Searcharxiv⌕ Search

arXiv subjects

Jinyang Li

Publications and source records attributed to Jinyang Li.

At least 19 recordsLinked to original sources
arXiv2026

Proactive Memory for Event-Driven LLM Agents

Proactive LLM agents anticipate service opportunities from ongoing observations and accumulated history. Existing approaches emphasize recent context or retrieval of related memories, providing limited support for validating recurring behavioral regularities and assessing their applicability to the current event. We propose Proactive Memory (ProactMem), which organizes history into recent events, temporal trajectories, and behavioral regularities. Trajectory facts are checked against source events, and regularities are assessed through historical replay before being matched to current conditions. We also construct ProactiveStream, a longitudinal benchmark covering three application scopes, where agents process isolated event streams chronologically under episode-aware opportunity labels. On ProactiveStream, ProactMem achieves an overall effectiveness (F1) of 52.6%, a 21.4% relative improvement over the strongest of five external baselines. Code is available at: https://anonymous.4open.science/r/ProactMem-5351/.

cs.AI↗
arXiv2026

The General Quantum Limit for and the Optimization of the Minimum Measurable Frequency Shift in a Laser

We show that, contrary to conventional understanding, the minimum measurable frequency shift (MMFS) for a single-mode ideal laser is determined by a combination of phase diffusion caused by SE and the shot noise caused by the VM. For all practical sensors, the MMFS is found to be given by the geometric mean of the measurement bandwidth and the Schwalow-Townes Linewidth, multiplied by a factor which can be much larger than unity under certain conditions. We determine the optimal value for the MMFS for three different sensing modalities, an unbalanced Mach-Zehnder Interferometer, a passive Fabry-Perot cavity (FPC), and heterodyning with a reference laser, and identify the conditions needed for reaching the optimal value.

quant-ph↗
arXiv2026

BIRD-RL: Scaling Agentic Reinforcement Learning over Stateful Data-Centric Environments

While agentic reinforcement learning (RL) has emerged as a promising paradigm for training small language models (SLMs), existing frameworks are predominantly tool-centric and struggle with data-centric environments. Specifically, applying tool-centric frameworks to stateful tasks like SQL debugging introduces severe spatial inefficiencies since proper environment isolation typically requires a separate database container per rollout. Naively reusing databases without tracking evolving states across multi-turn interactions causes trajectory contamination and cross-trajectory interference, which severely destabilizes policy updates. To overcome these bottlenecks, we introduce BIRD-RL, a novel four-stage agentic RL curriculum that alternately enhances foundational agent reasoning and trajectory-adaptive on-policy exploration to circumvent the instability inherently associated with direct end-to-end learning. Crucially, to support this curriculum, BIRD-RL is driven by a Trajectory-scoped Persistent Execution Infrastructure, which enables safe database replica reuse while preserving trajectory isolation. This design improves spatial RL efficiency by 12.8x, successfully enabling scalable agentic RL for complex data-centric tasks without needs of Kubernettes. Empirical evaluations on the BIRD-CRITIC-SQLite benchmark demonstrate that our 7B and 14B models achieve Success Rates of 44.40% and 48.00%, respectively, showing competitive performance against strong proprietary model-based agents such as Claude-Opus-4.6 Agent (48.20%). Furthermore, BIRD-RL successfully trains unified SLMs capable of both advanced SQL debugging and generation, highlighting its broader potential for multi-objective optimization across heterogeneous, real-world data-centric tasks.

cs.LG↗
arXiv2026

RouterWorld: Scaling Conversational AI with Latency-Aware LLM Routing

Conversational AI systems built on Large Language Models (LLMs) require routing mechanisms that balance answer quality and response latency at scale. However, existing LLM routing benchmarks assume full observability of all candidate model outcomes, under-emphasize latency as a first-class metric, and fragment evaluation across heterogeneous protocols misaligned with real-world settings. We introduce RouterWorld, a comprehensive benchmark for LLM routing in conversational GenAI. RouterWorld unifies 371k single-turn responses and 76k multi-turn agentic trajectories, adopts natural user interactions with a standardized LLM-as-a-judge protocol, and decouples efficiency from transient pricing via hardware-grounded latency annotations. Crucially, RouterWorld formalizes routing under partial observability through two setups: Cold-Start, which restricts learning to sparse, static logs, and Voyager, which models continuous data streams where routing decisions shape future training distributions. Evaluating 13 routing baselines, we find that most routers over-emphasize accuracy while ignoring latency under partial observability. Code and benchmark data are publicly available.

cs.AI↗
arXiv2026

Jianmu-Q: Bridging Quantum Field Theory and Quantum Hardware with AI Agents

Quantum computing offers a promising approach to non-perturbative simulations of quantum field theories, particularly for real-time dynamics and systems affected by the sign problem. Yet translating continuum field theories into quantum computations remains a challenging interdisciplinary task. Here we introduce \texttt{Jianmu-Q}, a multi-agent framework that transforms natural-language descriptions of quantum field theories into quantum simulation workflows. Specialized agents handle theoretical formulation, lattice discretization, qubit mapping, and circuit design, while physical consistency checks and independent numerical calculations provide validation throughout the workflow. We benchmark the framework on three field theories of increasing complexity: the $(1+1)$-dimensional Nambu--Jona-Lasinio and Schwinger models, and a $(2+1)$-dimensional $\mathbb{Z}_2$ gauge theory. Across these systems, noiseless quantum-circuit simulations converge toward independent exact-diagonalization and tensor-network results, while selected circuits are executed on IBM quantum hardware, demonstrating an end-to-end, validation-guided approach to quantum simulation of field theories.

hep-lat↗
arXiv2026

Learning Quantum Operator Dynamics from Short-Time Data

Real-time dynamics of quantum observables provide direct access to excitation spectra and correlation functions in quantum many-body systems, but currently available quantum devices are limited to short evolution times due to decoherence. We propose a neural ordinary differential equation (Neural ODE) framework with a physics-driven neural network to reconstruct long-time operator dynamics from short-time quantum data. By expanding observables in the Pauli basis and exploiting locality and symmetry constraints, the operator evolution is reduced to a tractable set of coefficients whose dynamics are learned from the measurement data. Applied to the Ising spin chain with mixed fields, the operator learning method accurately extrapolates its long-time correlation function and resolves its excitation spectrum. We validate the framework by collecting short-time data from IBM's superconducting quantum processor and successfully reconstruct the Fourier spectrum of an $N=8$ Ising spin chain with 5 times enhanced precision. Our results demonstrate a scalable and data-efficient strategy for extracting dynamical and spectral information from practical quantum hardware.

quant-ph↗
arXiv2026

When Interfaces Speak: Data-Aware Generative UI Harness for Active Interaction

Most human-agent interaction today remains text-based. Natural language can impose cognitive overload, ambiguity, information chaos, and slow input for complex tasks; ephemeral generative UIs can present structured information and guide users toward task completion. We propose GenUI-Harness, a multi-agent harness pairing a Tool Agent for information retrieval and task execution with a GUI Coder Agent that identifies ambiguities and generates front-end code for structured interfaces. Training the coder with reinforcement learning is challenging: verifiable rewards for interactive UI generation require costly execution, while LLM-as-a-Judge rewards are prone to reward hacking. We address the first challenge with Dynamic UX, a lightweight package for dynamic interaction and reward collection in a single sandbox, and the second with Reward Auditor, a meta-reward mechanism that monitors reward distributions and distills diagnostic patterns into a shared rubric and scoring specification. We introduce UI-TAU Bench, a benchmark for active human-agent interaction through generated UI code, built on 10 real-world domain databases constructed from public data sources and based on Tau-Bench tool-use settings, with Lite (300 tasks) and Full (1,000 tasks) splits. GenUI-Harness achieves an average Pass@3 gain of 4.48 percentage points over smolagents on Lite. Training with GenUI-Harness improves a 4B backbone from 9.33% to 58.00% Pass@3, outperforming larger frontier models such as Claude Opus 5 (46.67%). GenUI-Harness also remains robust on ambiguous and non-ambiguous queries. In a reviewer survey comparing communication channels, generated UIs reduce average dialogue rounds from 3.4 to 1.2. These results show that data-aware generative interfaces can support effective task completion and reduce dialogue rounds in evaluated database-backed workflows.

cs.AI↗
arXiv2026

Cleave: Scaling Tensor Program Optimization via Decoupled Algebraic Search and Operator Scheduling

Optimized kernels such as FlashAttention and FlashDecoding are crucial for accelerating today's large models. Most of them are handwritten by experts because existing ML compilers cannot match their efficiency. Producing such kernels requires fusing computations with multiple reductions, which requires both algebraic transformation of the computation graph and operator scheduling of the transformed graph. Unfortunately, searching the two jointly yields a space too large to navigate. We propose Cleave, an ML compiler built on symbolic decoupling: Cleave discovers transformations by performing superoptimization on a graph with symbolic shapes, and then schedules each resulting graph on concrete shapes. Representing shapes as symbols makes equivalence checking cheap and lets a new Split operator, with a symbolic split count, parallelize along a reduction dimension. Cleave's scheduler fuses graphs with multiple reductions through iterative tiling and horizontal fusion. Evaluation on common LLM subgraphs shows that Cleave generates kernels up to 2.8x faster than the best baseline (1.6x on average) and reduces compilation time by 5.9x on average compared to Mirage. For dynamic workloads captured from production serving traces, Cleave compiles each operator once and achieves geometric mean speedups of 1.4x and 1.7x over FlashInfer's handwritten FA2 and FA3 backends. Cleave's code is available at: https://github.com/nyu-systems/cleave

cs.PL↗
arXiv2026

Generative Modeling of Stochastic Dynamics for Long-Time Evolution

Exact stochastic equations for non-equilibrium dynamics are rarely accessible. We show that the long-time evolution of stochastic dynamics can be predicted from configuration pairs at a fixed short time lag, without knowledge of the equation of motion. Generative diffusion models learn the finite-time transition kernel from these pairs, and iterating it propagates the dynamics far beyond the training lag. For two-dimensional Model B, the diffusive dynamics of a conserved order parameter, the learned kernels reproduce dynamic critical scaling and self-similar $t^{1/3}$ coarsening. Agreement with direct simulations persists on lattices twice the largest training size and for initial ensembles absent from training. For driven colloids in a periodic optical potential, ten minutes of measured trajectories suffice to predict the particle current and mean passage time over the next twenty minutes within experimental uncertainty. Short-time observations thus contain the information needed to predict emergent non-equilibrium dynamics at much longer times.

cond-mat.stat-mech↗
arXiv2026

ORCA: Evaluating LLMs on Data Science Code Translation

Data Science Code Translation (DSCT) is the process of converting code between data science libraries while preserving functional equivalence and enabling interoperability across data science ecosystems. While Large Language Models (LLMs) have demonstrated considerable progress in Data Science Code Generation (DSCG), their performance in DSCT remains insufficiently studied. To address this gap, we introduce ORCA, a comprehensive benchmark with two complementary settings: ORCA-MAIN, which comprises 1,600 carefully curated grounding-level tasks across 3 representative domains: Data Querying, Data Manipulation, and Deep Learning; and ORCA-PROJECT, which contains 200 translation tasks over complete data science projects across 7 data science task types. Each task is accompanied by annotated reference translations and test cases for validating functional equivalence. We further incorporate a multi-stage quality verification process that thoroughly verifies task correctness and test case robustness. Experimental results demonstrate challenges in DSCT, with even frontier LLMs showing limited performance. Specifically, Claude-Opus-4.6 achieves a success rate of 56.92% on ORCA-MAIN and 33.67% on ORCA-PROJECT, indicating considerable room for improvement in DSCT. We also observe a clear directional preference in DSCT, where translation is consistently easier when the source code expresses the task through more explicit, fine-grained operations. Motivated by this, we propose an intent-augmented method, in which the model first infers source-code intent and then uses it as additional context for translation, achieving average absolute success-rate gains of 4.80% and 5.33% on ORCA-MAIN and ORCA-PROJECT, respectively.

cs.AI↗
arXiv2026

Aspen: Making Leaderless BFT Fast Paths Practical with Synchronized Clocks

No Byzantine Fault Tolerant (BFT) protocol can commit a request in less than the single round trip it takes for clients to reach the replicas and hear back. Some protocols approach this bound with a leaderless, speculative fast path, where clients broadcast requests directly to replicas and commit in two message delays ($2Δ$). However, such a fast path is extremely fragile: when clients submit requests concurrently, replicas receive them in different orders and end up in different states, forcing the protocol into costly recovery. This paper presents Aspen, a leaderless speculative BFT protocol that can maintain its fast path even while clients submit requests concurrently, at a near-optimal latency of $2Δ+ ε$. The $ε$ term is a short waiting delay that Aspen adds to give requests a best-effort, tentative initial order using loosely synchronized clocks and network delay estimates. Since real networks are not perfectly predictable, Aspen also tolerates replicas diverging: it adds extra replicas ($n = 3f + 2p + 1$) so that the fast path survives up to $p$ diverged replicas, a lightweight alignment subprotocol that returns those replicas to the fast path in the background, and a repair subprotocol that restores agreement when too many diverge. In experiments with geo-distributed replicas, Aspen reduces the latency of requests by $1.1\times$--$5.5\times$ compared to state-of-the-art BFT protocols, while maintaining throughput matching or exceeding throughput-optimized designs.

cs.DC↗
arXiv2026

Slow Light Augmented Unbalanced Interferometry for Extreme Enhancement in Sensitivity of Measuring Frequency Shift in a Laser

We demonstrate a slow-light augmented unbalanced Mach-Zehnder interferometer (MZI) which can be used to enhance very significantly the sensitivity of measuring the frequency shift in a laser. The factor of enhancement depends on the group index of the slow-light medium, the degree of imbalance between the physical lengths of the two arms of the MZI, and the spectral width of the laser. For a laser with a quantum noise limited spectral width, the group index has to be larger than the finesse of the laser cavity in order to achieve enhancement in measurement sensitivity. For the reported results, slow-light effect is produced by employing electro-magnetically induced transparency via coherent population trapping in a paraffin coated vapor cell of Rb atoms, with a maximum group index of ~2170. The maximum enhancement factor realized is ~70. This differs from the theoretically expected value of ~183 by a factor of ~2.6. This discrepancy can be attributed to effects of unidentified excess noise, and the fact that the test laser spectral width may not the quantum noise limited. Much larger values can potentially be obtained by modifying the apparatus, and using cold atoms for producing the slow-light effect. The sensitivity of any sensor that relies on measuring the frequency shift of a laser can be enhanced substantially using this technique. These include, but are not limited to, gyroscopes and accelerometers based on a conventional ring laser or a superluminal ring laser, and detectors for virialized ultra-light field dark matter.

quant-ph↗
arXiv2026

CS-Guard: Benchmarking LLM Guardrails for Code Generation Security

Large language models (LLMs) have been ex- ploited to generate malware, but the effective- ness of guardrails for code generation secu- rity remains unclear. We introduce CS-Guard, the first benchmark to systematically evalu- ate guardrails for code generation security. It covers 1) text-to-code generation with 1000 high-quality malware-generation prompts, 7 jailbreak attacks, and a novel fictional scenario attack (FSA) that embeds malicious intent in a legitimate fictional software-development sce- nario; and 2) code-to-code generation with 331 code prompts spanning code infilling, code completion, and code translation. We empiri- cally evaluate 9 guardrails across seven LLMs. We find that current guardrails perform poorly against malicious code-generation re- quests: for text-to-code, the average attack success rate (ASR) after jailbreaks reaches about 50% for many guardrails; for code-to- code, average ASR approaches 100% on base LLMs and remains high across many guardrails (14.4% to nearly 100%). Our FSA also achieves ASR close to 100% across many guardrails, raising major reliability concerns for real-world software development. To sup- port future research, CS-Guard uses a modular three-layer guardrail taxonomy that lets devel- opers register guardrails for evaluation. We release the benchmark and data to enable fur- ther community evaluation.

cs.CR↗
arXiv2026

MUX-USCT: A Noise-Robust Neural Network for Ultrasound Computed Tomography

Deep neural networks (DNNs) have shown strong potential for ultrasound computed tomography (USCT) reconstruction in ideal noise-free environments, yet existing DNNs are vulnerable to the noisy conditions in clinical practice, as they equally treat inputs that suffer mild, moderate, or severe noise. More challenging, the distributions of noise shift along with the environment, indicating the less effectiveness of noise-aware training, which injects a specific noise distribution into the training data. We rethink these challenges and observe that the DNN models can become more robust to noise if we know the noise sources and filter them out. This filtering operation is very alike the Multiplexers (or MUX), a fundamental combinational circuit in digital logic design. However, the challenge here is that noise can happen randomly during inference; as a result, the manually predefined MUX cannot work. To address these challenges, we propose MUX-USCT, a novel encoder-decoder DNN architecture that encodes the known acoustic acquisition geometry with an "adaptive MUX" that can automatically identify and filter noise, where the attention mechanism is applied in reconstructing the speed-of-sound map. On the OpenPros benchmark, MUX-USCT reaches 6.88 m/s MAE with 17% fewer parameters than the leading baseline with 7.65 m/s of MAE. Under simulated clinical noise, it remains stable across diverse degradation types that cause geometry-agnostic baselines to fail. Results show that the attention distributions in MUX-USCT provide interpretable indicators of the signal quality between pairs of transducers.

eess.IV↗
arXiv2026

Silicon-compatible ideal antiferroelectricity with large digital electromechanical responses enabled by thermal-strain domain engineering

Antiferroelectrics exhibit reversible antipolar-polar transformations, offering a compelling platform for multiple functionalities in modern nanoelectronics, yet deterministic control of antiferroelectric domains and switching pathways remain elusive. Moreover, their integration with ubiquitous silicon-based electronic devices has been limited by the structural and chemical incompatibilities of conventional oxide platforms. Here, we convert the conventional drawback of thermal mismatch into a functional advantage and realize ideal antiferroelectricity in epitaxial PbZrO3 thin films on silicon through thermal tensile-strain engineering, a strain regime unattainable on conventional perovskite substrates. Combined theoretical and experimental studies show that tensile strain stabilizes the (004)o domain, enabling a direct one-step switching, whereas compressive-strain-stabilized (240)o domains switch through intermediate ferrielectric states. The resulting films exhibit near-zero remanent polarization, square double hysteresis, nanosecond switching (~75ns), large reversible electrostrain (~0.6%) and robust operation windows. These findings provide key insights into domain-engineered ideal antiferroelectricity on silicon, opening a viable route toward high-performance antiferroelectric nano-electronic devices.

cond-mat.mtrl-sci↗
arXiv2026

TRL-Bench: Standardizing Cross-Paradigm Representation-Level Evaluation of Tabular Encoders

Tabular encoders are usually evaluated inside task-specific end-to-end pipelines, so models from different training paradigms are difficult to compare directly even when they operate on similar tabular signals. We introduce TRL-Bench, a multi-granular tabular representation learning (TRL) benchmark that standardizes cross-paradigm representation-level evaluation: each encoder exports row-, column-, or table embeddings through its supported wrapper, and shared lightweight heads probe them across three suites: TRL-CTbench (column/table), TRL-Rbench (row), and TRL-DLTE (compositional Data-Lake Table Enrichment spanning all three granularities). To support this standardized setting, we release curated benchmark assets and task reformulations, including 50 OpenML tables with 123 verified targets, 16 row-pair linkage rewrites, and a 47,772-table DLTE lake derived from 1,379 parent tables. Across 20 models and 16 tasks, TRL-Bench shows that once downstream conditions are standardized, encoder quality is capability-specific rather than captured by a single leaderboard. In TRL-CTbench, generic text encoders often lead on tasks with strong surface-text signal, while tabular specialists win where their pretraining objective aligns with the task. In TRL-Rbench, within-table prediction and cross-table linkage favor different training regimes, with atomic linkage performance correlating strongly with the row-matching stage of DLTE pipelines. In TRL-DLTE, the strongest pipelines combine capability-matched specialists rather than reuse a single encoder, and top end-to-end quality depends on non-additive compositional fit rather than per-stage marginal rank alone. TRL-Bench provides a common protocol for measuring reusable signal in exported tabular representations under shared downstream conditions. Code and data: https://github.com/LOGO-CUHKSZ/TRL-Bench

cs.AI↗
arXiv2026

Impact of spontaneous emission on spin-squeezed quantum sensors

The echo squeezing protocols (ESPs) are techniques that amplify the phase shift in a quantum sensor with one-axis-twist squeezing (OATS). For atomic sensors, spontaneous emission (SE) in the OATS operations is an important imperfection, which, however, is prohibitively difficult to study. SE can transfer atoms to all the Zeeman substates, making the number of relevant collective states excessively large. In this paper, we focus on a relatively simple isotope, namely Sr-88, which only has one Zeeman substate in each of the two ground states. Nevertheless, studying the effect of SE is still challenging because SE will populate all collective states, either symmetric or asymmetric, thus putting the ensemble into a mixed state. Based on analytical derivation and numerical simulations, we conclude that the GESP is more resistant to SE than the conventional echo squeezing protocol (CESP). This is another advantage of the GESP that has not been realized formerly. We also find that for the GESP employing Sr-88, the SE-induced reduction in the signal contrast is the same as that for a Ramsey protocol without squeezing and the quantum noise is suppressed by SE. This is an unexpectedly favorable result. The SE-induced suppression of quantum noise constitutes a previously unrecognized effect that challenges both earlier conclusions and intuitive expectations.

quant-ph↗
arXiv2026

Modified One-Axis-Twist Squeezing for Deterministic Orientation Control of Schrödinger Cat States with Unknown Atom-Number Parity

The Schrödinger cat state protocol (SCSP) makes use of a Schrödinger cat (SC) state generated with one-axis-twist squeezing (OATS) to enhance the sensitivity. In principle, the SCSP can magnify the phase shift by a factor N, the number of atoms under interrogation, accompanied by a quantum noise amplification by a factor of root-N, enabling an atomic sensor to reach the Heisenberg limit. However, the current SCSP can reach this benchmark only if the parity of N is known, which is almost impossible for a large ensemble. The reason is the orientation of the SC state generated with the conventional OATS depends on the parity of N. In this paper, we propose a modified OATS (MOATS) operation that generates the SC state aligned in a certain direction regardless of the parity of N.

quant-ph↗