Searcharxiv⌕ Search

arXiv subjects

Wentao Wang

Publications and source records attributed to Wentao Wang.

At least 19 recordsLinked to original sources

Shared Phase and Retention Control for Efficient Adaptive Spectral Recurrence

Fixed-state spectral recurrences offer constant-memory inference and strong long-range retention through heterogeneous complex modes, but their static transitions cannot selectively filter distractors or adapt to evolving dynamics. Existing input-dependent recurrences introduce selectivity, yet real-valued gates lack adaptive rotation, while fine-grained complex transitions incur substantial projection and data-movement overhead and can expose persistent modes to cumulative drift. To address these limitations, we introduce Shared Phase and Retention Control for Efficient Adaptive Spectral Recurrence (SPARC), which modulates a layer's heterogeneous complex memory through two input-dependent scalars. Each mode retains its learned baseline timescale, frequency, and write response, while a retention-coupled clock damps phase corrections in persistent modes. The same shared signals regulate incoming content through modal gates and retention-dependent normalization. The resulting diagonal affine recurrence supports parallel associative scans for full-sequence backpropagation through time (BPTT) and exact structured Real-Time Recurrent Learning (RTRL), with sensitivity storage and updates linear in the recurrent layer's parameter count. We evaluate SPARC on partially observable continuous control, POPGym memory benchmarks, sequence classification, and language modeling. SPARC places among the top two means across all ten reinforcement-learning tasks. Compared with RG-LRU, it improves permuted-MNIST and sequential-CIFAR-10 accuracy by $10.13$ and $10.66$ percentage points, respectively, reduces held-out language-model perplexity by $2.00\%$, and accelerates mixer forward--backward computation by up to $1.69\times$ at width $D=2048$. Code is available at https://github.com/Botwwt/sparc.

cs.LG↗

Single-spike superradiant free-electron lasers

Free-electron lasers (FELs) deliver high-brilliance radiation from the infrared to X-rays, enabling atomic-scale studies on femtosecond-to-attosecond timescales. Yet conventional facilities rely on large, costly radio-frequency accelerators. Laser wakefield accelerators (LWFAs), combining high accelerating gradients with ultra-high peak currents, offer a compact alternative and open up new prospects for advanced FEL operation. Here, we demonstrate a tabletop superradiant FEL driven by single-spike, pre-bunched electron beams from a LWFA. By manipulating the energy chirp during acceleration and compressing the bunch to sub-wavelength duration, we obtain coherent radiation extending into the vacuum-ultraviolet (VUV) regime within a remarkably compact 3-metre beamline--from gas target to radiator. Under optimized conditions, the measured photon yield reaches 2.8x10^10 per pulse after transmission through a 500-um slit, with a reproducibility of approximately 70%. This approach uniquely exploits the energy chirp and high-current characteristics inherent to LWFA beams without recourse to exponential gain, substantially minimizing device length while obviating the need for external synchronization and offering enhanced tolerance to jitter--thereby establishing a practical route toward efficient, compact FELs suitable for applications in academic, industrial, and clinical settings.

physics.acc-ph↗

High-quality electron beam generation from laser wakefield accelerators for driving compact free electron lasers

Despite the successful demonstration of compact free electron lasers (FELs) driven by laser wakefield accelerators (LWFAs), the pursuit of further enhancements in high-gain compact FELs presents a challenge due to the limitations in electron beam quality. In this work, we pinpoint the pivotal physics and optimization strategies for high-quality single-stage LWFAs that are crucial for high-gain FELs. We have delved into the synergistic injection mechanism, where the self-evolution injection threshold is far from reached at the injection position, with both the shock front and self-evolution of the laser playing a role in the injection process. A thorough discussion has been provided on the beam-quality degradation and optimization strategies, in terms of global (slice) energy spread and projected (slice) emittance. With the goal of achieving high-gain FELs driven by LWFAs, we have also explored the synthesis quality of the electron beam to determine an optimized power gain length. A comprehensive start-to-end simulation has been conducted, demonstrating the effectiveness of compact FELs powered by these high-quality electron beams. The resulting radiation reaches the saturation regime after a 4.5-meter-long undulator, with an energy of 17.4 μJ and a power of 6.0 GW at a wavelength of 23.9 nm. This proposed scheme offers not only a framework for optimizing beam quality in LWFAs, but also a promising path for future compact LWFA-driven FELs to achieve saturated regimes, opening up new possibilities for widespread applications.

physics.plasm-ph↗

EvaOmni: An Evidence-Grounded Benchmark for Audio-Visual Reasoning

Omni LLMs can naturally locate key audio and visual evidence and integrate information across modalities for audiovisual reasoning. Existing audio-visual reasoning benchmarks that mainly rely on multiple-choice questions usually score models solely based on answer correctness, making it difficult to assess whether a model truly forms a reasoning chain from evidence to the correct answer. In addition, textual or unimodal shortcuts may allow models to answer correctly without analyzing the necessary audio-visual evidence, further weakening the reliability of evaluation benchmarks. To address these issues, we introduce EvaOmni, an evidence-grounded benchmark containing 392 videos and 1,027 purely human-annotated question-answer pairs, covering 15 capabilities and three levels: retrieval and localization, basic understanding, and complex reasoning. We design a series of rigorous annotation procedures in which experts annotate audio and visual evidence and construct video-based distractors; humans review whether questions meet strict inspection criteria, and finally models filter textual and unimodal shortcuts. Beyond high-quality question-answer pairs, we require models to simultaneously return the answer and supporting evidence intervals in order to evaluate the complete reasoning chain. We propose the Coverage Grounded Score (C.G.) to measure how much annotated evidence a model identifies, and the Bottleneck Grounded Score (B.G.) to penalize poor temporal localization in either modality; both are incorporated into answer correctness. Across 7 open-weight and 17 proprietary Omni LLMs, Gemini3.7-Flash achieved the highest overall accuracy (79.4%), C.G. (70.8%), and B.G. (35.1%). At Level 3, it reached 79.1% accuracy, but its B.G. was only 12.3%. These results show that even the most advanced models still have room for development in grounded audio-visual reasoning.

cs.CV↗

Planning-Coding Separation in LLM-Based Code Generation: An Empirical Study

Planning has been identified as an important factor for improved code generation with Large Language Models (LLMs). However, the interaction between planning capability and coding capability remains largely underexplored, especially when the two capabilities are from different LLMs. In this paper, we conduct the first comprehensive empirical study of planning-coding interaction in LLM-based code generation, where planning and coding are assigned to different LLMs (i.e., planning-coding separation). Specifically, we explore different planner-coder combinations to investigate the effectiveness for coders to utilize planning information from various planners and the impact of further training of coders on planner-coder combinations. Our results demonstrate that almost all planner-coder combinations are as competitive as using the same LLM for planning and coding, and the combination of a stronger planner and a stronger coder typically achieves improved effectiveness. Furthermore, using a large model as the planner paired with a much smaller model as the coder can sometimes be as competitive as using the large model as both the planner and the coder, and further training of the coder can typically boost various planner-coder combinations. These findings provide empirical evidence that planning and coding can be separated to a considerable extent in LLM-based code generation, which may offer a more flexible way to construct code-generation systems.

cs.AI↗

Rate-Distortion-Perception Tradeoff for Vector Gaussian Sources with Limited Common Randomness: Exact Characterization and Approximate Universality

The distortion-rate-common randomness-perception (DRCP) theory characterizes how limited common randomness (CR) improves reconstruction fidelity under perceptual constraints, thereby providing an information-theoretic foundation for generative compression. In this paper, we characterize the DRCP tradeoff for vector Gaussian sources with limited CR under the squared-error distortion measure and the squared Wasserstein-2 perception measure. Specifically, we first establish a generalized reverse waterfilling characterization of globally optimal allocations. Second, we derive an analytical upper bound on the DRCP function through sequential greedy allocation of rate and CR. For fixed rate and CR budgets, the resulting representation is independent of the perception threshold. Finally, we establish its approximate universality through uniform performance guarantees: the achievable distortion exceeds the optimal distortion by at most $2ρ^*\approx0.2897$ times the total source variance, whereas the optimal-to-greedy distortion ratio is at least $η^*\approx0.7517$.

cs.IT↗

LinearARD: Linear-Memory Attention Distillation for RoPE Restoration

The extension of context windows in Large Language Models is typically facilitated by scaling positional encodings followed by Continued Pre-Training (CPT). While effective, this paradigm is notoriously data-hungry and computationally expensive, requiring massive long-text corpora to recalibrate the model to the shifted positional distribution. We propose LinearARD, a restoration distillation method that restores Rotary Position Embedding (RoPE)-scaled students through attention-structure consistency with a frozen native-RoPE teacher. Rather than next-token prediction or opaque hidden-state matching, LinearARD aligns row-wise distributions of dense $Q/Q$, $K/K$, and $V/V$ self-relation matrices from the final attention layer to directly supervise attention dynamics. To remove the quadratic memory bottleneck of $n \times n$ relation maps, we introduce a linear-activation-memory kernel that stores only per-token log-sum-exp statistics and recomputes logits in the backward pass to obtain exact Kullback-Leibler divergence gradients. Across LLaMA2-7B, LLaMA3-8B, and Mistral-7B-v0.1 extended to 32K context, LinearARD recovers 93.1%/94.2%/94.3% of native short-context performance and achieves strong long-context robustness on RULER using only 4.25M tokens---amounting to just 1.6% (a $\sim$60$\times$ reduction) of the 256M-token budget required by state-of-the-art baselines. Under this severely constrained budget, CPT and LongReD remain near zero on RULER, demonstrating that relation-level restoration provides a substantially better efficiency-quality tradeoff. Our code is available at https://github.com/gracefulning/LinearARD.

cs.CL↗

Agents Don't Just Agree, They Remember: Benchmarking Persistent Sycophancy in Self-Improving Personal Agents

Self-improving personal agents write profiles, memories, and reusable skills that carry over from one chat to the next. Prior works ask whether a model agrees with the user in its next reply. Yet these agents can also write down a claim the user made, and a later chat can read that note back as trusted context. We call this persistent sycophancy. We introduce the Personal Agent Sycophancy Benchmark, PASB, with 1,600 tasks run on two agent frameworks, Hermes-Agent and OpenClaw, across sixteen models that span scales and generations. Each task separates a first chat containing the claim from a follow-up chat in a fresh session, so the first chat can reach the follow-up only through a note the agent chose to write. Downstream failure, the rate at which follow-up answers still rely on the claim, reaches 68.1% when the agent saved the claim against 35.6% when it did not. Later generations fail less often, yet a saved note still more than doubles their failure rate. In 38.8% of runs the saved note states the claim with more authority than the user gave it, and a saved claim still shapes answers in a different domain. Among four mitigations, editing the stored note works better than correcting the agent in conversation, cutting same-domain downstream failure from 36.3% to 30.3% for Hermes-Agent and from 37.9% to 31.6% for OpenClaw. A strong instruction against saving brings failure back to the level of runs that saved nothing. Self-improving agents need controls on what they write to memory as well as on what they say. Our benchmark is available at https://github.com/henrymao2004/agent-sycophancy.

cs.AI↗

MC-TRCM: Observation-Aware Recursive Fusion for Incomplete Mobile and Wearable Mental-Health Feature Views

Public mobile and wearable mental-health datasets often provide summarized feature tables rather than synchronized raw sensor streams. In these releases, each anchor corresponds to a survey or label time and may combine phone or wearable summaries, prior symptom scores, demographics, clinical variables, and source-availability indicators. We propose the Modality-Conditioned Temporal Recursive Context Model (MC-TRCM), which preserves each feature source as a separate token and incorporates missingness as part of the input context. Observed sources are encoded with values and missingness summaries, absent sources use learned absence tokens, dataset and task embeddings condition fusion, and a recursive prediction head refines each output over validation-selected steps. We evaluated MC-TRCM on six predefined endpoints from DepreST-CAT and Prediction of Severity Change-Depression (PSYCHE-D) using participant-level splits and validation-only model selection. MC-TRCM achieved the lowest mean absolute error on DepreST-CAT Patient Health Questionnaire-9 (PHQ-9) and Generalized Anxiety Disorder-7 (GAD-7) severity, improving over the best tabular reference by 0.181 and 0.217 scale points. Classification endpoints showed task-dependent behavior: MC-TRCM matched the best rounded GAD-7 category balanced accuracy, was numerically highest by 0.002 balanced-accuracy points on PSYCHE-D multiclass prediction, and remained close to the strongest references on PHQ-9 category and PSYCHE-D binary prediction. Ablations support Feature-wise Linear Modulation, absence tokens, missingness projections, and recursive refinement, while calibration and feature-source controls characterize endpoint behavior. Our code is available at https://github.com/Botwwt/MC-TRCM.

cs.LG↗

ActGuard: Pre-execution Action Auditing against Indirect Prompt Injection in LLM Agents

Large language model (LLM) agents interact with external environments through tool invocation, but tool outputs can also expose them to indirect prompt injection (IPI) attacks. Existing defenses mainly rely on prompt hardening, content filtering, pre-generated plans, or permission constraints. These approaches often struggle with complex tasks or over-sanitize external content, making it difficult to balance security and utility. The key challenge is therefore to preserve execution flexibility while precisely identifying and removing the malicious content that actually induces unsafe actions. To address this challenge, we propose ActGuard, a pre-execution action auditing framework. Rather than judging whether external content is inherently suspicious, ActGuard assesses whether it causes the current action to deviate from a locally reasonable expectation. At each step, ActGuard predicts the tools likely to be used by the upcoming action and constructs a local tool prior without constraining the execution trajectory. Before execution, it compares the candidate action against this prior and performs tool-level contrastive analysis and parameter-level evidence localization to identify deviations in tool selection and action parameters. A verifier then examines the localized evidence, masks only spans confirmed as malicious, and regenerates the action from the sanitized context. This design preserves legitimate planning flexibility while minimizing information loss from indiscriminate filtering. We evaluate ActGuard on challenging benchmarks for tool-using agents. Results show that ActGuard reduces attack success rates to a level comparable to state-of-the-art defenses while maintaining task utility close to the no-attack setting, achieving a favorable security-utility trade-off. Our code is publicly available at: https://github.com/binzhwang/ActGuard.

cs.CR↗

GT-Space: Enhancing Heterogeneous Collaborative Perception with Ground Truth Feature Space

In autonomous driving, multi-agent collaborative perception enhances sensing capabilities by enabling agents to share perceptual data. A key challenge lies in handling {\em heterogeneous} features from agents equipped with different sensing modalities or model architectures, which complicates data fusion. Existing approaches often require retraining encoders or designing interpreter modules for pairwise feature alignment, but these solutions are not scalable in practice. To address this, we propose {\em GT-Space}, a flexible and scalable collaborative perception framework for heterogeneous agents. GT-Space constructs a common feature space from ground-truth labels, providing a unified reference for feature alignment. With this shared space, agents only need a single adapter module to project their features, eliminating the need for pairwise interactions with other agents. Furthermore, we design a fusion network trained with contrastive losses across diverse modality combinations. Extensive experiments on simulation datasets (OPV2V and V2XSet) and a real-world dataset (RCooper) demonstrate that GT-Space consistently outperforms baselines in detection accuracy while delivering robust performance. Our code will be released at https://github.com/KingScar/GT-Space.

cs.LG↗

Crossing-Free Probabilistic K-Line Forecasts Without Retraining

Probabilistic K-line forecasting describes uncertainty in four complementary prices, namely open--high--low--close (OHLC). However, it introduces two consistency problems: quantile crossing and K-line crossing. Quantile crossing occurs when a higher-quantile forecast falls below a lower-quantile forecast, while K-line crossing occurs when the forecast low exceeds the open or close, or the forecast high falls below the open or close. Existing solutions generally address only one problem through output reordering, specialized architectures, or penalized training objectives. We propose K-line--Quantile Sequential Projection (KQSP), a parameter-free and training-free reconciliation method applicable to forecasts produced by any model. Compared with other crossing solutions, KQSP preserves predictive accuracy while producing substantially smaller corrections to the original forecasts. To mitigate model bias, we evaluate KQSP using various models, including pretrained foundation models. KQSP reduces both quantile and K-line crossing rates to zero for all test data undertaken. These results show that probabilistic K-line consistency can be enforced independently of forecast generation and without retraining.

stat.ML↗

Generation of bright quantum high-order harmonic driven by combined coherent and bright squeezed vacuum light

Attosecond quantum light, formed by the superposition of high-order harmonics driven by intense quantum light, opens new routes to probe quantum-mechanical correlations in matter. In this study, we have investigated the macroscopic propagation effects of quantum high-order harmonics generated by the combination of strong coherent and weak bright squeezed vacuum (BSV) lasers interacting with atomic gas. Our results reveal that the pressure-dependent intensity of harmonics arising from absorbing or emitting BSV photons differs from that of harmonics generated using only strong coherent pulses. Macroscopic propagation simulations indicate that the action phase of harmonics is perturbed by the weak BSV pulses. This perturbation modulates the phase mismatch of sub-cycle attosecond bursts and affects their quantum properties when the gas pressure varies. The ability to generate bright quantum high-order harmonics lays a foundation for the establishment and application of attosecond quantum spectroscopy.

quant-ph↗

Deep Learning for Electricity Price Forecasting: A Review of Day-Ahead, Intraday, and Balancing Electricity Markets

Electricity price forecasting (EPF) plays a critical role in power system operation and market decision making. While existing review studies have provided valuable insights into forecasting horizons, market mechanisms, and evaluation practices, the rapid adoption of deep learning has introduced increasingly diverse model architectures, output structures, and training objectives that remain insufficiently analyzed in depth. This paper presents a structured review of deep learning methods for EPF in day-ahead, intraday, and balancing markets. Specifically, We introduce a unified taxonomy that decomposes deep learning models into backbone, head, and loss components, providing a consistent evaluation perspective across studies. Using this framework, we analyze recent trends in deep learning components across markets. Our study highlights the shift toward probabilistic, microstructure-centric, and market-aware designs. We further identify key gaps in the literature, including limited attention to intraday and balancing markets and the need for market-specific modeling strategies, thereby helping to consolidate and advance existing review studies.

q-fin.CP↗

A Market-Rule-Informed Neural Network for Efficient Imbalance Electricity Price Forecasting

Accurate and efficient imbalance electricity price forecasting is critical for industrial energy trading systems, especially as battery assets and automated bidding pipelines increasingly participate in balancing markets. However, real-time forecasting is complicated by nonlinear market-rule-based price formation, heterogeneous input signals, and incomplete data availability caused by communication delays, publication lags, and measurement outages. This paper proposes a market-rule-informed neural forecasting framework that embeds imbalance price formation rules into the latent space of an expressive neural network. The proposed framework preserves raw signal information while exploiting transparent market-rule priors. We further analyze operational robustness by removing price-component information and characterize how forecasting performance scales with input length and forecasting horizon. Experimental results show that the proposed model achieves competitive forecasting performance with substantially fewer trainable parameters and shorter training time than generic deep learning baselines. Experimental results show that the proposed model achieves competitive forecasting performance with substantially fewer trainable parameters and shorter training time than generic deep learning baselines, demonstrating that market-rule priors and expressive neural networks should be jointly used for accurate and computationally sustainable forecasting in industrial energy trading applications. The implementation is publicly available at https://runyao-yu.github.io/MRINN/.

q-fin.CP↗

A Temporal Retrieval Method for Modulated Electron Bunches via Adaptive Kernel Reconstruction

Femtosecond electron beams with complex modulation play a crucial role in applications such as X-ray Free Electron Lasers (XFELs) and plasma wakefield accelerators. However, diagnostics for the electron beam current profile still face challenges with complex structure. In this letter, we propose a novel temporal retrieval algorithm for the coherent transition radiation (CTR) diagnostics of complex modulated electron beams. Starting from the time-frequency analysis of the electron bunch train, the algorithm separates and reconstructs the high- and low-frequency components. A temporal kernel was derived from the inverse sampling of the measured spectrum to construct the high-frequency component, while the low-frequency envelope was composed of several basis functions. Tested on the electron bunch trains from the complex multi-gaussian model and bunching-enhanced coherent harmonic generation, the algorithm successfully reconstructed the temporal signals and achieves better performance than the Kramers-Kronig method. This method is expected to crucial provide temporal evidence for potential electron beam modulation schemes, and will enable broad prospects for future applications.

physics.acc-ph↗

Enhancing SignSGD: Small-Batch Convergence Analysis and a Hybrid Switching Strategy

SignSGD compresses each stochastic gradient coordinate to a single bit, offering substantial memory and communication savings, but its 1-bit quantization removes magnitude information and is known to leave a generalization gap relative to well-tuned SGD. We revisit SignSGD from a 1-bit quantization and dithering perspective and contribute three improvements. First, we derive a small-batch convergence rate for SignSGD under unimodal symmetric gradient noise using a signal-to-noise weighted stationarity measure, removing the large-batch assumption of prior analyses. Second, we inject annealed Gaussian noise before the sign operator, which acts as a classical dithering mechanism and probabilistically restores magnitude information lost to hard thresholding. Third, we adapt the SWATS strategy to sign-based updates with a projection-based learning-rate calibration that smoothly transitions from SignSGD to SGD. Single-worker experiments on ResNet-18 isolate optimizer effects from communication aspects: pre-sign dithering surpasses Adam on CIFAR-100, and the calibrated switch reaches 92.18% test accuracy on CIFAR-10, outperforming both pure SGD 91.38% and pure SignSGD with momentum 90.82%.

cs.LG↗

RELIC: Evaluating Complex Reasoning via the Recognition of Languages In-Context

Large language models (LLMs) are increasingly used to solve complex tasks where they must retrieve and compose many pieces of in-context information in long reasoning chains. For many real-world tasks it is hard to accurately gauge how model performance and strategy change as task complexity grows. To evaluate models' complex reasoning capability in a scalable and verifiable way, we introduce RELIC (Recognition of Languages In-Context), a framework that evaluates an LLM's ability to decide whether a given string belongs to the context-free language (CFL) generated by a grammar presented in-context. CFL recognition allows us to modulate the intrinsic complexity of the problem by varying grammar size and string length and translate this asymptotic complexity into predictions for ideal LLM performance. We find that even the most advanced reasoning models perform poorly on RELIC, not only failing to appropriately scale their inference compute to keep pace with task difficulty, but even reducing the number of reasoning tokens they use as task complexity increases. We find that these decreases in compute accompany changes in reasoning strategy, as models move from identifying and implementing algorithmic solutions to guessing. For models whose full completions go uninspected, this manifests as ``quiet quitting'' on hard tasks.

cs.CL↗