arXiv subjects

Lei Gao

Publications and source records attributed to Lei Gao.

At least 19 recordsLinked to original sources

arXiv2026

Non-Hermitian Light Beams

Non-Hermitian systems provide remarkable features actively studied in modern photonics. Non-Hermiticity is often related to the properties of open structures and devices, which involve lossy and gain channels. Here, we reveal that electromagnetic fields themselves can be designed in a non-Hermitian way, proposing an additional degree of freedom in non-Hermitian physics. Within classical electromagnetic wave theory, we show that the non-Hermitian behaviors may stem not only from the dispersive properties of the waves, but also from the peculiarities of the light spatial spectrum and local structure of the fields. For the nondiffracting light beams, we identify exceptional points detaching symmetric and asymmetric non-Hermitian phases related to the shape of the beam's intensity profile, the non-Hermiticity being well recognized only for non-paraxial beams. In the local structure of the electromagnetic fields and Poynting vector, we determine exceptional lines differentiating non-Hermitian phases stemming from the behavior of streamlines. We believe that the concept of non-Hermitian light beams will enrich our knowledge of the light-matter interaction.

physics.optics↗

arXiv2026

STAIR: Semantic-Temporal Automaton for Interpretable Reasoning in Temporal Question Answering

By leveraging large-scale pretraining, LLMs can interpret diverse temporal expressions and question formulations without task-specific training. However, existing prompt-based neuro-symbolic systems continue to rely on LLMs for both semantic interpretation and exact temporal inference. Consequently, discrete decisions regarding intervals, time anchors, and ordered states remain vulnerable to probabilistic errors and difficult to verify. We present STAIR, a \textbf{S}emantic-\textbf{T}emporal \textbf{A}utomaton for \textbf{I}nterpretable \textbf{R}easoning. STAIR separates semantic interpretation from precise temporal inference: an answer-free LLM adapter maps complex question formulations to normalized temporal intents, while a deterministic temporal automaton with finite control and guarded transitions executes the corresponding policies over canonicalized evidence. Following a rule-first design, STAIR resolves standard questions without invoking an LLM and applies semantic adaptation only when the rule path fails to produce an executable intent. This approach reduces free-form reasoning, making temporal decisions verifiable and interpretable. Specifically, guarded execution supports precise point-time containment and before/after selection, while semantic adaptation handles non-exact intervals and time-anchored queries. Across the TimeQA-Easy, TimeQA-Hard, TempReason-L2, and TempReason-L3 datasets, STAIR consistently outperforms strong baselines in the TQA task using matched model settings, achieving average F1 improvements of 16.57\% and 3.10\% when utilizing the Qwen2.5-7B and GPT-4o-mini models, respectively. Furthermore, ablations and diagnostic analyses demonstrate that STAIR excels at handling both boundary-sensitive and order-sensitive queries, while its guarded execution and semantic adaptation ensure precise point-time reasoning and inexact intervals, respectively.

cs.CL↗

arXiv2026

Bidirectional Context Self-Distillation for Reinforcement Learning of Skill-Based LLM Agents

External natural-language skills provide large language model (LLM) agents with reusable and editable guidance for solving complex tasks. Yet their effectiveness depends not only on skill quality, but also on whether the policy can translate the provided guidance into appropriate actions. However, methods specifically designed to improve this skill-utilization ability remain largely underexplored. In practice, skill-based agents are commonly trained with reinforcement learning objectives centered on task-level rewards, which offer limited supervision and struggle to capture subtle differences in how effectively the policy uses the provided skills. We propose BCSD (Bidirectional Context Self-Distillation), a framework that combines self-distillation with reinforcement learning to train LLM agents to use external skills more effectively. Unlike prior self-distillation methods that rely on a single privileged context, BCSD evaluates each trajectory from two complementary skill-context views. The augmented view introduces higher-level Meta-Skill guidance, while the reduced view prunes general guidance to highlight task-specific skills. Their complementary token-level signals are combined to rescale the RL advantage. Experiments on ALFWorld and WebShop demonstrate that BCSD achieves the strongest overall performance across model scales, enabling agents to utilize external skills more effectively. Ablation studies further verify the complementary contributions of the augmented and reduced context views. Code will be released to ensure full reproducibility.

cs.AI↗

arXiv2026

Agent Harness Distillation: Inference-Time Harness Extraction and Exploitation in Autonomous Multi-Agent Systems

Autonomous multi-agent systems (AMAS) built on large language models (LLMs), such as Hermes, increasingly rely on inference-time harnesses to coordinate reasoning and action. Constructing these harnesses requires substantial engineering effort and computational resources, as they are iteratively optimized over a combinatorial search space while co-evolving with the underlying LLM. Inference-time harnesses therefore constitute valuable intellectual property (IP). Although prior work has investigated IP leakage in static multi-agent systems with pre-configured architectures, it remains unclear whether similar risks arise in AMAS, where harness behavior emerges dynamically during inference. To address this gap, we introduce Agent Harness Distillation (AHD), a framework for studying the security risks arising from inference-time harness extraction in AMAS. We formalize harness extraction as a new security problem and develop an evaluation framework for quantifying such risks. AHD extracts inference-time harness capabilities from a target agent through black-box interactions and consists of two stages. In the pre-distillation stage, AHD infers inference-time harness behaviors from the responses of the target agent and constructs an initial harness. In the post-distillation stage, AHD iteratively refines the initial harness to align with the behavioral patterns of the target agent. Experiments on real-world AMAS across multiple backbone LLMs demonstrate the effectiveness of AHD and reveal substantial IP leakage risks. We further propose a deception-based defense that reduces harness extraction effectiveness while preserving the utility of the protected agent. Our findings uncover a previously underexplored security threat to AMAS.

cs.CR↗

arXiv2026

General and concise operator approach to the dyadic Green's function of layered media

Dyadic Green's function is an important tool of computational photonics, giving deeper insights into light-matter interaction. We present an operator approach to the derivation of the dyadic Green's function of a generic anisotropic planarly-layered medium for both electric and magnetic fields. The resulting Green's function is expressed through the evolution operators (a kind of transfer matrices) of the comprising layers and the surface impedance tensors, the singular term being naturally separated from other terms. The operator approach to the Green's function simplifies both the conceptual understanding of the problem and the subsequent practical applications, some of which are demonstrated here. The proposed approach can be easily generalized to the case of spherical and cylindrical layers, as well as bi-anisotropic layered media. The obtained results can be applied in nanophotonics engineering problems.

physics.optics↗

arXiv2026

Segment-Aligned Policy Optimization for Multi-Modal Reasoning

Existing reinforcement learning approaches for Large Language Models typically perform policy optimization at the granularity of individual tokens or entire response sequences. However, such formulations often misalign with the natural step-wise structure of reasoning processes, leading to suboptimal credit assignment and unstable training in multi-modal reasoning tasks. To bridge this gap, we propose Segment-Aligned Policy Optimization (SAPO), a novel reinforcement learning paradigm that treats coherent reasoning steps, rather than tokens or full sequences as fundamental units of policy update. SAPO introduces a step-wise Markov decision process abstraction over reasoning segments, accompanied by segment-level value estimation, advantage computation, and importance sampling mechanisms that are semantically aligned with reasoning boundaries. Experiments on representative reasoning benchmarks demonstrate that SAPO consistently outperforms token-level and sequence-level policy optimization methods, achieving significant accuracy improvements while exhibiting better training stability and value estimation consistency. Our work underscores the importance of aligning reinforcement learning updates with the intrinsic structure of reasoning, paving the way for more efficient and semantically grounded policy optimization in complex reasoning tasks. Codes and models will be released to ensure full reproducibility.

cs.AI↗

arXiv2026

A Gesture-Based Visual Learning Model for Acoustophoretic Interactions using a Swarm of AcoustoBots

AcoustoBots are mobile acoustophoretic robots capable of delivering mid-air haptics, directional audio, and acoustic levitation, but existing implementations rely on scripted commands and lack an intuitive interface for real-time human control. This work presents a gesture-based visual learning framework for contactless human-swarm interaction with a multimodal AcoustoBot platform. The system combines ESP32-CAM gesture capture, PhaseSpace motion tracking, centralized processing, and an OpenCLIP-based visual learning model (VLM) with linear probing to classify three hand gestures and map them to haptics, audio, and levitation modalities. Validation accuracy improved from about 67% with a small dataset to nearly 98% with the largest dataset. In integrated experiments with two AcoustoBots, the system achieved an overall gesture-to-modality switching accuracy of 87.8% across 90 trials, with an average end-to-end latency of 3.95 seconds. These results demonstrate the feasibility of using a vision-language-model-based gesture interface for multimodal human-swarm interaction. While the current system is limited by centralized processing, a static gesture set, and controlled-environment evaluation, it establishes a foundation for more expressive, scalable, and accessible swarm robotic interfaces.

cs.RO↗

arXiv2026

Vehicle-as-Prompt: A Unified Deep Reinforcement Learning Framework for Heterogeneous Fleet Vehicle Routing Problem

Unlike traditional homogeneous routing problems, the Heterogeneous Fleet Vehicle Routing Problem (HFVRP) involves heterogeneous fixed costs, variable travel costs, and capacity constraints, rendering solution quality highly sensitive to vehicle selection. Furthermore, real-world logistics applications often impose additional complex constraints, markedly increasing computational complexity. However, most existing Deep Reinforcement Learning (DRL)-based methods are restricted to homogeneous scenarios, leading to suboptimal performance when applied to HFVRP and its complex variants. To bridge this gap, we investigate HFVRP under complex constraints and develop a unified DRL framework capable of solving the problem across various variant settings. We introduce the Vehicle-as-Prompt (VaP) mechanism, which formulates the problem as a single-stage autoregressive decision process. Building on this, we propose VaP-CSMV, a framework featuring a cross-semantic encoder and a multi-view decoder that effectively addresses various problem variants and captures the complex mapping relationships between vehicle heterogeneity and customer node attributes. Extensive experimental results demonstrate that VaP-CSMV significantly outperforms existing state-of-the-art DRL-based neural solvers and achieves competitive solution quality compared to traditional heuristic solvers, while reducing inference time to mere seconds. Furthermore, the framework exhibits strong zero-shot generalization capabilities on large-scale and previously unseen problem variants, while ablation studies validate the vital contribution of each component.

cs.LG↗

arXiv2026

Nonlocal Optomechanics: Hybrid Anapole Opens a New Route to Optical Tweezing

Optical tweezers confine a particle in an intensity-defined potential well by engaging its local multipoles. In this picture, eliminating far-field scattering from the particle should cancel the optical force, as the multipole moments underpinning the conventional optomechanical response vanish. We show that certain resonant states, such as, e.g., the hybrid anapole state, enable qualitatively different optical manipulation, nonlocal by nature, where the optical force exhibits nontrivial spatial variations absent in conventional tweezing, establishing a new framework for manipulating resonant nanoparticles.

physics.optics↗

arXiv2026

Suppressed excitonic effects enable high mobility, high-yield photoconductivity in a two-dimensional polymer crystal with axial pyridine coordination

Two-dimensional polymers (2DPs) and their layer-stacked covalent organic frameworks (2D COFs) offer modular, atomically precise platforms for organic optoelectronics, yet their photoconductive responses remain fundamentally constrained by strong excitonic effects and localized charge transport. Here, we demonstrate that a diyne-linked 2DP crystal with axial pyridine coordination overcomes this limitation, enabling simultaneous efficient free-carrier generation and band-like transport. Introducing pyridine ligands that axially coordinate to Cu-porphyrin nodes transforms weak van der Waals stacking into a pyridine-bridged architecture with pronounced interlayer band dispersion and substantially reduced carrier effective masses. The resulting strong interlayer electronic coupling suppresses the exciton binding energy to well below thermal energy, such that optical excitation directly populates delocalized electronic states. Time-resolved terahertz spectroscopy reveals Drude-type photoconductivity with room-temperature mobilities approaching 500 cm^2 V^-1 s^-1 and a photon-to-free-carrier conversion ratio of ~0.4, yielding a photoconductive response that exceeds that of state-of-the-art organic and many inorganic photoactive materials. These results establish interlayer coordination as a powerful strategy for mitigating excitonic effects and accessing inorganic-like charge transport in organic 2D crystals, opening a pathway toward highly efficient photo-to-electricity conversion in organic-based systems.

cond-mat.mtrl-sci↗

arXiv2026

Infrastructure for Valuable, Tradable, and Verifiable Agent Memory

Every API token you spend is your accumulated wealth; once you can prove its value and the effort behind it, you can resell it. As autonomous agents repeatedly call models and tools, they accumulate memories that are your intellectual property. But today these memories remain private and non-transferable, as there is no way to validate their value. We argue that agent memory can serve as an economic commodity in the agent economy, if buyers can verify that it is authentic, effort-backed, and produced in a compatible execution context. To realize this idea, we propose clawgang, which binds memory to verifiable computational provenance, and meowtrade, a market layer for listing, transferring, and governing certified memory artifacts. Together, they transform one-shot API token spending into reusable and tradable assets, enabling timely memory transfer, reducing repeated exploration, and opening a memory trade market.

cs.CR↗

arXiv2026

Accurate and Efficient Multi-Channel Time Series Forecasting via Sparse Attention Mechanism

The task of multi-channel time series forecasting is ubiquitous in numerous fields such as finance, supply chain management, and energy planning. It is critical to effectively capture complex dynamic dependencies within and between channels for accurate predictions. However, traditional method paid few attentions on learning the interaction among channels. This paper proposes Linear-Network (Li-Net), a novel architecture designed for multi-channel time series forecasting that captures the linear and non-linear dependencies among channels. Li-Net dynamically compresses representations across sequence and channel dimensions, processes the information through a configurable non-linear module and subsequently reconstructs the forecasts. Moreover, Li-Net integrates a sparse Top-K Softmax attention mechanism within a multi-scale projection framework to address these challenges. A core innovation is its ability to seamlessly incorporate and fuse multi-modal embeddings, guiding the sparse attention process to focus on the most informative time steps and feature channels. Through the experiment results on multiple real-world benchmark datasets demonstrate that Li-Net achieves competitive performance compared to state-of-the-art baseline methods. Furthermore, Li-Net provides a superior balance between prediction accuracy and computational burden, exhibiting significantly lower memory usage and faster inference times. Detailed ablation studies and parameter sensitivity analyses validate the effectiveness of each key component in our proposed architecture. Keywords: Multivariate Time Series Forecasting, Sparse Attention Mechanism, Multimodal Information Fusion, Non-linear relationship

cs.AI↗

arXiv2026

Dynamic Dispersion Accumulation in Fiber Loops for Realizing Record-High Frequency Resolution or Ultra-Low Signal Sampling Rate in Dispersion-Based Photonics-Assisted Wideband Microwave Measurement Systems

Dispersion-based photonics-assisted microwave measurement systems provide immense potential for real-time analysis of wideband and dynamic signals. However, they face two critical challenges: a difficulty in achieving high frequency resolution over a wideband analysis bandwidth, and a reliance on large-bandwidth-and-high-sampling-rate oscilloscopes to capture the resulting ultra-narrow pulses. We introduce a dynamic dispersion accumulation technique to overcome these limitations. By circulating the optical signal in fiber loops containing a dispersion-compensating fiber, we achieve a high accumulated dispersion of -215700 ps/nm. This high dispersion relaxes the required chirp rate of the chirped optical signal, enabling two distinct advantages: When the analysis bandwidth is fixed, a lower chirp rate enables a longer temporal period, yielding a record-high frequency resolution of 27.9 MHz; When the temporal period is fixed, a lower chirp rate enables a smaller bandwidth, generating a wider pulse and thus relaxing pulse sampling requirements at the expense of analysis bandwidth. This sacrifice in analysis bandwidth can be compensated by a duty-cycle-enabling technique, which holds the potential to extend the analysis bandwidth beyond 100 GHz. This work breaks the performance and hardware limitations in dispersion-based systems, paving the way for high frequency resolution, wideband microwave measurement systems that are both real-time and cost-effective.

physics.optics↗

arXiv2026

Broadband tunable narrow-linewidth laser based on scattering-enhanced fiber covering E-S-C-L bands

This work demonstrates a broadband tunable narrow-linewidth laser based on scattering-enhanced fiber, covering the E-S-C-L wavelength bands from 1337.47 nm to 1631.39 nm, with a total tuning span of 293.92 nm. The laser employs two semiconductor optical amplifiers (SOAs) centered at 1420 nm and 1550 nm, which are connected into a single ring resonator via polarization multiplexing. Wavelength selection and tunability is realized using an ultra-broadband tunable filter based on a blazed grating. To suppress side longitude modes, an 18-meter-long femtosecond-laser-empowered random scattering fiber is utilized inside the cavity as a feedback medium, yielding an output linewidths between 1.54 kHz and 2.61 kHz. Benefited from the fast response of the galvanometer mirror and short relaxation time of SOAs, wavelength switching time is less than 1 ms under different tuning channels among the wavelength range of near 300 nm. The stable single-longitude-mode operation is maintained across the entire tuning range. The exceptionally broad tuning range and high spectral purity of the laser endow it with significant application potentials across a wide range of fields.

physics.optics↗

arXiv2025

Broadband tunable microwave photonic radar for simultaneous detection of human respiration, heartbeat, and speech with deep learning-based speech recognition

Multimodal vital sign monitoring and speech detection hold significant importance in medical health, public safety, and other fields. This study proposes a broadband tunable microwave photonic radar system that can simultaneously monitor respiration, heartbeat, and speech. The system works by generating broadband radar signals to detect subtle skin displacements caused by these physiological activities. It then utilizes phase variations in radar echo signals to extract and reconstruct the corresponding physiological signals. In order to enhance the processing capability for speech signals, a convolutional neural network with a dual-channel feature fusion model is incorporated, enabling high-precision speech recognition. In addition, the system's frequency-tunable characteristic allows it to flexibly switch frequency bands to adapt to different working environments, greatly improving its practicality and environmental adaptability. In concept-verification experiments, speech signals were reconstructed and recognized in the Ku, K, and Ka bands, achieving recognition accuracies of 97.20%, 98.07%, and 97.43%, respectively. The system's capability to detect multimodal vital signs was also thoroughly validated using a respiratory and heartbeat simulator. During a 20-second monitoring period, while accurately reconstructing speech, the maximum average error counts for respiratory and heartbeat monitoring were 0.39 and 0.87, respectively, proving its reliability and effectiveness in multimodal vital sign monitoring.

physics.optics↗

arXiv2025

Striking the Right Balance between Compute and Copy: Improving LLM Inferencing Under Speculative Decoding

With the skyrocketing costs of GPUs and their virtual instances in the cloud, there is a significant desire to use CPUs for large language model (LLM) inference. KV cache update, often implemented as allocation, copying, and in-place strided update for each generated token, incurs significant overhead. As the sequence length increases, the allocation and copy overheads dominate the performance. Alternate approaches may allocate large KV tensors upfront to enable in-place updates, but these matrices (with zero-padded rows) cause redundant computations. In this work, we propose a new KV cache allocation mechanism called Balancing Memory and Compute (BMC). BMC allocates, once every r iterations, KV tensors with r redundant rows, allowing in-place update without copy overhead for those iterations, but at the expense of a small amount of redundant computation. Second, we make an interesting observation that the extra rows allocated in the KV tensors and the resulting redundant computation can be repurposed for Speculative Decoding (SD) that improves token generation efficiency. Last, BMC represents a spectrum of design points with different values of r. To identify the best-performing design point(s), we derive a simple analytical model for BMC. The proposed BMC method achieves an average throughput acceleration of up to 3.2x over baseline HuggingFace (without SD). Importantly when we apply BMC with SD, it results in an additional speedup of up to 1.39x, over and above the speedup offered by SD. Further, BMC achieves a throughput acceleration of up to 1.36x and 2.29x over state-of-the-art inference servers vLLM and DeepSpeed, respectively. Although the BMC technique is evaluated extensively across different classes of CPUs (desktop and server class), we also evaluate the scheme with GPUs and demonstrate that it works well for GPUs.

cs.DC↗

arXiv2025

DuetServe: Harmonizing Prefill and Decode for LLM Serving via Adaptive GPU Multiplexing

Modern LLM serving systems must sustain high throughput while meeting strict latency SLOs across two distinct inference phases: compute-intensive prefill and memory-bound decode phases. Existing approaches either (1) aggregate both phases on shared GPUs, leading to interference between prefill and decode phases, which degrades Time-Between-Tokens (TBT); or (2) disaggregate the two phases across GPUs, improving latency but wasting resources through duplicated models and KV cache transfers. We present DuetServe, a unified LLM serving framework that achieves disaggregation-level isolation within a single GPU. DuetServe operates in aggregated mode by default and dynamically activates SM-level GPU spatial multiplexing when TBT degradation is predicted. Its key idea is to decouple prefill and decode execution only when needed through fine-grained, adaptive SM partitioning that provides phase isolation only when contention threatens latency service level objectives. DuetServe integrates (1) an attention-aware roofline model to forecast iteration latency, (2) a partitioning optimizer that selects the optimal SM split to maximize throughput under TBT constraints, and (3) an interruption-free execution engine that eliminates CPU-GPU synchronization overhead. Evaluations show that DuetServe improves total throughput by up to 1.3x while maintaining low generation latency compared to state-of-the-art frameworks.

cs.LG↗

arXiv2025

Achieving Empirical Potential Efficiency with DFT Accuracy: A Neuroevolution Potential for the $\alpha$-Fe--C--H System

A neuroevolution potential (NEP) for the ternary $\alpha$-Fe--C--H system was developed based on a database generated from spin-polarized density functional theory (DFT) calculations, achieving empirical potential efficiency with DFT accuracy. At the same power consumption, simulation speeds using NEP are comparable to, or even faster than, those with bond order potentials. The NEP achieves DFT-level accuracy across a wide range of scenarios commonly encountered in studies of $\alpha$-Fe- and $\alpha$-Fe--C under hydrogen environments. The NEP enables large-scale atomistic simulations with DFT-level accuracy at the cost of empirical potentials, offering a practical tool to study hydrogen embrittlement in steel.

cond-mat.mtrl-sci↗