Searcharxiv⌕ Search

arXiv subjects

Ming Xiao

Publications and source records attributed to Ming Xiao.

At least 19 recordsLinked to original sources
arXiv2026

Large Language Models Meet Bayesian Optimization: Towards Self-evolving Wireless Digital Twins

Wireless digital twins (DTs) are increasingly integrated with artificial intelligence to support online modeling, prediction, and network optimization. However, most existing approaches perform adaptation within predefined learning or optimization structures, which may become unsuitable when wireless tasks and operating conditions change. In this article, we investigate the integration of large language models (LLMs) and Bayesian optimization (BO) to extend wireless DTs from online adaptation towards algorithm-level self-evolution. We introduce an LLM-assisted BO framework in which BO provides a stable and sample-efficient optimization backbone, while the LLM determines whether selected BO components should be retained or adapted based on the current task and accumulated optimization experience. In simulations, we apply the framework to Gaussian process (GP) kernel discovery for a multi-user wireless beam-selection problem. Simulation results show that the LLM-discovered kernel improves the final BO performance by 19.7% on the training tasks compared with the strongest standard kernel baseline. The results demonstrate the potential of combining LLM flexibility with structured BO for self-evolving wireless DTs.

eess.SY↗
arXiv2026

NGC 3227 Blinks: An Extreme Changing-Look Event in a Seyfert Galaxy

We report a rapid optical changing-look event in the Seyfert galaxy NGC 3227: the optical continuum faded by more than 70% and broad H-beta emission nearly disappeared within weeks, before recovering within two months. Damped random walk simulations calibrated to the source's normal variability show that such a decline is unlikely for stochastic accretion fluctuations (p = 0.002). Simultaneous Swift monitoring reveals X-ray spectral hardening, pointing to enhanced line-of-sight absorption. During the dim state, reverberation mapping yields a persistent 6-8 day lag between the continuum and broad H-beta, showing that the broad-line region continued to reverberate and placing the obscurer interior to the broad-line region. Photoionization calculations require a fully ionized, Compton-thick column with N_H ~ 1.5 x 10^24 cm^-2 to reproduce the optical continuum suppression. Together, these observations demonstrate that compact, high-column-density obscurers interior to the broad-line region can produce extreme optical changing-look behavior on week timescales.

astro-ph.GA↗
arXiv2026

Resource-aware Sub-network Allocation for Heterogeneous Decentralized Learning

System heterogeneity in computation, communication, and memory can substantially limit synchronous decentralized learning, where each training round is constrained by resource-limited nodes. Existing approaches provide limited support for jointly determining node-specific model capacities from deployment resources. We propose RASA-DL, a resource-aware sub-network allocation framework that treats the maximum sub-network width assigned to each node as a deployment variable. RASA-DL combines offline hardware profiling of computation and memory costs with deployment-aware profiling of learning utility under heterogeneous decentralized training. Using these profiles together with node- and link-level resource information, we formulate deployment as a bottleneck-latency minimization problem subject to per-node memory constraints and a required aggregate learning utility. Exploiting the monotone feasibility of the latency threshold, RASA-DL efficiently identifies the deployment through a binary-search-based allocation procedure. Experiments under controlled resource heterogeneity demonstrate favorable accuracy--latency tradeoffs. Under the Moderate and Severe regimes, the minimum-latency operating point improves accuracy over the largest globally feasible uniform deployment by 2.80 and 2.43 percentage points, respectively, at the same modeled round latency. The advantage is further amplified under strong Non-IID data heterogeneity.

cs.DC↗
arXiv2026

To Defer or Not to Defer? Online Learning with LLMs and Congested Experts

Modern AI systems increasingly combine large language models (LLMs) with human experts, using the LLM to handle routine queries while deferring uncertain cases for expert review. When expert capacity is limited, however, deferral decisions must account not only for the reliability of the LLM but also for congestion in the expert queue. We formulate this as an online sequential decision-making problem in which, for each query, the system observes the confidence signal of the LLM's answer and must decide whether to output the answer or defer the query to a capacity-constrained expert. The reliability of the confidence signal is initially unknown and must be learned from expert feedback, which is both selectively observed (only deferred queries are labeled) and delayed by queuing. We propose \texttt{KL-PBD}, a learning algorithm that uses Kullback--Leibler lower confidence bounds on the unknown reliability parameters to guide deferral decisions. For $K$ confidence levels and horizon $T$, \texttt{KL-PBD} achieves $O(K\log T)$ gap-dependent regret when the accuracy at every confidence level is bounded away from the deferral boundary, and $\widetilde{O}(\sqrt{KT})$ worst-case regret. We establish matching lower bounds, showing that these rates are optimal up to logarithmic factors. Finally, we validate our approach on synthetic instances and real-world question-answering datasets, namely GSM8K, NuminaMath, and MedMCQA, using Qwen LLMs.

cs.LG↗
arXiv2026

Heterogeneity-Aware Client Sampling for Optimal and Efficient Federated Learning

Federated learning (FL) commonly involves clients with diverse communication and computational capabilities. Such heterogeneity can significantly distort the optimization dynamics and lead to objective inconsistency, where the global model converges to an incorrect stationary point potentially far from the pursued optimum. Despite its critical impact, the joint effect of communication and computation heterogeneity has remained largely unexplored, due to the intrinsic complexity of their interaction. In this paper, we reveal the fundamentally distinct mechanisms through which heterogeneous communication and computation drive inconsistency in FL. To the best of our knowledge, this is the first unified theoretical analysis of general heterogeneous FL, offering a principled understanding of how these two forms of heterogeneity jointly distort the optimization trajectory under arbitrary choices of local solvers. Motivated by these insights, we propose Federated Heterogeneity-Aware Client Sampling, FedACS, a universal method to eliminate all types of objective inconsistency. We theoretically prove that FedACS converges to the correct optimum at a rate of $O(1/\sqrt{R})$, even in dynamic heterogeneous environments. Extensive experiments across multiple datasets show that FedACS outperforms state-of-the-art and category-specific baselines by 4.3%-36%, while reducing communication costs by 22%-89% and computation loads by 14%-105%, respectively.

cs.LG↗
arXiv2026

Channel-coded Over-the-Air Computation

AirComp enables efficient wireless data aggregation, however, its computation accuracy is vulnerable to channel impairments. As a promising solution, the role of redundancy for AirComp has been largely unexplored, leaving a critical gap in reliable AirComp design. To address this, we propose a channel-coding scheme tailored for AirComp that preserves the aggregation structure while mitigating channel impairments through bandwidth expansion. We show that the computation error decreases with the coding rate and can vanish. Both theoretical and simulation results demonstrate that the proposed scheme significantly improves computation performance.

cs.IT↗
arXiv2026

Reliability assessment and multicenter clinical application of magnetic resonance methods for knee cartilage quantification

Background: This study evaluated interreader agreement and longitudinal performance of MRI methods for knee cartilage volume, thickness, and defect-area quantification. Methods: AI-presegmented masks from 1,189 phase III examinations underwent independent correction by two readers and adjudication. Cartilage volume, three-dimensional ray-tracing thickness (3D-RT), and ray-based defect area (3D-RBA), defined by a 1.5-mm thickness threshold, were calculated. Agreement was assessed using segmentation metrics, intraclass correlation coefficients (ICCs), repeated-measures Bland-Altman analysis, and minimal detectable change at 95% confidence (MDC95). The 3D-RBA framework was evaluated in 120 digital-phantom experiments from 40 participants. Longitudinal analyses included 374 participants, alternative-method comparisons included 65, and retrospective phase II analysis included 24 participants with four visits. Results: Overall AI-to-adjudicated-mask Dice was 0.964 +/- 0.029. Interreader ICCs for volume, thickness, and defect area were 0.956, 0.904, and 0.932; corresponding MDC95 values were 1,596.9 mm^3, 0.227 mm, and 147.4 mm^2. Geometric mean absolute percentage error for defect area was 5.62%, with spatial Dice of 0.961. In 374 participants, volume changes correlated positively with thickness changes (rho=0.431) and negatively with defect-area changes (rho=-0.221). Within-participant phase II correlations followed the same directions in both groups. Conclusions: The workflow demonstrated good interreader agreement. Controlled geometric results and longitudinal associations supported the feasibility of threshold-based defect-area estimation. Volume, thickness, and defect area provide complementary measures of cartilage structure.

eess.IV↗
arXiv2026

Componentwise rigidity of holomorphic isometric maps between bounded symmetric domains with negative conformal factors

Let $D$ be an irreducible bounded symmetric domain of complex dimension at least two. We study local holomorphic maps from $D$ into products of irreducible bounded symmetric domains satisfying a metric identity with positive and negative conformal factors. Under a noncancellation condition on these factors, we prove that every nonconstant component extends to a holomorphic isometric embedding up to a positive integer factor. This proves a componentwise rigidity conjecture formulated by Cheng--Hao--Yuan--Zhang.

math.CV↗
arXiv2026

CoCurve: Cross-Module Co-Pruning Curvature for Structured LLM Pruning

Resource-constrained deployment requires sustaining large language model (LLM) capabilities as models scale under fixed memory and computation budgets. Structured pruning advances this deployment frontier with smaller dense checkpoints, yet deciding what to prune remains bottlenecked by interactions among joint removals. We introduce Cross-Module Co-Pruning Curvature (CoCurve), which formulates structured pruning as set-dependent predictive risk over a unified inventory of attention heads and feed-forward groups. A co-pruning graph built from single-unit forward ablations assigns individual risk to nodes and reinforcement or cancellation to interaction edges, conditioning each decision on the units already removed. We evaluate 6 LLMs (3B--70B) and 3 vision--language models (VLMs) across 3 perplexity corpora, 12 language tasks, and 7 multimodal benchmarks. Across the five-model 20--40% grid and the 70B 10--50% sweep, CoCurve ranks first in 53/60 corpus comparisons (15/15 at 70B); matched-quality interpolation permits 2.2--6.6 more pruning points in 9/10 cases, while CoCurve leads all 6 VLM Avg$_7$ blocks. After the same lightweight recovery, its retained structures remain strongest through 50% pruning, where the 8B checkpoint regains 10.8 Avg$_{12}$ points; physical slicing delivers $1.58\times$ dense prefill throughput with 41% lower peak memory. Mechanism analysis across 10 LLMs and 7 VLMs finds organized within- and cross-module edge structure; matched low-saliency, high-coupling removals degrade 19/20 capability groups by up to 23.7 points.

cs.LG↗
arXiv2026

State of Thought Enables Endogenous Reasoning

Test-time compute has emerged as a major approach to improving the capabilities of Large Language Models (LLMs). However, existing test-time reasoning paradigms rely heavily on externally imposed control, either through fixed reasoning programs or through costly expansion in constrained search spaces, limiting both generalization and efficiency. We propose State of Thought (SoT), a new reasoning paradigm that enables endogenous reasoning in LLMs, with the model's internal reasoning state governing how reasoning unfolds. Concretely, SoT extracts a compact dynamics-geometric state from the model's internal information transfer and uses a 582-parameter controller on frozen backbones to selectively activate historical reasoning support useful under the current reasoning state, framing reasoning as a state-conditioned process over evidence rather than an externally prescribed token chain. Across quantitative (1.34x), general (1.62x), symbolic-and-code (1.76x), and long-context (2.51x) reasoning on 3 LLMs and 16 datasets, SoT consistently improves mean-baseline accuracy while reducing generated tokens by 62.6% and end-to-end latency by 44.6%. Across 2 VLM scales and 3 reasoning tasks, it improves mean accuracy by 3.8 points over reasoning baselines, with 74.9% fewer completion tokens and 73.5% lower latency than search-based methods. Under constrained access, SoT retains 38.2%/36.5% mean accuracy gains in training-free/embedding-only settings, while trajectory-only judging reaches 84.1% agreement across 3 API models. Together, endogenous state-driven reasoning provides a generalizable and efficient alternative.

cs.CL↗
arXiv2026

Conditional Co-Ablation: Recovering Self-Repair Backups in Transformer Circuits

Mechanistic interpretability seeks to explain transformer behavior through circuits: sets of internal components that causally support a behavior. However, self-repair creates a blind spot: ablating a primary component can activate a dormant backup, so a circuit that explains behavior in the intact model can become incomplete under the intervention used to test it. We formulate this gap as conditional circuit completion: given a primary set, identify components that become causally important after its removal. We introduce conditional co-ablation (CoAx), which ranks candidates by growth in ablation effect after primary-set removal. We show that a perfectly dormant backup can be indistinguishable from an irrelevant component to per-unit intact-state scores, whereas its conditional effect change exactly aggregates all interaction orders linking it to the removed set. On GPT-2-small's Indirect Object Identification (IOI) circuit, CoAx recovers the documented backup heads at 0.941 ROC-AUC, versus 0.815 for the strongest intact-state attribution baseline and 0.758 for the matched conditional-energy control. Recovery drops to 0.40 +/- 0.13 AUC for alternative component sets matched in behavioral effect, output displacement, and depth, showing that recovery is specific to the removed circuit. Beyond recovery, the CoAx-selected heads are causally load-bearing: freezing them after primary removal sharply reduces the IOI margin, while adding them to the incomplete circuit reduces incompleteness from 0.75 to 0.21. More broadly, conditional growth aligns with intervention-derived repair in 11/12 held-out instances across 4 mechanism clusters, and CoAx completions outperform matched random completions on all 8 non-GPT-2 models spanning 6 architecture families. Together, causal explanations of self-repairing transformers must account for backup circuitry when primary components fail.

cs.LG↗
arXiv2026

Degree-Three Rational Sphere Maps: Sharp Denominator Region and Gram Normal Forms

We study degree-three rational sphere maps in two complex variables. After a standard normalization, the denominator of such a map takes the form \[ g_σ(z)=1+σ_1 z_1^2+σ_2 z_2^2, \qquad σ_1,σ_2\geq 0. \] A basic question is: which pairs $(σ_1,σ_2)$ can actually occur as the denominator of a degree-three rational sphere map? The first main result of the paper gives a complete answer: such a denominator occurs if and only if \[ 0\leq σ_1,σ_2<1, \qquad \sqrt{1-σ_1^2}+\sqrt{1-σ_2^2}>1. \] Our approach converts the sphere-mapping condition into a finite-dimensional Gram-matrix positivity problem. Furthermore, for each admissible parameter $ σ=(σ_1,σ_2), $ we determine all possible minimal target dimensions in which the corresponding denominator $g_σ$ can be realized. We also give a Gram-matrix normal form for maps with a fixed denominator and compute, for each admissible $σ$, the dimension of the moduli space of equivalence classes of rational sphere maps realizing $g_σ$. Finally, we extend the Gram-matrix method to arbitrary source dimension and obtain a general sufficient condition for the existence of degree-three rational sphere maps.

math.CV↗
arXiv2026

A new strong rigidity phenomenon for the Bergman metric

We establish a new local-to-global rigidity phenomenon for the Bergman metric. Namely, under natural geometric hypotheses, a local conformal identification of Bergman metrics determines the underlying complex manifold globally, up to the unavoidable ambiguity of removing Bergman-negligible subsets. More precisely, let $Ω\subseteq\mathbb C^n$ be a bounded domain with a complete Bergman metric, and suppose that the Bergman metric of a complex manifold $M$ is locally conformal, via a holomorphic map $f$, to that of $Ω$. We prove that the given local map $f$ extends to a biholomorphism $F\colon M\to D$ onto a subdomain $D\subseteqΩ$ in two complementary settings. If $M$ is Stein, then $Ω\setminus D$ is a closed pluripolar set. If $M$ is a bounded domain and $Ω$ satisfies a natural symmetry condition expressed in terms of its automorphism orbits, then $Ω\setminus D$ is Bergman-negligible. In particular, this applies when $Ω$ is a bounded homogeneous domain and yields a characterization, up to Bergman-negligible sets, of bounded domains with locally symmetric Bergman metrics. The latter answers a question raised by Loi--Palmieri and Zimmer. A key ingredient in the proof is a new Calabi-type extension theorem tailored to Bergman metrics.

math.CV↗
arXiv2026

Resource-efficient Semantic Coding Schemes with Manifold-constrained Hyper-connections

Semantic communication (SemCom) and task-oriented communication (TOC) can reduce wireless resource consumption by focusing on transmitting semantic or task-relevant information instead of raw messages. In practice, a main challenge is to make transmitting information robust to channel noise and fading while keeping it compact. Existing learning-based transceivers often improve reliability by using larger encoders or higher-dimensional channel features, which increase computation complexity and channel uses. Therefore, optimized system design needs explicit rate control to balance performance and transmitting resources e.g., bandwidth and power. For this purpose, we propose a manifold-constrained hyper-connection (mHC) coding scheme with an entropy bottleneck (EB) for resource-efficient SemCom and TOC over wireless channels. Instead of using a single residual path of existing encoders, the proposed mHC-based semantic encoder applies multiple residual streams and constrains their interaction by doubly stochastic (DS) mixing matrices. The new structure improves representation diversity and training stability with negligible parameter and floating-point overhead. The EB quantizes the channel features and estimates the entropy-coded rate, enabling end-to-end rate--distortion/task optimization under bandwidth and transmit-power constraints. We further show that DS-constrained stream mixing does not increase the differential entropy of the transmitted features. This implies no increase in the ideal EB coding length. Experiments on SemCom and TOC under additive white Gaussian noise (AWGN), Rayleigh fading, Rician fading, and imperfect channel state information (CSI) show that the proposed scheme improves semantic/task performance, communication robustness, and convergence stability over residual and unconstrained HC baselines, while requiring no additional channel uses.

cs.IT↗
arXiv2026

Coding-Enforced Robust Secure Aggregation for Federated Learning Under Unreliable Communication

This work studies privacy-preserving federated learning (ppFL) under unreliable communication. In ppFL, zero-sum privacy noises enables privacy protection without sacrificing model accuracy, effectively overcoming the privacy-utility trade-off. However, in practice, unreliable communication can randomly disrupt the coordination of zero-sum noises, leading to aggregation errors and unpredictable partial participation, which severely harm the model accuracy and learning performance. To overcome these challenges, we propose a robust coding-enforced structured secure aggregation method, termed secure cooperative gradient coding (SecCoGC), which enables exact reconstruction of the global model under unreliable communication while allowing for arbitrarily strong privacy preservation. In this paper, a complete problem formulation and constructions of real-field zero-sum privacy noise are presented, and fairness is introduced as a privacy metric. Privacy across all protocol layers in SecCoGC is evaluated, accounting for the correlation among privacy noises and their linear combination under unreliable communication. Moreover, a distinct convergence analysis for the FL algorithm with a binary outcome for global model recovery is provided. Experimental results demonstrate that SecCoGC achieves strong resilience to unreliable communication while maintaining varying levels of privacy preservation, yielding test accuracy improvements of up to 20%-70% over existing benchmark methods.

cs.IT↗
arXiv2026

LingxiDiagBench: A Multi-Agent Framework for Benchmarking LLMs in Chinese Psychiatric Consultation and Diagnosis

Mental disorders are highly prevalent worldwide, but the shortage of psychiatrists and the inherent subjectivity of interview-based diagnosis create substantial barriers to timely and consistent mental-health assessment. Progress in AI-assisted psychiatric diagnosis is constrained by the absence of benchmarks that simultaneously provide realistic patient simulation, clinician-verified diagnostic labels, and support for dynamic multi-turn consultation. We present LingxiDiagBench, a large-scale multi-agent benchmark that evaluates LLMs on both static diagnostic inference and dynamic multi-turn psychiatric consultation in Chinese. At its core is LingxiDiag-16K, a dataset of 16,000 EMR-aligned synthetic consultation dialogues designed to reproduce real clinical demographic and diagnostic distributions across 12 ICD-10 psychiatric categories. Through extensive experiments across state-of-the-art LLMs, we establish key findings: (1) although LLMs achieve high accuracy on binary depression--anxiety classification (up to 92.3%), performance deteriorates substantially for depression--anxiety comorbidity recognition (43.0%) and 12-way differential diagnosis (28.5%); (2) dynamic consultation often underperforms static evaluation, indicating that ineffective information-gathering strategies significantly impair downstream diagnostic reasoning; (3) consultation quality assessed by LLM-as-a-Judge shows only moderate correlation with diagnostic accuracy, suggesting that well-structured questioning alone does not ensure correct diagnostic decisions. We release LingxiDiag-16K and the full evaluation framework to support reproducible research at https://github.com/Lingxi-mental-health/LingxiDiagBench.

cs.AI↗
arXiv2026

The Kalman Evolve: Closing the Gap in Kalman Filtering via Interpretable Algorithm Discovery

State estimation is a fundamental problem in control and signal processing, for which the Kalman Filter provides an optimal solution under linear dynamics, Gaussian noise, and known noise covariances. However, these assumptions often fail in realistic sensing settings such as Doppler radar and LiDAR. In these cases, the optimal estimator is inherently nonlinear, which leads to systematic performance degradation. This creates a performance gap that cannot be eliminated by tuning the noise covariance parameters (i.e., the process and measurement noise in the Kalman Filter) alone. To address this limitation, we propose Kalman Evolve, a framework for discovering improved filtering algorithms by jointly optimizing both noise parameters and the update structure. Our approach leverages large language models (LLMs) as a structured prior over program space, enabling the generation of interpretable, non-affine modifications to the classical Kalman filter while preserving its recursive form. We provide analytical results establishing the suboptimality of affine estimators under common nonlinear sensing models, motivating the need for structure-aware updates. Across a range of synthetic and real-world tracking benchmarks, including Doppler radar, LiDAR-based localization, and pedestrian tracking, the discovered algorithms consistently improve over strong baselines such as the Optimized Kalman Filter, achieving up to 12\% reduction in RMSE. These results suggest that optimizing the structure of the Kalman filter, rather than only its parameters, provides a practical and interpretable way to improve state estimation.

cs.LG↗
arXiv2026

Abundance of Bergman metrics with constant positive holomorphic sectional curvature

An outstanding open question, which has attracted renewed attention following the pioneering work of Huang--Li--Treuer, is whether, for a given positive integer $m$, there exists a complex manifold whose Bergman metric is locally isometric to $m$ times the Fubini--Study metric. Previously, this question had only been resolved in the case $m=1$. In this paper, we construct, for any pair of positive integers $(m,n)$ with $n \geq 2$, an $\mathbb{R}$-parameter (hence uncountable) family of Reinhardt domains in $\mathbb{C}^n$ whose Bergman metrics are all locally isometric to $m$ times the Fubini--Study metric. Moreover, we show that the domains in this family are mutually Bergman inequivalent. This not only answers the folklore question, but also suggests that a reasonable classification of the geometry of such complex manifolds is infeasible. We also note such examples cannot exist in dimension one. The results complete the remaining open case in the study of complex manifolds whose Bergman space separates points and whose Bergman metric has constant holomorphic sectional curvature. Our approach differs from existing methods in the literature. We reduce the construction to a mapping problem and apply a Brouwer fixed point argument to establish the existence of the desired domains.

math.CV↗