, a benchmark containing 2,327 expert-verified QA pairs from 179 SEC 10-K filings of 87 S\&P 500 companies across fiscal years 2022--2024, with labels for intent, difficulty, and structural evidence attributes. Experiments on public financial QA benchmarks and Multi-Doc-2025 show that HC-RAG improves both answer quality and evidence localization, especially in long-document, table-related, and cross-document settings. HC-RAG outperforms RAPTOR by 6.6 F1 points on DocFinQA and GraphRAG by 10.9 F1 points on Multi-Doc-2025. Evidence-level analysis and ablation studies show that the improvements mainly come from more accurate section localization, table grounding, cross-document evidence aggregation, and intent-aware text-table routing.

cs.CL↗

arXiv2026

SIRI: Self-Internalizing Reinforcement Learning with Intrinsic Skills for LLM Agent Training

Long-horizon LLM agents can benefit from reusable skills, yet existing skill-based methods often rely on external skill generators during training or persistent skill retrieval at inference, increasing engineering complexity, context length, and deployment latency. We propose Self-Internalizing Reinforcement learning with Intrinsic skills (SIRI), a three-phase framework that enables agents to discover, validate, and internalize skills without external skill generators or inference-time skill banks. SIRI first warms up the policy with GiGPO to acquire basic interaction ability and collect successful skill-free trajectories. It then performs self-skill mining, where the current policy summarizes compact skills from its own successful plain rollouts and validates them through paired skill-augmented and skill-free rollouts. Finally, SIRI distills only beneficial skill-guided action tokens into the plain policy using trajectory-level utility and action-level advantage. At inference, the agent runs with the original prompt only. On ALFWorld and WebShop with Qwen2.5-7B-Instruct, SIRI improves GiGPO from 0.908 to 0.930 on ALFWorld and from 0.728 to 0.813 on WebShop, outperforming prompt-based, RL-based, and memory-augmented baselines. Further analysis shows that our self-mining strategy can achieve performance comparable to distillation with closed-source large model. Our code is available at https://github.com/kirito618/SIRI.

cs.AI↗

arXiv2026

Closing Trajectories: Equation-Free Cyclic Animation via Koopman Surrogates

Cyclic animation is widely used in computer graphics and interactive content.It supports seamless playback in games, VR, and interactive simulation,where short clips must repeat smoothly over long durations. Achievingphysically plausible cyclic synthesis from an input sequence is challengingbecause the endpoint states of the observed sequence rarely match exactly,and the governing equations of the underlying system are often unavailable.We therefore propose an equation-free framework that identiffes a Koopmansurrogate from the observed trajectory and computes a cyclic trajectory byapplying a Fourier-parameterized, time-varying control force under a hardtemporal periodicity constraint. The resulting formulation reduces cyclicsynthesis to a linearly constrained quadratic program that can be solvedefffciently through a structured KKT system. Our method is applicable toa diverse range of examples, including N-body systems, cloth, deformableobjects, shallow water, etc.

cs.GR↗

arXiv2026

DualMem: Bypassing the Objectness Bottleneck for Calibrated Unknown-Stream Filtering in Open-World Object Detection

Open-world object detection (OWOD) requires detectors to localize known classes while identifying unknown objects for future incremental learning. We find that the unknown prediction streams of strong OWOD detectors are heavily polluted: on M-OWODB, across PROB, OW-DETR, and HypOW, future-task positive unknowns make up less than 10% of unknown predictions, whereas background false positives account for 46-71%. We show that this is not a missing-information problem, but an information bottleneck at the objectness head. On PROB Task 1, a linear probe on the 256-D decoder query achieves an AUROC of 0.908 for positive-versus-negative unknown discrimination, but the final one-dimensional objectness scalar drops to 0.642. A frozen SigLIP feature, without access to the detector, independently recovers much of this proposal-level separability at the filtering stage (AUROC = 0.871). Motivated by this finding, we propose DualMem, a calibrated post-hoc filter that assumes a small image-disjoint annotated calibration split of held-out future-task objects and performs a non-parametric likelihood ratio test in frozen SigLIP feature space. DualMem uses a k-nearest-neighbor positive memory to protect future-task objects and a negative memory to suppress background-like proposals. Its decision threshold is chosen by Neyman-Pearson calibration, giving users an explicit trade-off between false-unknown suppression and novel recall. Across PROB, OW-DETR, and HypOW on M-OWODB Task 1, DualMem reduces background-type false unknown proposals per image by 44.9%-66.3%, with a mean reduction of 56.6%. On PROB Task 1, it more than doubles the reduction achieved by a natural K-means prototype baseline, while leaving known-class mAP unchanged because known detections bypass the filter.

cs.CV↗

arXiv2026

WorldParticle: Unified World Simulation of Lagrangian Particle Dynamics via Transformer

A unified simulator that can model diverse physical phenomena without solver-specific redesign is a long-standing goal across simulation science. We present a learning-based particle simulator built on a single transformer architecture to model cloth, elastic solds, Newtonian and non-Newtonian fluids, granular materials, and molecular dynamics. Our model follows a prediction-correction design on a shared Lagrangian particle representation. An explicit predictor first advances particles under the known external forces, producing an intermediate state that captures externally driven motion but not inter-particle interactions. A learned corrector then predicts the residual position and velocity updates through three stages: a particle tokenizer that encodes local particle-particle, particle-boundary, and topology-guided interactions; a super-token encoder that hierarchically merges particle tokens into a compact set of super tokens via alternating self-attention and token merging; and a super-token decoder that lifts these super tokens back to particle resolution through cross-attention to predict per-particle position and velocity corrections. Progressive token merging reduces the attention cost at successive encoder layers by halving the token count at each level, and the decoder communicates through the compact super-token set rather than full particle-to-particle attention. Across the six dynamics categories, the same architecture generalizes to unseen materials, boundary configurations, initial conditions, and external forces. We further demonstrate downstream interactive control, inverse design, and learning from real-world manipulation data, reducing the need for per-phenomenon solver engineering.

cs.GR↗

arXiv2026

Constraints on Ultralight Scalar and Dark Photon Dark Matter from PPTA-DR3 and EPTA-DR2

The cold dark matter model successfully describes the Universe on large scales, yet faces challenges at sub-galactic scales. Ultralight dark matter (ULDM), with particle masses around $10^{-22} \mathrm{eV}$, offers a promising solution to these small-scale issues. Pulsar Timing Arrays (PTAs), designed to detect nanohertz gravitational waves, can also provide a sensitive probe for ULDM signals. In this work, we perform a Bayesian search for ULDM using PTA data sets, focusing on two types of signals: the oscillatory gravitational potential from scalar ULDM and the fifth-force interaction mediated by dark photon dark matter (DPDM). We incorporate pulsar distances in the analysis to better model the ULDM density. No statistically significant evidence for ULDM has been found, therefore we place 95% confidence-level upper limits on the relevant parameters. For scalar ULDM, our analysis does not exclude the scenario in which ULDM constitutes all of dark matter. The constraints from PPTA-DR3 show significant improvements over the earlier PPTA-DR2 (2018 Preview) across most of the mass range, and are consistent with the recent uncorrelated limits from other PTAs. We also present for the first time the DPDM constraints using EPTA data. The obtained bounds on the DPDM from the EPTA-DR2 and PPTA-DR3 are comparable to existing constraints.

astro-ph.CO↗

arXiv2026

CAGE-SGG: Counterfactual Active Graph Evidence for Open-Vocabulary Scene Graph Generation

Open-vocabulary scene graph generation (SGG) aims to describe visual scenes with flexible and fine-grained relation phrases beyond a fixed predicate vocabulary. While recent vision-language models greatly expand the semantic coverage of SGG, they also introduce a critical reliability issue: predicted relations may be driven by language priors or object co-occurrence rather than grounded visual evidence. In this paper, we propose an evidence-rounded open-vocabulary SGG framework based on counterfactual relation verification. Instead of directly accepting plausible relation proposals, our method verifies whether each candidate relation is supported by relation-pecific visual, geometric, and contextual evidence. Specifically, we first generate open-vocabulary relation candidates with a vision-language proposer, then decompose predicate phrases into soft evidence bases such as support, contact, containment, depth and state. A relation-conditioned evidence encoder extracts predicate-relevant cues, while a counterfactual verifier tests whether the relation score decreases when necessary vidence is removed and remains stable under irrelevant perturbations. We further introduce contradiction-aware predicate learning and graph-level preference optimization to improve fine-grained discrimination and global graph consistency. Experiments on conventional, open-vocabulary, and panoptic SGG benchmarks show that our method consistently improves standard recall-based metrics, unseen predicate generalization, and counterfactual grounding quality. These results demonstrate that moving from relation generation to relation verification leads to more reliable, interpretable, and evidence-grounded scene graphs.

cs.CV↗

arXiv2026

Positron Transport System for Muonium-to-Antimuonium Conversion Experiment

Muonium-to-Antimuonium Conversion Experiment (MACE) aims to find the charged lepton flavor violation (cLFV) process. A key component of MACE is the positron transport system (PTS) to collect and transport atomic positrons from antimuonium decays, which consists of an electrostatic accelerator and a solenoid beamline. Through field simulations in \textsc{COMSOL} and particle transport simulations based on \textsc{Geant4}, the PTS can achieve a geometric acceptance of the signal at 65.81(4)\% along with a position resolution of 88(1)~$μ$m~$\times$~102(1)~$μ$m. The system achieves 322.4(1)~ns transit time with a spread of 6.9(1)~ns, which allows for a TOF-based rejection of internal conversion backgrounds by a factor of $10^{-7}$. These promising results pave the way for new-physics signal identifications and background rejections in MACE and offer a novel paradigm for internal transport system in high-intensity frontiers.

hep-ex↗

arXiv2026

To understand the radiative processes of pulsars and fast radio bursts with the FAST

The radiative mechanism of coherent radio emission has remained an enigma since the discovery of pulsars, even the emergence of fast radio bursts (FRBs), which exhibit similarities to the single-pulse behavior of pulsars and have opened a new view for deciphering the long-standing mystery. Besides tremendous efforts in modelling, advanced facilities matter for solving the problem. The authors review the observational breakthroughs from the Five-hundred-meter Aperture Spherical radio Telescope (FAST), which are providing pivotal insights to unravel the underlying physics of pulsars and FRBs. This study offers a novel perspective in the era when pulsars meet FRBs, and further investigations are encouraged to utilize the highly sensitive telescope, the FAST.

astro-ph.HE↗