arXiv subjects

Han Xu

Publications and source records attributed to Han Xu.

At least 19 recordsLinked to original sources

arXiv2026

Harbor Adapters and Harbor-Index: Infrastructure and a Curated Meta-Dataset for Large-Scale Agentic Evaluation

Evaluating agents on the growing number of agentic benchmarks is challenging because they often require complex environments and agent integrations. We introduce Harbor Adapters, a unified evaluation infrastructure for agentic benchmarks. Our work makes three contributions. First, we develop benchmark adapters that port more than 80 benchmarks to evaluate arbitrary agents, and validate them through rigorous code review and parity experiments. Second, we conduct a large-scale evaluation of 8 models spanning capability tiers across 54 benchmarks; every model is run with Terminus-2 and with one of 3 native harnesses. This enables a broader analysis of agent capabilities and failure modes than was previously possible. Third, we introduce Harbor-Index, a curated set of 82 difficult, diverse, and high-quality tasks spanning 29 benchmarks, refined from the adapted suite through difficulty filtering, AI and human audit, and an audit-and-fix loop. Harbor-Index preserves the challenge and breadth of large-scale agentic evaluations while being affordable to run; no evaluated model-harness configuration exceeds 30% pass rate, and the strongest (GPT-5.5 with Codex) reaches 28.0%. We release the adapters, evaluation results, in-depth analysis, and Harbor-Index as open-source artifacts to support more reliable and comprehensive evaluation of language-model agents.

cs.AI↗

arXiv2026

STAMP: Provenance-Guided Credit Assignment for Deep Search Agents

Reinforcement learning for deep-search agents has largely focused on trajectory-level scoring -- outcome correctness, citation-aware rewards, and evidence coverage. Yet the actions that expose supporting documents receive no targeted credit, a gap we call the reward-credit mismatch. We propose STAMP, in which a reference-based verifier judges whether each cited document supports an entity or relation in a training-time evidence graph, and first-exposure attribution traces each supported citation back to the action that first surfaced it. This step credit is injected through sign-preserving advantage modulation, which redistributes advantage across steps without changing the trajectory-level reward or the relative ranking of trajectories within each group. On BrowseComp, BrowseComp-ZH, and xbench-DS, STAMP improves the GRPO baseline by +2.0/+5.5/+3.0 points under matched SFT initialization, training data, and search tools, and composes with both outcome-only and citation-rubric base rewards. Component ablations confirm that the provenance-based credit signal and the sign-preserving advantage modulation each contribute to the gains.

cs.AI↗

arXiv2026

SCOPE-RL: Optimizing Reasoning Paths Before and After Success

Reinforcement learning with verifiable rewards (RLVR) optimizes LLMs using sparse verifiable final-answer rewards. This sparse anchor reliably verifies whether a trajectory succeeds but provides no direct feedback on the reasoning path that produced it. Before success, prerequisite progress on hard problems receives no reward signal; after success, outcome rewards cannot distinguish well-organized correct trajectories from redundant or locally flawed ones. We introduce SCOPE-RL (Scaffolded Chain Optimization with Process Efficiency), a two-stage framework that densifies this anchor while retaining the GRPO update: Adaptive Scaffolded RL adds prefix-decomposed verifiable rewards on answer-hidden sub-question chains before success, and Quality-Aware Process RL applies correctness-gated process-shape rewards to refine correct trajectories after success. An expert-validated Step-Quality Evaluation Protocol evaluates useful-step density, error localization, and token efficiency beyond final-answer accuracy. On Qwen3-8B-Instruct trained on DAPO-Math and Big-Math, SCOPE-RL improves average accuracy by up to 11.2 pp and reduces reasoning tokens by up to 27.1% over outcome-only GRPO; the gains hold under GSPO and on Qwen3-0.6B-Instruct, indicating that reward-signal densification is complementary to policy-update-level RLVR advances. Code and data are available at https://github.com/tokencraft-lab/SCOPE-RL.

cs.LG↗

arXiv2026

Kleene Algebra with Transitive Commutativity Conditions

Kleene algebra (KA) provides a foundational algebraic framework for reasoning about program structure and control flow. To capture equivalences arising from reordering or independence of actions, Kozen [1996] purposed that KA can be extended with commutativity conditions, that is, equations of the form { ab = ba | (a,b) \in C }, where C is a binary relation on constant symbols. This paper studies the following question: for which relations C is the equational theory of KA+C decidable? Early related work [Bertoni et al. 1982; Ibarra 1978] showed that regular languages modulo commutativity conditions C are decidable if and only if C is transitive. For Kleene algebra KA and commutativity conditions C, however, the situation is substantially more difficult. Only very recently, Kuznetsov [2023] showed that the equational theory of Kleene algebra KA+C is undecidable under certain specific commutativity conditions, settling the first nontrivial cases more than 25 years after the corresponding problem for KA* +C was resolved by Kozen [1996]. Nevertheless, the decidability problem of KA+C remained open. In this work, we resolve this question completely by showing that the equational theory of KA+C is decidable if and only if C is transitive. Moreover, we strengthen the result in both directions. On the negative side, we show that when C is not transitive, the universality problem for KA+C is already undecidable. On the positive side, we show that for transitive C, the equational theories of KA* +C and KA+C coincide.

cs.PL↗

arXiv2026

Network Analysis with Parametric NetKAT

Network engineers often need to perform network diagnosis and inference tasks, which frequently require answers to enumeration questions such as "Which packets from the Internet arrive at host C?" or "Which single-link failures disconnect my network?" Parametric NetKAT is a new domain-specific language that combines elements of NetKAT, Relational NetKAT, and Weighted NetKAT into a single system and extends them with parameters, allowing users to pose such enumeration questions directly over network models. This paper presents the design and semantics of Parametric NetKAT and illustrates its utility through a series of examples. It shows how to compile Parametric NetKAT into NetKAT automata, develops new algorithms for efficiently collecting satisfying valuations, and proves the correctness of these procedures. Finally, it evaluates the performance of Parametric NetKAT on a collection of benchmarks drawn from industrial sources.

cs.PL↗

arXiv2026

Burst Spiking Neural Networks

A central goal of current Spiking Neural Network (SNN) research is to improve their accuracy toward becoming low-power alternatives to Artificial Neural Networks (ANNs). This work further argues that realizing this ambition requires improving not only accuracy but also robustness, defined as the ability to maintain correct predictions under input perturbations. We identify two key issues in existing SNN methods that undermine robustness. First, binary spiking activations can produce large activation-state changes under small perturbations. Second, the lack of effective weight constraints makes network outputs more sensitive to input variations. To this end, we propose Burst Spiking Neural Networks (BuSNNs), built upon Burst-enhanced Spiking Neurons (BSNs) and a Dynamic Weight Constraint (DWC) mechanism. BSNs incorporate burst firing to provide a graded spiking pattern. This spiking mechanism mitigates perturbation-induced transitions in activation states and thereby enhances robustness. DWC penalizes connection weights based on activation states, effectively reducing weight magnitudes and improving robustness while preserving accuracy. We provide theoretical analyses to support these robustness effects. Experimental results further show that, on smaller-scale benchmarks such as CIFAR-10, BuSNNs outperform both SNN and ANN counterparts in accuracy and robustness. On large-scale ImageNet, BuSNN with the MS ResNet-34 backbone further improves top-1 accuracy and corruption robustness over the corresponding SNN baseline by 3.18% and 2.66%, respectively. Despite using spike-based activations, BuSNNs surpass 4-bit activation-quantized ANN baselines and approach 8-bit ANN baselines on ImageNet. They also preserve SNNs' low-power advantage. This work studies the accuracy-robustness problem in SNNs, advancing their practical viability in robust and energy-efficient applications.

cs.NE↗

arXiv2026

UHD-MFF: Shattering Barriers in Multi-Focus Ultra-High-Definition Image Fusion via Learnable Lookup Tables

With the advancement of imaging technology, ultra-high-definition images have become increasingly essential in modern visual applications. However, existing multi-focus image fusion remains largely confined to low-resolution images and faces three major barriers in UHD scenarios, namely data availability, model adaptability, and deployment feasibility, which severely hinder its practical application. To shatter these barriers, first, we propose the UHD-MFF dataset, the first large-scale ultra-high-resolution multi-focus fusion dataset. Second, we propose a scale-specialized lookup-table framework tailored for ultra-high-resolution images, termed as UMF-LUT. It consists of Coarse-Region Lookup Table (C-LUT) and Detail-Edge Lookup Table (D-LUT). Specifically, C-LUT performs joint queries of multiple gradient cues and semantic cues at low-resolution scales to enable region-level decision-making. Also, D-LUT operates at high-resolution scales, leveraging efficient Laplacian cues to provide complementary edge-level decision information. Such a design makes the model particularly well-suited for ultra-high-resolution multi-focus image fusion. Finally, it offers strong deployability with minimal computational overhead, enabling real-time 4K multi-focus fusion and showing promising potential for smartphone. Extensive experiments demonstrate that it outperforms SOTA methods in both visual fidelity and quantitative metrics. It effectively advances the development of multi-focus image fusion toward ultra-high-resolution imaging scenarios. The code is available at https://github.com/zyb5/UHD-MFF.

cs.CV↗

arXiv2026

Data and Evaluation Closed-Loop for Model Capability Enhancement

Model capability is the central variable in LLM pre-training, yet is never observed directly: data shapes it prospectively, while evaluation reveals it only retrospectively, compressing samples, prompts, decoding, and scoring rules into one noisy score. Practical optimization runs this backward: a failure is observed first, and the engineer must infer the corpus fix. The two sides speak incompatible vocabularies -- benchmark names and per-sample correctness versus data sources, domains, and quality labels -- so this inference is usually intuition, not method. We close this gap with the \emph{capability slice}: a group of evaluation samples sharing background condition, task type, solving operation, and output constraint -- precise enough to localize a single weakness yet stable enough to survive aggregation, unlike a benchmark name, too coarse, or a single sample, too noisy. Built around this unit, an evaluation taxonomy, a non-instruction data taxonomy, and mapping rules form a closed loop turning a benchmark-level failure into a targeted, testable data intervention. We test this loop on two case studies pulling in opposite directions. First, the loop rules the data out: continued pre-training drives BBH down by $-46.82\%$, but diagnosis traces this to a single masked \texttt{\textless EOS\textgreater} loss rather than weakened reasoning; restoring it recovers BBH to $66.44$, above the original checkpoint, without changing the data. Second, the loop rules the data in: a persistent math-reasoning weakness is decomposed by solving operation into specific failing combinations, and a weakness-targeted sampling procedure built from it lifts AIME2025/AIME2026 Pass@128 from $6.67$/$0.00$ to $26.67$ each. The same unmodified loop reaches opposite, correct verdicts in both cases, showing the evaluation-to-data inference can be routine, auditable, and experimentally validated rather than intuitive.

cs.AI↗

arXiv2026

Let the Data Decide: Supervision Analysis, Capability Trade-offs, and Adaptive Objective Routing in Continued Pre-Training via Off-Policy Distillation

Off-policy distillation is now central to large language model pre-training, yet how training data, objective parameterization, and model capabilities interact remains poorly characterized. We studies top-$k$-truncated, temperature-scaled off-policy distillation by decomposing this problem into two questions: an \emph{objective-to-capability} analysis of how the training objective shapes token-level supervision and downstream performance, and a \emph{data-to-objective} analysis of how data heterogeneity should inform objective routing. We first show that the language-modeling objective ($L_{\mathrm{LM}}$) and the knowledge-distillation objective ($L_{\mathrm{KD}}$) induce systematically different capability profiles, and trace this divergence to a gradient-level tension between \emph{direct observed-token reinforcement} and \emph{teacher-supported alternative supervision}. To quantify this tension, we introduce diagnostic metrics -- support coverage, observed-token probability mass, and teacher-distribution concentration -- and show via controlled sweeps that the support size $k$ governs a coverage-sharpness trade-off, while distillation temperature controls within-support probability allocation. We then examine adaptive objective routing: a domain-level policy that applies $L_{\mathrm{LM}}$ to math and code and $L_{\mathrm{KD}}$ to general-domain data yields consistent gains over both single-objective baselines, whereas token-level routing based on observed-token probability mass or teacher entropy fails to consistently match the single-objective baseline. These results suggest that effective objective routing depends less on routing granularity than on the quality of the routing signal, reframing continued pre-training via off-policy distillation as a structured, data-conditional supervision-design problem rather than a global hyperparameter choice.

cs.LG↗

arXiv2026

Sub-8-nm resolution AKB-mirror-based hard X-ray ptychography via generalized Wirtinger projections

Hard X-ray ptychography has become increasingly essential in both the life and physical sciences. However, pushing resolution down to a few nanometres often requires highly customized, chromatic diffractive or refractive X-ray nanofocusing optics, significantly limiting the practical broadband energy-scan applications. Here, we present the first known hard X-ray ptychographic imaging with a half-pitch resolution below 8 nm using total-reflection Advanced Kirkpatrick-Baez (AKB) mirror nanofocusing optics at the high energy photon source (HEPS), with clear potential for further extension. Despite leveraging the benefits of enhanced instrumentation, such as the high coherent flux of 4th-generation diffraction-limited storage rings (DLSR), state-of-the-art beamline X-ray optics and detectors, this is made possible by developing a reconstruction algorithm termed Generalized Wirtinger Projections (GWP). We derive the theory of GWP and experimentally demonstrate its capability for improved partial-coherence reconstruction and enhanced spatial resolution over conventional methods through imaging experiments on a Siemens star test chart at 12.4 keV. GWP provides a highly compact framework for jointly accounting for multiple coupled uncertainties that degrade resolution, enabling straightforward extension to other imaging modalities, such as burst ptychography, while delivering nearly an order-of-magnitude improvement in GPU memory efficiency. Furthermore, the ability to combine nanometre-scale spatial resolution with the inherently achromatic nanofocusing optics demonstrated in this work potentially opens new opportunities for in situ or operando broadband, energy-scan 3D spectroscopic imaging with element- or chemical-state specificity in complex environments at the nanoscale, holding significant promise for a wide range of applications from electronics and energy science to neuroscience.

physics.optics↗

arXiv2026

Filter-assisted quantum subspace diagonalization via wavefunction sparsity engineering

Subspace diagonalization techniques based on quantum sampling, such as quantum selected configuration interaction (QSCI) and sample-based quantum diagonalization (SQD), have recently emerged as promising quantum-centric approaches for approximating ground-state energies of many-body systems. However, their performance is fundamentally limited by an intrinsic trade-off between sampling efficiency and the sparsity of the ground-state wavefunction, which becomes particularly severe in strongly correlated systems. Here, we introduce a filter-assisted SQD protocol that engineers wavefunction sparsity via a quantum filter, i.e., a unitary transformation of the Hamiltonian designed to concentrate the ground-state weight onto a small number of computational basis states. Using the Gini coefficient as a robust sparsity measure, we establish a quantitative relationship between wavefunction sparsity and the resource requirements of SQD, providing theoretical bounds on the required subspace dimension and sampling cost. To realize the quantum filter, we employ a tensor-network-based circuit-encoding algorithm that maps target states to quantum circuits with controllable fidelity. We benchmark our approach on the quantum Ising model with transverse and longitudinal fields using both numerical simulations and quantum hardware experiments. Our results demonstrate that, compared with standard SQD, the proposed protocol significantly enhances wavefunction sparsity, reduces ground-state energy estimation errors by orders of magnitude, and substantially lowers sampling overhead. These findings establish filter-assisted subspace diagonalization as a powerful and scalable framework for quantum many-body calculations in the strongly correlated regime.

quant-ph↗

arXiv2026

Crossing the 12,000-atom barrier with heterogeneous quantum-classical supercomputing: quantum chemistry of protein-ligand complexes

We develop a workflow decomposing a molecule into fragments via quantum embedding and simulating them with a heterogeneous quantum-classical (HQC) method. We sample fragment electronic configurations on two 156-qubit quantum processors (ibm$\_$cleveland, ibm$\_$kobe), using up to 94 qubits, running 21,006 circuits for over 239 hours, collecting $3.0 \cdot 10^9$ measurement outcomes -- the most resource-intensive HQC computation for quantum chemistry to date. We compute fragment wavefunctions via optimized subspace diagonalization on supercomputers Fugaku, Miyabi-G, and ROQUO, achieving 72.5$\%$ parallel efficiency with scalable distributed linear algebra kernels. We simulate two protein-ligand complexes spanning dispersion- and electrostatics-dominated regimes (11,608, 12,635 atoms), demonstrate $>40\times$ increase in system size and up to $210\times$ improvement in accuracy over previous state-of-the-art, with HQC matching coupled-cluster (CCSD) accuracy in fragment energies. We present the first HQC protein-ligand binding prediction using a mixed-basis set and an automated end-to-end workflow enabling practical HQC calculations of large protein systems.

quant-ph↗

arXiv2026

GLM-5V-Turbo: Toward a Native Foundation Model for Multimodal Agents

We present GLM-5V-Turbo, a step toward native foundation models for multimodal agents. As foundation models are increasingly deployed in real environments, agentic capability depends not only on language reasoning, but also on the ability to perceive, interpret, and act over heterogeneous contexts such as images, videos, webpages, documents, GUIs. GLM-5V-Turbo is built around this objective: multimodal perception is integrated as a core component of reasoning, planning, tool use, and execution, rather than as an auxiliary interface to a language model. This report summarizes the main improvements behind GLM-5V-Turbo across model design, multimodal training, reinforcement learning, toolchain expansion, and integration with agent frameworks. These developments lead to strong performance in multimodal coding, visual tool use, and framework-based agentic tasks, while preserving competitive text-only coding capability. More importantly, our development process offers practical insights for building multimodal agents, highlighting the central role of multimodal perception, hierarchical optimization, and reliable end-to-end verification.

cs.CV↗

arXiv2026

SpikingBrain2.0: Brain-Inspired Foundation Models for Efficient Long-Context and Cross-Platform Inference

Scaling context length is reshaping large-model development, yet full-attention Transformers suffer from prohibitive computation and inference bottlenecks at long sequences. A key challenge is to design foundation models that maintain performance and long-context efficiency with minimal training overhead. We introduce SpikingBrain2.0 (SpB2.0), a 5B model that advances both architecture and training efficiency of its predecessor. Our contributions are two-fold. (1) Architectural Innovation: We propose Dual-Space Sparse Attention (DSSA), an inter-layer hybrid of Sparse Softmax Attention (MoBA) and Sparse Linear Attention (SSE), achieving an improved performance-efficiency trade-off for long-context modeling. SpB2.0 further supports dual quantization paths: INT8-Spiking coding enables sparse event-driven computation, while FP8 coding accelerates inference on modern GPUs. (2) Enhanced Training Strategy: We develop an optimized Transformer-to-Hybrid (T2H) pipeline with dual conversion paths for LLMs and VLMs using curated open-source data. Empirically, SpB2.0-5B and SpB2.0-VL-5B recover most of the base Transformer (Qwen3-4B) capability with under 7k A100 GPU hours. SpB2.0 achieves a 10.13x TTFT speedup at 4M context and supports over 10M tokens on 8 A100 GPUs under vLLM, where full-attention models exceed memory limits. It also demonstrates strong cross-platform compatibility, enabling FP8 GPU inference (2.52x speedup at 250k) and efficient neuromorphic execution (64.31% sparsity, with 70.6% and 46.5% area and power reduction at 500MHz). Overall, SpikingBrain2.0 provides a practical pathway for lightweight, multimodal, spiking foundation models, highlighting the potential of combining brain-inspired mechanisms with efficient architectures for resource-constrained and edge scenarios.

cs.LG↗

arXiv2026

SpikeMLLM: Spike-based Multimodal Large Language Models via Modality-Specific Temporal Scales and Temporal Compression

Multimodal Large Language Models (MLLMs) have achieved remarkable progress but incur substantial computational overhead and energy consumption during inference, limiting deployment in resource-constrained environments. Spiking Neural Networks (SNNs), with their sparse event-driven computation, offer inherent energy efficiency advantages on neuromorphic hardware, yet extending them to MLLMs faces two key challenges: heterogeneous modalities make uniform spike encoding insufficient, and high-resolution image inputs amplify timestep unfolding overhead. We propose SpikeMLLM, the first spike-based framework for MLLMs, which unifies existing ANN quantization methods in the spiking representation space and incorporates Modality-Specific Temporal Scales (MSTS) guided by Modality Evolution Discrepancy (MED) and Temporally Compressed LIF (TC-LIF) for timestep compression from T=L-1 to T=log2(L)-1. Experiments on four representative MLLMs across diverse multimodal benchmarks show that SpikeMLLM maintains near-lossless performance under aggressive timestep compression (Tv/Tt=3/4), with average gaps of only 0.72% and 1.19% relative to the FP16 baseline on InternVL2-8B and Qwen2VL-72B. We further develop a dedicated RTL accelerator tailored to the spike-driven datapath, observing 9.06x higher throughput and 25.8x better power efficiency relative to an FP16 GPU baseline under a deployment-oriented co-design setting, suggesting the promise of algorithm-hardware co-design for efficient multimodal intelligence.

cs.NE↗

arXiv2026

Spike-driven Large Language Model

Current Large Language Models (LLMs) are primarily based on large-scale dense matrix multiplications. Inspired by the brain's information processing mechanism, we explore the fundamental question: how to effectively integrate the brain's spiking-driven characteristics into LLM inference. Spiking Neural Networks (SNNs) possess spike-driven characteristics, and some works have attempted to combine SNNs with Transformers. However, achieving spike-driven LLMs with billions of parameters, relying solely on sparse additions, remains a challenge in the SNN field. To address the issues of limited representational capacity and sparsity in existing spike encoding schemes at the LLM level, we propose SDLLM, a spike-driven large language model that eliminates dense matrix multiplications through sparse addition operations. Specifically, we use the plug-and-play gamma-SQP two-step spike encoding method to ensure that the quantization process aligns with the model's semantic space, mitigating representation degradation caused by binary spikes. Furthermore, we introduce bidirectional encoding under symmetric quantization and membrane potential clipping mechanisms, leading to spike trains with no or low firing counts dominating, significantly reducing the model's spike firing rate, while halving the number of time steps. Experimental results show that SDLLM not only significantly reduces inference costs but also achieves state-of-the-art task performance under the spike-based paradigm. For example, compared to previous spike-based LLMs, SDLLM reduces energy consumption by 7x and improves accuracy by 4.2%. Our model provides inspiration for the architecture design of the next generation of event-driven neuromorphic chips.

cs.NE↗

arXiv2026

Phase transitions in parametrized quantum circuits

Phase transitions are among the most intriguing phenomena in physical systems, yet their behavior near criticality remain challenging to study using classical algorithms. Parameterized quantum circuits (PQCs) offer a promising approach to investigating such regimes on practical quantum computers. However, in order to use it to probe critical behavior, a PQC itself should be non-trivial and exhibit a phase transition and non-analyticity -- a property that has not yet been clearly identified. In this work, we identify a mechanism for generating non-analyticities intrinsically in PQCs. As a concrete realization, we construct a class of sequential PQCs whose observable expectation value is a non-analytic function of the circuit parameter in the infinite volume limit, showing that the prepared PQC states undergo a phase transition at the non-analytic points. The entanglement and the identified order parameter have distinct behaviors in different phases, revealing a phase diagram of the PQC state. We show that classical simulation of this PQC based on tensor networks and Pauli propagation gets less efficient in the vicinity of the phase transition point, indicating a physically motivated route towards practical quantum advantage using PQCs with phase transitions.

quant-ph↗

arXiv2026

Data-Driven Successive Linearization for Optimal Voltage Control

Power distribution systems are increasingly exposed to large voltage fluctuations driven by intermittent renewable generation and time varying loads (e.g., electric vehicles and storage). To address this challenge, a number of advanced controllers have been proposed for voltage regulation. However, these controllers typically rely on fixed linear approximations of voltage dynamics. As a result, the solutions may become infeasible when applied to the actual voltage behavior governed by nonlinear power flow equations, particularly under heavy power injection from distributed energy resources. This paper proposes a data-driven successive linearization approach for voltage control under nonlinear power flow constraints. By leveraging the fact that the deviation between the nonlinear power flow solution and its linearization is bounded by the distance from the operating point, we perform data-driven linearization around the most recent operating point. Convergence of the proposed method to a neighborhood of KKT points is established by exploiting the convexity of the objective function and structural properties of the nonlinear constraints. Case studies show that the proposed approach achieves fast convergence and adapts quickly to changes in net load.

eess.SY↗