arXiv subjects

Xing Li

Publications and source records attributed to Xing Li.

At least 19 recordsLinked to original sources

arXiv2026

Salvation Lies Within: Eliciting Inherent Style Transfer in Step-Distilled Diffusion Models

Adapting step-distilled text-to-image (T2I) models through post-training incurs additional computational costs and affects native few-step generation behavior. This motivates a complementary route beyond style-specific adaptation: drawing on the visual knowledge already encoded in step-distilled T2I models to elicit stylistic capabilities through language. Pursuing this direction requires textual guidance that captures how visual attributes jointly define a style and remain applicable as the depicted content changes. To explore this approach, we introduce StyleForge, a fully automatic, training-free framework that expresses reference styles as reusable rendering instructions. By integrating overall rendering characteristics with local color and lighting behavior, StyleForge organizes visual evidence from reference images into a coherent specification of how the target style should be expressed. The specification is then compiled into textual guidance that can be reused across content prompts, enabling frozen step-distilled T2I models to render different subjects and scenes in the reference style while retaining native few-step generation. Extensive experiments show relative gains of up to 29.47\% in generation quality scores over the strongest baseline, while Pareto analysis indicates that improved stylization is accompanied by strong adherence to the requested content. Code is available at https://github.com/sunshy-1/StyleForge.

cs.CV↗

arXiv2026

Embedding a de Rham Complex Structure into Eig-PIELM for Maxwell Eigenvalue Problems

The perfect electric conductor (PEC) boundary condition alone does not eliminate the gradient nullspace of the Maxwell curl-curl operator. Consequently, a boundary-only Maxwell implementation of Eig-PIELM, an eigenvalue extension of the physics-informed extreme learning machine (PIELM), may retain represented gradient modes in its reduced trial space, producing nonphysical, gradient-contaminated modes. We develop dR-Eig-PIELM, a finite-dimensional structure-preserving extension that retains the linear-coefficient, boundary-null-space, and one-shot multimode architecture of Eig-PIELM. It combines compatible tensor-product grad--curl trial spaces with a constrained null-space construction to restrict the sampled PEC space to the mass-orthogonal complement of the represented gradient image before the eigensolve. Symmetric weak-form Gram assembly yields a symmetric-definite reduced eigenproblem without an eigenvalue cutoff under the stated sampling and material-positivity assumptions. For rectangles and cuboids, we establish fixed-degree exactness at the PEC vector-field space and derive rank and dimension identities. Numerical experiments show that dR-Eig-PIELM excludes the spurious eigenmodes with significant gradient components observed in the boundary-only baseline. The method exhibits rapid convergence with increasing polynomial degree and competitive work--precision performance in the tested configurations. A covariantly mapped waveguide example illustrates the potential of dR-Eig-PIELM for integration into design optimization workflows through a constrained local geometry update driven by computed cutoff frequencies and their sensitivities. This work provides a mathematical and computational framework for separating represented gradient modes from physical Maxwell eigenmodes within Eig-PIELM and supports its use in waveguide spectral analysis and geometry design.

math.NA↗

arXiv2026

Scaling LLM Agent Tool Access: The MCP Gateway

LLM agents increasingly rely on tool calling to complete tasks. Model Context Protocol (MCP) standardizes tool calling between LLM agents and tool providers. It reduces integration cost and drives rapid growth in tool scale. In production deployments, tool access becomes difficult to scale. On the tool provider side, legacy services are not directly callable through MCP; the rapid protocol development creates ongoing compatibility cost. On the agent side, the number of mounted tools is limited by the LLM context window and inference overhead; mounting a large tool set increases token usage and inference latency and can reduce task success rate. Moreover, for stateful MCP backends with multiple replicas, preserving session affinity increases client-side complexity. We present a cloud-scale MCP Gateway system. It breaks the direct-connect model on the data plane and offloads legacy service integration, consolidating incompatible MCP variants, access control, tool recommendation, and session-aware routing at the gateway. The conversion module supports 315 OpenAPI services in the cloud and integrates MCP transport variants. Session-aware routing decouples the scalability of the gateway layer and the MCP service layer. The hybrid retrieval engine sustains $\sim$98\% Top-15 recall; it scales agent tool access to 3{,}000+ tools with improved tool selection accuracy. We also share the lessons learned from deploying the gateway system in production.

cs.DC↗

arXiv2026

A Clustered Equivalent-Source Method with Local Response Compression and Group-Symmetry Reduction for Antenna-Array Analysis

Full-wave analysis of finite antenna arrays is computationally demanding because detailed element responses must be resolved and coupled across the entire array. To reduce this cost, this paper presents a clustered equivalent-source method that combines local response compression with group-symmetry reduction. Local responses are represented in compressed equivalent-surface bases or coordinates inherited from child clusters. Symmetry constraints associated with the cluster partition and excitation reduce the coupled system, while the corresponding group transformations recover the complete response in the retained coordinates. To assess overall efficiency, we evaluate the costs of local construction, propagation, global solution, and recovery. Numerical examples involving layered and finite annular patch arrays show close agreement between the computed and reference complex far fields and shorter complete-run times than those of the corresponding symmetry-reduced direct solvers. Among the three equivalent-source implementations for the layered array, the one with the smallest coupled system has the longest runtime because local construction and propagation dominate the cost. These findings highlight the need to consider local construction and coupling costs when assessing the benefit of reducing the global system dimension.

math.NA↗

arXiv2026

Elliptic soliton solutions of lattice eigenfunction equations of the KdV type

It is known that some integrable lattice equations can be formulated by the eigenfunctions of the Lax triplet of the lattice potential Korteweg-de Vries (lpKdV) equation; the eigenfunctions satisfy a discrete Abel formula in which the `integration' constant defines the $δ$ parameter in some equations in the Adler-Bobenko-Suris (ABS) classification. In this paper, we present a full Darboux transformation for the Lax triplet of the lpKdV equation, and eigenfunctions are formulated in terms of Casoratians. When the eigenfunctions take the elementary discrete Lamé functions, we get the elliptic $N$-soliton solution for the lpKdV equation after $N$-fold Darboux transformation. We then determine the constant in the discrete Abel formula in this case. The eigenfunctions are used to formulate elliptic $N$-soliton solutions for various equations in the ABS list, such as H3($δ$), $Q1(Δ)$, Q2 and Q3($Δ$) equations.

nlin.SI↗

arXiv2026

ESPC-Transolver: An Equivalent-Surface Physics-Guided Coupling-Aware Transolver for 3D Multiple Scattering

Cross-geometry generalization is particularly challenging because geometry-conditioned multiple-scattering surrogates must infer unseen objects' scattering properties from geometric features while accurately capturing complex, phase-sensitive interactions arising from repeated wave exchange. We introduce ESPC-Transolver, an equivalent-surface physics-guided, coupling-aware Transolver that separates local scattering-response representations from inter-object interaction modeling. For each sound-hard scatterer, canonical scattered-pressure responses are constructed under a shared set of spherical-wave excitations. A local Dirichlet-to-Neumann (DtN) map and modal scattering conversion express these responses on an enclosing sphere, providing physically consistent input channels across geometries. Shared object-local Transolver blocks encode these responses and incident-field traces, while relative-position-aware slice attention models inter-object coupling. Predicted interface pressure and normal flux recovered from the local operator define equivalent sources for reconstructing exterior near- and far-field scattering. We evaluate the method on three-dimensional acoustic multiple-scattering problems at a fixed frequency, comparing it with geometry-conditioned predictors and alternative neural network backbones. ESPC-Transolver achieves lower scattered-field prediction errors on held-out instances of familiar geometry families and generalizes effectively to previously unseen Peanut and Torus geometries. Exterior-field reconstructions reproduce multiple-scattering interference patterns, linking cross-geometry interface prediction to near-field pressure and far-field scattering recovery.

math.NA↗

arXiv2026

Can Decision Models Understand Stance? Evaluating Jev Against General-Purpose LLMs

Stance detection requires identifying an author's attitude toward a given target, sometimes based on conversational context. Jev, a specialized decision model designed for structured decision-making, offers an alternative to general-purpose large language models (LLMs). In this work, we evaluate Jev on two stance detection datasets, VAST (English texts) and ZS-CSD (Chinese conversations), comparing it with four general-purpose LLMs and two fine-tuned models. Results show that Jev achieves competitive performance on VAST, matching GPT-5.6 and outperforming the other general-purpose LLMs. However, it falls behind stronger LLMs on ZS-CSD, particularly in distinguishing favor from against. Further analysis suggests that this limitation may be related to understanding reply relationships and stance direction rather than conversation length alone. These findings highlight both the potential and limitations of Jev for stance detection.

cs.CL↗

arXiv2026

Specialized Decision Models vs. General-Purpose LLMs: Benchmarking Jev Across Knowledge, Reasoning, and Multilingual Tasks

Jev is a "System One" model that returns a choice among given options instead of generating text. We study how such a specialized decision model compares with general-purpose large language models (LLMs). We evaluate Jev on 13 multiple-choice benchmarks covering knowledge, reasoning, and multilingual understanding, and compare it with 19 LLMs in three tiers: frontier, representative, and small. Jev is competitive with frontier LLMs on knowledge and commonsense benchmarks and obtains the best score on MMLU-Redux and ARC-Challenge. Outside mathematics, it also outperforms most representative LLMs and all small LLMs. However, it falls behind on mathematical word problems: on MathQA, it is 17.7 points below the frontier median and scores lower than all 19 LLMs. These results indicate that a specialized decision model can match general-purpose LLMs on decisions that rely mainly on knowledge, but not on decisions that require multi-step calculation.

cs.CL↗

arXiv2026

Multi-Track Time-Series Burst-Overlap Interferometry for Resolving Horizontal Deformation in Earthquake-Cycle Studies

Interferometric Synthetic Aperture Radar (InSAR) is intrinsically insensitive to the north-south component of crustal motion because of its near-polar geometry, which makes accurate quantification of this deformation a persistent limitation of InSAR. We introduce a Multi-track Time-Series Burst-Overlap Interferometry (MTSB) framework that delivers absolute, ITRF-referenced deformation fields across representative interseismic, coseismic, and postseismic applications by combining optimized phase linking, a unified time-series framework for separating deformation and residual misregistration, and block-wise spectral analysis. Case studies spanning co-, inter-, and postseismic phases demonstrate that MTSB accurately resolves along-track horizontal deformation, with primary sensitivity to north-south motion, while substantially reducing orbit-related artifacts. Independent comparisons with GNSS indicate centimeter-level agreement for postseismic and coseismic displacement estimates and millimeter-per-year agreement for interseismic velocities. The resulting deformation fields improve constraints on regional plate kinematics and fault kinematics, facilitate more reliable Euler-pole estimation, and provide by-product orbital corrections for conventional InSAR processing.

physics.geo-ph↗

arXiv2026

Experience Funnel: A State-Policy Alternating Loop for Self-Evolving Agents

Autonomous agents powered by large language models (LLMs) continuously accumulate experience through interaction, creating an opportunity to improve future behavior through self-evolution. A fundamental challenge is how to transform abundant, task-specific interaction experience into reusable model competence without sacrificing the ability to adapt rapidly to newly observed evidence. Explicit textual states, such as skills and agent harnesses, provide fast, human-readable and editable adaptation, but incur persistent dependence on external context; parametric policies provide compact and reusable competence, but are substantially slower to update. We present \textit{Experience Funnel}, a self-evolving framework that couples fast state adaptation with slow policy consolidation in an alternating loop. Interaction trajectories are first distilled into an explicit textual state, where newly acquired experience can be rapidly incorporated and validated. The framework then selectively identifies state-enabled behavior that remains useful across state revisions and consolidates it into the policy through transition-aware distillation. The updated state--policy pair subsequently generates new rollouts, providing fresh evidence for the next round of state adaptation and policy consolidation. Experiments across diverse agent benchmarks show that \textit{Experience Funnel} consistently improves agent capability over state-only evolution and policy-internalization approaches, while progressively converting useful explicit experience into autonomous policy competence.

cs.CL↗

arXiv2026

How Does mHC Use Its Residual Streams? Selective Routing and Near-Identity Mixing

Hyper-Connections and their manifold-constrained variant mHC widen a residual pathway from one stream to n, yet how trained models use this capacity remains unclear: how broadly blocks read and write, how strongly the residual pathway mixes streams, and whether the streams carry distinct representations. We examine these properties in the four-stream residual pathway of DeepSeek-V4-Flash using effective stream counts, cross-stream residual weights, and inter-stream cosine similarity. Read/write routing is concentrated but varies across depth: a typical attention or FFN site effectively uses about two streams, while the dominant stream changes across layers and the representations remain directionally distinct. Residual mixing is modest and occurs primarily in early layers; in layers 22-42, the pathway mostly carries each stream forward separately. Targeted interventions establish the functional significance of these patterns. Replacing the late mixers by identity increases C4 perplexity by only 1.9% and preserves the six-task average score, whereas replacing the early mixers increases perplexity by 41%. Fixing each early mixer to its C4 diagnostic mean increases perplexity by only 0.2% and reduces the average score by 0.25 percentage points, showing that its site-specific structure matters more than its token-wise variation on the evaluated metrics. Likewise, retaining the three largest routing weights per token at every site increases perplexity by at most 2.7% and changes the average score by at most 0.4 points. Thus, the studied model realizes only part of the flexibility afforded by four-stream mHC: individual blocks rarely require all four streams, and late residual mixing provides little measured benefit.

cs.LG↗

arXiv2026

Sharp Concave-Function Transfer for Lee-Type Schatten Norm Inequalities

We compare the sharp constants in Lee-type Schatten norm inequalities for finite families of complex matrices with the corresponding constants obtained after applying an arbitrary nonnegative concave function to the relevant absolute values. We prove that the nonlinear problem has exactly the same best constant as the underlying linear problem for every finite matrix dimension, every number of summands, and every Schatten exponent, including the operator-norm endpoint. For finite exponents, the proof proceeds through a lossless single-cap reduction, an exact representation as a positive mixture for finite cap combinations, and a noncommutative reassembly argument based on a weighted Schatten contraction and the Araki-Lieb-Thirring inequality. Finite-spectrum cap interpolation and a perturbation argument at zero then yield the general concave case. Thus every sharp linear result transfers without loss to the full class of nonnegative concave functions.

math.FA↗

arXiv2026

A negative exponent range for Audenaert's complementary McCarthy trace inequality

Audenaert introduced a class of complementary McCarthy type trace inequalities in his work on completely monotone functions and Bernstein functions; the same problem was later included in the problem list of Audenaert and Kittaneh. The known results cover the corresponding directions for $q\le -2$, $0<q\le 1$, $1\le q\le 2$, and $2\le q\le 3$, while the negative range $-2<q<0$ was left as a conjectural case. We prove the negative exponent inequality, in fact for every $q<0$. After the inversion $X=A^{-1}$, $Y=B^{-1}$, the problem reduces to a trace inequality for the parallel sum $X:Y$. The proof uses the Kubo--Ando mean chain, an Ando--Hiai type log-majorization for matrix geometric means, and a finite-dimensional Schatten Hölder inequality, including the quasi-norm range. We also record that the positive exponent side $q\ge 3$ follows directly from Audenaert's norm-compression inequality for positive semidefinite $2\times2$ block matrices, and in fact holds for all $q\ge 2$. Finally, we determine the equality case on the negative side: equality holds if and only if $A=B$.

math.RA↗

arXiv2026

Fiber Nonlinearity Compensation of Coherent Signals Using Deep Photonic Reservoir Computer

Photonic reservoir computer (PRC) is a promising optical computing framework for high-speed optical signal processing, and various reports have shown its functionality of linear equalization for intensity-modulation direct-detection communication links. However, coherent communication links suffer more from nonlinear impairment, whereas its nonlinear equalization is very challenging. Here we demonstrate the nonlinear equalization of coherent 16-level quadrature amplitude modulation (16-QAM) signals using a deep PRC in experiment. The deep PRC consists of cascading injection-locked semiconductor lasers with optical feedback loops. For 16-QAM signals with a transmission rate of 240 Gbps and a launch power of 12 dBm, the single-channel PRC with 3 hidden layers raises the Q factor by as high as 0.58 dB, while the dual-channel PRC with 2 hidden layers raises the Q factor by 0.55 dB. In addition, we prove that the deep PRC is able to equalize 16-QAM signals of different launch powers and transmission distances.

physics.optics↗

arXiv2026

How Powerful are LLMs in Generating Formal Program Specifications?

Formal verification provides strong guarantees of software correctness, but its adoption is limited by the high cost of writing precise formal specifications. While recent large language models (LLMs) have shown strong capabilities in theorem proving and verified code generation, their true ability to generate program specifications remains unclear. Existing evaluations require either verifying implementation conformance or proving semantic equivalence between specifications, both of which are formidably difficult and may conflate proof difficulty with specification quality. To address this problem, we introduce Coins, a Rocq based evaluation framework that assesses specification quality by instantiating specifications under evaluation on trusted test cases and generating concrete proof obligations. This design aligns with the asymmetric nature of formal reasoning, where successful proofs provide reliable evidence while proof failures are inherently ambiguous. Using Coins, we conduct a large scale study on HumanEval with a curated set of human written Rocq specifications. Our results show that specification generation remains a formidable challenge, and that verification complexity can obscure genuine differences in specification quality. Overall, we find that accurate specification evaluation, rather than model scaling alone, is central to understanding the power of LLMs for specification synthesis, and that test case based formal reasoning offers a more faithful and discriminative measure of progress.

cs.SE↗

arXiv2026

Getting the Parameters Right: A Difficulty-Graded Benchmark and Probe-Guided Training for LLM Tool Calls

Large language model agents derive much of their capability from tool use. Existing research on tool use has largely focused on selecting the right tool and orchestrating the order of calls. However, correctly filling the parameters of a tool call is equally critical for successful execution and has received far less attention. In domains such as cloud networking, even frontier models correctly complete fewer than half of tool calls. Inspired by recent analyses showing that LLM hidden states encode rich information about model predictions, we discover that while the model generates a parameter value, its hidden state contains a strong correctness signal: a simple linear probe can accurately predict whether the value will be correct. Based on this observation, we propose a unified probe-guided framework with two complementary approaches: probe-filtered bootstrapped training (PBT), which uses the probe to filter reliable self-generated calls for fine-tuning, and probe-guided reranking (PGR), which uses the probe to select better candidates during inference. To support systematic evaluation, we release ParamBench, a benchmark built from real cloud-network APIs that categorizes every instance into five difficulty levels according to parameter nesting depth, cross-parameter dependencies, and the reasoning required to derive values from earlier calls. Extensive experiments across 5 open models on ParamBench and 6 external benchmarks demonstrate that our method substantially improves parameter generation, raising the average exact match from 19.7% to 59.6%.

cs.AI↗

arXiv2026

Kernel Foundry: A Diagnosis-driven Evolutionary Kernel Optimizer with Multi-Experts

Generating high-performance GPU kernels remains challenging due to the need for both correctness and hardware-aware optimization. While large language models (LLMs) show promise in code generation, they often fail to produce kernels that are both correct and efficient. We propose Kernel Foundry, a diagnosis-driven evolutionary framework for automatic GPU kernel optimization. Our method combines expert-guided, retrieval-augmented initialization with a multi-island evolutionary search, where candidate kernels are iteratively refined using structured diagnostic feedback. A centralized experience library accumulates reusable optimization knowledge to guide subsequent evolution, while explicit mechanisms prevent cheating behaviors that bypass kernel-level computation. Experiments on KernelBench show that our method consistently improves both correctness and performance over strong baselines, achieving up to 100% correctness on Level~2.

cs.NE↗

arXiv2026

SwiftMem: Fast Agentic Memory via Query-aware Indexing

Agentic memory systems have become critical for enabling LLM agents to maintain long-term context and retrieve relevant information efficiently. However, existing memory frameworks often perform query-agnostic retrieval over the full memory embedding space even when their storage layer is backed by efficient vector indexes such as HNSW. This full-scope retrieval path creates latency bottlenecks as memory grows, hindering real-time agent interactions. We propose SwiftMem, a query-aware agentic memory system that narrows retrieval to query-relevant memory subsets through specialized indexing over temporal and semantic dimensions. Our temporal index enables logarithmic-time range queries for time-sensitive retrieval, while the semantic DAG-Tag index maps queries to relevant topics through hierarchical tag structures. To address memory fragmentation during growth, we introduce an embedding-tag co-consolidation mechanism that reorganizes storage based on semantic clusters to improve locality. Across LoCoMo and LongMemEval$_S$, SwiftMem reaches 10.8/12.7 ms search latency while maintaining competitive LLM-judge accuracy against strong HNSW-backed memory systems. On the calibrated benchmark, LoCoMo Refined, SwiftMem remains close to the top LLM-judge score while preserving an order-of-magnitude latency advantage.

cs.CL↗