arXiv subjects

Jiaxu Li

Publications and source records attributed to Jiaxu Li.

At least 19 recordsLinked to original sources

arXiv2026

Bridging the Detection-to-Abstention Gap in Reasoning Models under Insufficient Information

We highlight a failure mode of large reasoning models on questions with insufficient information: models may recognize that a problem is under-specified, yet still continue reasoning and produce unsupported final answers instead of abstaining. We formalize this mismatch as the detection-to-abstention gap, where detected insufficiency fails to translate into final abstention. This gap is especially concerning in high-risk domains such as medical AI, where answers based on incomplete evidence can be more harmful than refusal. To close this gap, we propose Judge-Then-Solve (JTS), a trajectory-level reasoning-control framework that trains models to make an explicit answerability commitment before solution generation. Rather than treating abstention as a final-answer style, JTS casts it as a control decision: the model either proceeds to solve or terminates early based on its answerability judgment. We instantiate this policy through supervised warm-up and missing-premise reinforcement learning with consistency and length-shaping rewards. Experiments on dense and MoE reasoning models show that JTS substantially improves reliable abstention across datasets and pushes Abstention@Detection (A@D) to near-saturation, indicating that models not only detect missing information but also act on that detection. By terminating unanswerable trajectories immediately after the answerability judgment, JTS reduces unnecessary reasoning and improves inference efficiency when continued deliberation would amplify unsupported assumptions. We also observe that missing-premise training can alter reasoning behavior on difficult but answerable problems, reducing unproductive self-reflection. These results suggest that abstention under insufficient information is a key form of reasoning control for deploying reasoning models safely and efficiently.

cs.AI↗

arXiv2026

DeepAFL: Deep Analytic Federated Learning

Federated Learning (FL) is a popular distributed learning paradigm to break down data silo. Traditional FL approaches largely rely on gradient-based updates, facing significant issues about heterogeneity, scalability, convergence, and overhead, etc. Recently, some analytic-learning-based work has attempted to handle these issues by eliminating gradient-based updates via analytical (i.e., closed-form) solutions. Despite achieving superior invariance to data heterogeneity, these approaches are fundamentally limited by their single-layer linear model with a frozen pre-trained backbone. As a result, they can only achieve suboptimal performance due to their lack of representation learning capabilities. In this paper, to enable representable analytic models while preserving the ideal invariance to data heterogeneity for FL, we propose our Deep Analytic Federated Learning approach, named DeepAFL. Drawing inspiration from the great success of ResNet in gradient-based learning, we design gradient-free residual blocks in our DeepAFL with analytical solutions. We introduce an efficient layer-wise protocol for training our deep analytic models layer by layer in FL through least squares. Both theoretical analyses and empirical evaluations validate our DeepAFL's superior performance with its dual advantages in heterogeneity invariance and representation learning, outperforming state-of-the-art baselines by up to 5.68%-8.42% across three benchmark datasets.

cs.LG↗

arXiv2026

Strong solutions to the initial-boundary-value problem of compressible MHD equations with degenerate viscosities and far field vacuum in 3D exterior domains

This paper concerns the initial-boundary-value problem (IBVP) of the compressible Magnetohydrodynamic (MHD) equations in 3D exterior domains with Navier-slip boundary conditions for the velocity and perfect conducting conditions for the magnetic field. For the case that the density approaches far-field vacuum initially and the viscosities are power functions of the density (ρ}δ with 0 < δ < 1), the local existence and uniqueness of strong solutions to the IBVP is established for regular large initial data. In particular, in contrast to the local theory of compressible Navier-Stokes equation Li-Lü-Yuan [24], we show that the magnetic field maintains the initial quality of decaying faster rate than density throughout the time evolution, which reveals the role of the magnetic field in handling singularities arising from density-dependent viscosities.

math.AP↗

arXiv2026

Local well-posedness of strong solutions to the compressible Navier-Stokes equations with degenerate viscosities and far field vacuum in 3D exterior domains

The isentropic compressible Navier-Stokes system subject to the Navier-slip boundary conditions is considered in a general three-dimensional exterior domain. For the density approaches far-field vacuum initially and the viscosities are power functions of the density(ρ^δ with 0 < δ< 1), the local well-posedness of strong solutions is established in this paper. In particular, the method we adopt can not only simultaneously handle the difficulties caused by boundary terms and far-field vacuum, but also make the selection of δ independent of the gas coefficient γ.

math.AP↗

arXiv2026

TeleMem: Building Long-Term and Multimodal Memory for Agentic AI

Large language models (LLMs) excel at many NLP tasks but struggle to sustain long-term interactions due to limited attention over extended dialogue histories. Retrieval-augmented generation (RAG) mitigates this issue but lacks reliable mechanisms for updating or refining stored memories, leading to schema-driven hallucinations, inefficient write operations, and minimal support for multimodal reasoning.To address these challenges, we propose TeleMem, a unified long-term and multimodal memory system that maintains coherent user profiles through narrative dynamic extraction, ensuring that only dialogue-grounded information is preserved. TeleMem further introduces a structured writing pipeline that batches, retrieves, clusters, and consolidates memory entries, substantially improving storage efficiency, reducing token usage, and accelerating memory operations. Additionally, a multimodal memory module combined with ReAct-style reasoning equips the system with a closed-loop observe, think, and act process that enables accurate understanding of complex video content in long-term contexts. Experimental results show that TeleMem surpasses the state-of-the-art Mem0 baseline with 19% higher accuracy, 43% fewer tokens, and a 2.1x speedup on the ZH-4O long-term role-play gaming benchmark.

cs.CL↗

arXiv2025

Revisiting the Data Sampling in Multimodal Post-training from a Difficulty-Distinguish View

Recent advances in Multimodal Large Language Models (MLLMs) have spurred significant progress in Chain-of-Thought (CoT) reasoning. Building on the success of Deepseek-R1, researchers extended multimodal reasoning to post-training paradigms based on reinforcement learning (RL), focusing predominantly on mathematical datasets. However, existing post-training paradigms tend to neglect two critical aspects: (1) The lack of quantifiable difficulty metrics capable of strategically screening samples for post-training optimization. (2) Suboptimal post-training paradigms that fail to jointly optimize perception and reasoning capabilities. To address this gap, we propose two novel difficulty-aware sampling strategies: Progressive Image Semantic Masking (PISM) quantifies sample hardness through systematic image degradation, while Cross-Modality Attention Balance (CMAB) assesses cross-modal interaction complexity via attention distribution analysis. Leveraging these metrics, we design a hierarchical training framework that incorporates both GRPO-only and SFT+GRPO hybrid training paradigms, and evaluate them across six benchmark datasets. Experiments demonstrate consistent superiority of GRPO applied to difficulty-stratified samples compared to conventional SFT+GRPO pipelines, indicating that strategic data sampling can obviate the need for supervised fine-tuning while improving model accuracy. Our code will be released at https://github.com/qijianyu277/DifficultySampling.

cs.CV↗

arXiv2025

Field-Free Superconducting Diode Enabled by Geometric Asymmetry and Perpendicular Magnetization

The superconducting diode effect (SDE)- manifested as directional, dissipationless supercurrents - is pivotal for realizing energy-efficient superconducting logic and memory technologies. Achieving high-efficiency SDE without external magnetic fields, however, remains a fundamental challenge. Here, we report a strongly enhanced, field-free SDE in Pt/Co/Nb heterostructures, enabled by the interplay of engineered geometric asymmetry and stray fields from a perpendicularly magnetized Co layer. This configuration promotes directional vortex entry and spatially selective pinning, yielding diode efficiencies that exceed all previously reported field-free values. Temperature- and field-dependent transport measurements, supported by micromagnetic simulations, reveal that the enhanced nonreciprocity stems from three cooperative mechanisms: asymmetric vortex entry, localized magnetic pinning, and Lorentz-force imbalance. These findings establish a scalable, CMOS-compatible platform for high-performance superconducting rectifiers, offering new opportunities for cryogenic spintronics and quantum electronics.

cond-mat.supr-con↗

arXiv2025

Weak-anti-localization-to-spin-dependent scattering at a proximity-magnetized heavy metal interface

A change in a materials electrical resistance with magnetic field (magnetoresistance) results from quantum interference effects and, or spin-dependent transport, depending on materials properties and dimensionality. In disordered conductors, electron interference leads to weak localization or anti-localization; in contrast, ferromagnetic conductors support spin-dependent scattering, leading to giant magnetoresistance (GMR). By varying the thickness of Au between 4 and 28 nm in a EuS/Au/EuS spin-switches, we observe a crossover from weak anti-localization to interfacial GMR. The crossover is related to a magnetic proximity effect in Au due to electron scattering at the insulating EuS interface. The proximity-induced exchange field in Au suppresses weak anti-localization, consistent with Maekawa-Fukuyama theory. With increasing Au thickness, GMR emerges along with spin Hall magnetoresistance. These findings demonstrate spin transport governed by interfacial exchange fields, building a framework for spintronic functionality without metallic magnetism.

cond-mat.mtrl-sci↗

arXiv2025

Is-NeRF: In-scattering Neural Radiance Field for Blurred Images

Neural Radiance Fields (NeRF) has gained significant attention for its prominent implicit 3D representation and realistic novel view synthesis capabilities. Available works unexceptionally employ straight-line volume rendering, which struggles to handle sophisticated lightpath scenarios and introduces geometric ambiguities during training, particularly evident when processing motion-blurred images. To address these challenges, this work proposes a novel deblur neural radiance field, Is-NeRF, featuring explicit lightpath modeling in real-world environments. By unifying six common light propagation phenomena through an in-scattering representation, we establish a new scattering-aware volume rendering pipeline adaptable to complex lightpaths. Additionally, we introduce an adaptive learning strategy that enables autonomous determining of scattering directions and sampling intervals to capture finer object details. The proposed network jointly optimizes NeRF parameters, scattering parameters, and camera motions to recover fine-grained scene representations from blurry images. Comprehensive evaluations demonstrate that it effectively handles complex real-world scenarios, outperforming state-of-the-art approaches in generating high-fidelity images with accurate geometric details.

cs.GR↗

arXiv2025

CAT: Concept-level backdoor ATtacks for Concept Bottleneck Models

Despite the transformative impact of deep learning across multiple domains, the inherent opacity of these models has driven the development of Explainable Artificial Intelligence (XAI). Among these efforts, Concept Bottleneck Models (CBMs) have emerged as a key approach to improve interpretability by leveraging high-level semantic information. However, CBMs, like other machine learning models, are susceptible to security threats, particularly backdoor attacks, which can covertly manipulate model behaviors. Understanding that the community has not yet studied the concept level backdoor attack of CBM, because of "Better the devil you know than the devil you don't know.", we introduce CAT (Concept-level Backdoor ATtacks), a methodology that leverages the conceptual representations within CBMs to embed triggers during training, enabling controlled manipulation of model predictions at inference time. An enhanced attack pattern, CAT+, incorporates a correlation function to systematically select the most effective and stealthy concept triggers, thereby optimizing the attack's impact. Our comprehensive evaluation framework assesses both the attack success rate and stealthiness, demonstrating that CAT and CAT+ maintain high performance on clean data while achieving significant targeted effects on backdoored datasets. This work underscores the potential security risks associated with CBMs and provides a robust testing methodology for future security assessments.

cs.CV↗

arXiv2025

APFL: Analytic Personalized Federated Learning via Dual-Stream Least Squares

Personalized Federated Learning (PFL) has presented a significant challenge to deliver personalized models to individual clients through collaborative training. Existing PFL methods are often vulnerable to non-IID data, which severely hinders collective generalization and then compromises the subsequent personalization efforts. In this paper, to address this non-IID issue in PFL, we propose an Analytic Personalized Federated Learning (APFL) approach via dual-stream least squares. In our APFL, we use a foundation model as a frozen backbone for feature extraction. Subsequent to the feature extractor, we develop dual-stream analytic models to achieve both collective generalization and individual personalization. Specifically, our APFL incorporates a shared primary stream for global generalization across all clients, and a dedicated refinement stream for local personalization of each individual client. The analytical solutions of our APFL enable its ideal property of heterogeneity invariance, theoretically meaning that each personalized model remains identical regardless of how heterogeneous the data are distributed across all other clients. Empirical results across various datasets also validate the superiority of our APFL over state-of-the-art baselines, with advantages of at least 1.10%-15.45% in accuracy.

cs.LG↗

arXiv2025

Global strong solution of the 3D compressible liquid crystal flows with density-dependent viscosity and large velocity

This paper concerns the Cauchy problem of three-dimensional compressible liquid crystal flows with density-dependent viscosity. When the viscosity coefficients $μ_1(ρ),μ_2(ρ)$ are power functions of the density with the power larger than $1$, it is proved that the system exists a unique global strong solution as long as the initial density is sufficiently large and $L^3$-norm of the derivative of the initial director is sufficiently small. This is the first result concerning the global strong solution for three-dimensional compressible liquid crystal flows without smallness of velocity.

math.AP↗

arXiv2025

Highly integrated broadband entropy source for quantum random number generators based on vacuum fluctuations

In this work, we designed and experimentally verified a highly integrated broadband entropy source for a quantum random number generator (QRNG) based on vacuum fluctuations. The core of the entropy source is a hybrid laser-and-silicon-photonics chip, which is only 6.3 $ \times $ 2.6 $ \times $ 1.5 mm$^{3}$ in size. A balanced homodyne detector based on cascaded radio-frequency amplifiers in the entropy source achieves a 3 dB bandwidth of 2.4 GHz and a common-mode rejection ratio above 25 dB. The quantum-to-classical-noise ratio is 9.51 dB at a photoelectron current of 1 mA. The noise equivalent power and equivalent transimpedance are 8.85$\,\text{pW}/\sqrt{\text{Hz}}$ , and 22.8 k$Ω$, respectively. After optimization using equalizer technology that eliminates the dependence of adjacent samples, the quantum random number generation rate reaches 67.9 Gbps under average conditional minimum entropy and 61.9 Gbps under the worst-case conditional minimum entropy. The developed hybrid chip enhances the integrability and speed of QRNG entropy sources based on vacuum fluctuations.

quant-ph↗

arXiv2025

TS-ACL: Closed-Form Solution for Time Series-oriented Continual Learning

Time series classification underpins critical applications such as healthcare diagnostics and gesture-driven interactive systems in multimedia scenarios. However, time series class-incremental learning (TSCIL) faces two major challenges: catastrophic forgetting and intra-class variations. Catastrophic forgetting occurs because gradient-based parameter update strategies inevitably erase past knowledge. And unlike images, time series data exhibits subject-specific patterns, also known as intra-class variations, which refer to differences in patterns observed within the same class. While exemplar-based methods fail to cover diverse variation with limited samples, existing exemplar-free methods lack explicit mechanisms to handle intra-class variations. To address these two challenges, we propose TS-ACL, which leverages a gradient-free closed-form solution to avoid the catastrophic forgetting problem inherent in gradient-based optimization methods while simultaneously learning global distributions to resolve intra-class variations. Additionally, it provides privacy protection and efficiency. Extensive experiments on five benchmark datasets covering various sensor modalities and tasks demonstrate that TS-ACL achieves performance close to joint training on four datasets, outperforming existing methods and establishing a new state-of-the-art (SOTA) for TSCIL.

cs.LG↗

arXiv2025

Geometric Asymmetry-Enhanced Nonreciprocal Supercurrent Transport Revealed by Second-Harmonic Response

Nonreciprocal transport in superconducting systems serves as a powerful probe of symmetry-breaking mechanisms, with the superconducting diode effect emerging as a key manifestation enabling cryogenic rectification. While theoretical models have extensively explored superconducting nonreciprocity, experimental verification remains challenging, as conventional transport measurements struggle to disentangle intrinsic and extrinsic contributions. Nonlinear transport analysis, particularly second-harmonic response, offers an alternative approach by providing a sensitive probe for detecting spatial inversion symmetry breaking in the presence of time-reversal symmetry violation. Here, we systematically investigate the influence of geometric symmetry on nonreciprocal transport by comparing two triangular-extended Hall bar configurations with a symmetric Hall bar control. Second-harmonic nonlinear transport measurements reveal that the triangular extension significantly enhances nonreciprocal response, exhibiting a clear dependence on the base angle of the extension. These findings establish a direct connection between mesoscopic geometry and macroscopic nonreciprocity, demonstrating how spatial symmetry and vortex dynamics govern nonlinear transport. This insight offers a guiding principle for designing superconducting rectification architectures.

cond-mat.supr-con↗

arXiv2025

CFSSeg: Closed-Form Solution for Class-Incremental Semantic Segmentation of 2D Images and 3D Point Clouds

2D images and 3D point clouds are foundational data types for multimedia applications, including real-time video analysis, augmented reality (AR), and 3D scene understanding. Class-incremental semantic segmentation (CSS) requires incrementally learning new semantic categories while retaining prior knowledge. Existing methods typically rely on computationally expensive training based on stochastic gradient descent, employing complex regularization or exemplar replay. However, stochastic gradient descent-based approaches inevitably update the model's weights for past knowledge, leading to catastrophic forgetting, a problem exacerbated by pixel/point-level granularity. To address these challenges, we propose CFSSeg, a novel exemplar-free approach that leverages a closed-form solution, offering a practical and theoretically grounded solution for continual semantic segmentation tasks. This eliminates the need for iterative gradient-based optimization and storage of past data, requiring only a single pass through new samples per step. It not only enhances computational efficiency but also provides a practical solution for dynamic, privacy-sensitive multimedia environments. Extensive experiments on 2D and 3D benchmark datasets such as Pascal VOC2012, S3DIS, and ScanNet demonstrate CFSSeg's superior performance.

cs.CV↗

arXiv2025

Formulation a new SIR model with non-local mobility

In this manuscript, we develop a mobility-based Susceptible-Infectious-Recovered (SIR) model to elucidate the dynamics of pandemic propagation. While traditional SIR models within the field of epidemiology aptly characterize transitions among susceptible, infected, and recovered states, they typically neglect the inherent spatial mobility of particles. To address this limitation, we introduce a novel dynamical SIR model that incorporates nonlocal spatial motion for three distinct particle types, thereby bridging the gap between epidemiological theory and real-world mobility patterns. This paper primarily focuses on analyzing the long-term behavior of this dynamic system, with specific emphasis on the computation of first and second moments. We propose a new reproduction number $R_0^m$ and compare it with the classical reproduction number $R_0$ in the traditional SIR model. Furthermore, we rigorously examine the phenomenon of intermittency within the context of this enhanced SIR model. The results contribute to a more comprehensive understanding of pandemic spread dynamics, considering both the interplay between disease transmission and population mobility and the impact of spatial motion on the system's behavior over time.

math.DS↗

arXiv2024

Global strong solution of the 3D inhomogeneous liquid crystal flows with density-dependent viscosity and large velocity

This paper concerns the initial boundary value problem of three-dimensional inhomogeneous incompressible liquid crystal flows with density-dependent viscosity. When the viscosity coefficient $μ(ρ)$ is a power function of the density with the power larger than $1$, that is $μ(ρ)=μρ^α$ with $α>1$, it is proved that the system exists a unique global strong solution as long as the initial density is sufficiently large and $L^3$-norm of the derivative of the initial director is sufficiently small. This is the first result concerning the global strong solution for three-dimensional inhomogeneous liquid crystal flows without smallness of velocity.

math.AP↗