Searcharxiv⌕ Search

arXiv subjects

Tao Zhang

Publications and source records attributed to Tao Zhang.

At least 73 records · Page 4Linked to original sources
arXiv2026

From Bilinear to Linear: Differentially Private Federated LoRA via Low-Dimensional Parameterization

Federated Low-Rank Adaptation (LoRA) provides an efficient solution for finetuning large language models across distributed and privacy-sensitive data. However, despite avoiding raw data sharing, federated LoRA remains vulnerable to privacy leakage through transmitted model updates. Differential privacy (DP) mitigates such leakage, but integrating DP into federated LoRA introduces two fundamental challenges: aggregation mismatch from independently averaging low-rank factors, and quadratic noise amplification when noise is injected into both factors. To address these challenges, we propose FedHSIP, a differentially private federated LoRA framework based on a unified low-dimensional parameterization. FedHSIP reformulates all LoRA parameters into a shared low-dimensional trainable vector, enabling clients to optimize and communicate only low-dimensional updates. This reformulation transforms federated LoRA from a bilinear factor aggregation problem into a unified linear parameter space, thereby eliminating aggregation mismatch and preventing the quadratic amplification of DP noise. To further handle non-IID data, we introduce a heterogeneity- and sensitivity-aware isometric projection, constructed from warm-up statistics, which groups coordinates with compatible cross-client update patterns while balancing sensitivity, update energy, and heterogeneity across the low-dimensional space. Extensive experiments on natural language understanding and generation benchmarks show that FedHSIP consistently outperforms existing federated LoRA methods under both private and non-private settings, achieving up to 3-4% improvements under differential privacy while reducing communication cost by over 80% and maintaining robustness under heterogeneous data distributions.

cs.CR↗
arXiv2026

Ego2Robot: Scalable Robot Data Synthesis from Egocentric Human Data

Learning generalizable robot manipulation policies requires large-scale and diverse demonstration data. Egocentric human manipulation videos offer rich scene and task diversity, and prior work has shown that retargeting and rendering such videos into robot-format data can yield effective per-task policies at small scale. However, whether this approach can provide pretraining benefits for vision-language-action models at scale remains unexplored. We present \textbf{Ego2Robot}, a scalable pipeline that converts egocentric human manipulation videos into robot training data through action retargeting, robot-arm visual synthesis, and multi-level quality curation. Ego2Robot supports both curated datasets and in-the-wild videos, producing 18,561 hours of robot training data spanning 15 robot morphologies, making it the largest ego-to-robot dataset to date. To evaluate generalization, we extend RoboTwin2.0 with disentangled perturbation axes covering visual appearance, scene layout, embodiment morphology, and task semantics. Experiments show that joint pretraining on Ego2Robot-synthesized and robot data consistently improves out-of-distribution generalization across multiple perturbation types, with benefits validated on real-robot deployment. Project page: https://www-ye.github.io/ego2robot_blog/

cs.RO↗
arXiv2026

Third-order nonlinear transport in a percolative two-dimensional superconductor

Percolative superconductivity frequently arises in two-dimensional van der Waals materials due to reduced dimensionality, enhanced quantum fluctuations, and complex electron-phonon interactions, providing a unique platform where normal electrons coexist with Cooper pairs. We report the observation of substantial third-order nonlinear transport in a trilayer $1T^\prime$-MoTe$_2$ superconductor within its percolative transition regime. The third-harmonic longitudinal voltage ($V_{\|}^{3ω}$) exhibits a clear cubic dependence on excitation current below a threshold, with both its magnitude and nonlinear coefficient strongly correlated with the superconducting state. This nonlinear response is semiquantitatively captured by the superconducting fluctuation within the time-dependent Ginzburg-Landau theory, where nonlinear transport arises due to fluctuating Cooper pairs. Our results demonstrate that third-order nonlinear transport serves as a sensitive probe of superconducting transitions in percolative systems and establish a foundation for exploring higher-order transport phenomena in strongly correlated systems.

cond-mat.mes-hall↗
arXiv2026

Cross-scale Interaction between Microturbulence and Fishbone in Fusion Plasmas

Global gyrokinetic simulations are performed for the first time to investigate cross-scale interactions between electromagnetic ion temperature gradient (ITG) turbulence and fishbone instability in tokamak plasmas. The investigation of fluctuation response in the multiscale simulation including both instabilities indicates a strong impact of fishbone on ITG turbulence. Detailed analysis reveals that fishbone-driven zonal radial electric fields at nonlinear saturation significantly suppress electromagnetic ITG turbulence, reducing ion thermal transport close to the neoclassical level. The simulation results agree well with experimental observations that turbulence suppression during fishbone bursts. These findings advance understanding of multiscale interactions that enhance thermal confinement in fusion plasmas.

physics.plasm-ph↗
arXiv2026

Anchoring and Steering Diffusion: Enhancing the Faithfulness of Text-to-Image Generation at Inference Time

While text-to-image diffusion models achieve impressive visual quality, they frequently struggle to maintain precise alignment with complex compositional prompts. An effective strategy is to improve the inference process of diffusion models, thereby better leveraging their pretrained priors to address misalignment. Existing training-free methods can be divided into two categories. The first category focuses on improving the randomly sampled initial noise, either performing costly search over noise pools or manipulating sampled noise without ensuring reliable semantic injection. The second category focuses on improving the denoising trajectory, lacking explicit mechanisms to timely diagnose and correct semantic errors. we propose \textbf{AnchorSteer}, a training-free framework that exerts fine-grained control over \textbf{both initialization} and \textbf{the denoising trajectory}. AnchorSteer consists of two synergistic components: \textbf{Semantic Anchoring} replaces uninformative Gaussian noise with text-aligned initializations via CLIP-based prior extraction and a novel Latent-Prior Score Distillation Sampling (LP-SDS) objective. Specifically, LP-SDS distills CLIP visual priors into the knowledge distribution of diffusion models, mitigating the domain gap between CLIP-based priors and diffusion-based priors. \textbf{Reflective Steering} transforms passive denoising with an active Think--Erase--Retouch loop that enables mid-generation self-correction. It leverages VLM-based diagnosis to detect semantic deviations and performs targeted latent refinement to suppress erroneous content and recover missing attributes. Extensive experiments on GenEval and T2I-CompBench++ demonstrate that AnchorSteer consistently outperforms existing baselines in text--image alignment while preserving high visual quality.

cs.CV↗
arXiv2026

GeoMFD: Continual Drone-View Geo-Localization with Geometry-Aware Adapter and Margin-Field Distillation

Existing drone-view geo-localization (DVGL) methods are mainly developed under a static training paradigm, where models are optimized for fixed environments with all training data available in advance. However, this paradigm is difficult to extend to real-world deployment, where drones may encounter diverse environments and require multiple environment-specific models, resulting in additional storage and model-selection costs. Directly adapting a single model to new environments also risks distorting previously learned cross-view embedding geometry and causing forgetting. To address these challenges, we formalize the continual drone-view geo-localization (C-DVGL) setting and propose GeoMFD, a geometry-aware continual adaptation method for DVGL. GeoMFD combines a cold-start bootstrapping strategy (CBS), a geometry-aware adapter (Geo-Adapter), and margin-field distillation (MFD) to balance adaptation and cross-view geometry preservation. CBS initializes a stable embedding space, Geo-Adapter enables environment adaptation through controlled residual corrections, and MFD preserves similarity margins between positive pairs and hard negatives to alleviate cross-view geometry forgetting. Extensive experiments demonstrate that GeoMFD effectively mitigates forgetting and achieves competitive performance with environment-specific DVGL methods using a single continuously updated model.

cs.CV↗
arXiv2026

Nanbeige4.2-3B: Unlocking Agentic Capabilities in a Compact Model

We present Nanbeige4.2-3B, a compact general agentic model with 3B non-embedding parameters. It delivers strong performance across code-agent, office-agent, and complex tool-use tasks while maintaining highly competitive reasoning capabilities in mathematics, coding, and science. Nanbeige4.2-3B is pretrained from scratch on 28T tokens with a Looped Transformer that reuses the layer stack to increase capacity without adding parameters. For SFT data and trajectory construction, we expand the diversity of executable environments, task assets, and agentic scaffolds through real-world deployment and large-scale synthesis. Our RL pipeline applies mixed-mode RLHF over Think and Non-Think responses to improve overall model quality and reduce failure cases, length-controlled reasoning RL to balance accuracy and reasoning efficiency, and agentic RL with outcome and process rewards to stabilize long-horizon training. Extensive evaluations show that Nanbeige4.2-3B outperforms larger models, including Qwen3.5-9B and Gemma4-12B, across diverse agentic benchmarks while remaining competitive on reasoning and alignment tasks. Performance with OpenClaw further supports its use as a compact local personal assistant.

cs.AI↗
arXiv2026

RP-OPSD: Resolution-Privileged On-Policy Self-Distillation for Multimodal Large Language Models

On-Policy Self-Distillation (OPSD) uses privileged information available only to the teacher to provide dense token-level supervision on trajectories generated by the student. However, existing methods often rely on verified solution traces, explanations generated by external models, or manually localized visual evidence, which limits their scalable application to multimodal large language models. To address this issue, we exploit the information gap between high- and low-resolution views of the same image and propose RP-OPSD (Resolution-Privileged On-Policy Self-Distillation for Multimodal Large Language Models). During training, the student policy generates on-policy trajectories from images at one-quarter of the original resolution, while the teacher policy provides supervision using the original-resolution images. By minimizing the divergence between their output distributions along the student trajectories, the student learns the predictive behavior of the teacher under high-resolution inputs, thereby strengthening its low-resolution capability and transferring the learned improvement to original-resolution inference. RP-OPSD requires neither additional human annotations nor external models to generate solution traces but only image--question pairs. Experiments on Qwen3.5-9B show that RP-OPSD achieves a 5.45\% relative improvement in average performance at the original resolution and a $1.78\times$ training speedup over OPSD. These results demonstrate that resolution differences can serve as a simple and scalable source of privileged information, providing an effective and efficient approach to on-policy self-distillation for multimodal large language models.

cs.CV↗
arXiv2026

Organizing Principles for Moiré Quantum Matter

Moiré flat bands in van der Waals bilayers are usually discussed through a small set of mechanisms associated with the $Γ$ and $K$ valleys of hexagonal crystals, and more recently with $M$-valleys systems. Here we show that this view is incomplete. The momentum-space location and effective local orbital character of the monolayer's band edge, in conjunction with the moiré symmetry and the symmetry representations of the resulting bands, provide a general set of organizing variables for the emergent low-energy moiré Hamiltonian. Applying fully relaxed first-principles calculations, band unfolding and symmetry-representation analysis to more than 600 commensurate twisted bilayers spanning all 2D lattice classes, we identify several routes to moiré quantum matter beyond the conventional single-orbital paradigm. The resulting flat bands realize trigonal, honeycomb, square, checkerboard and kagome-like Hubbard models with single-orbital, multi-orbital and multi-site Hilbert spaces; spin-orbit-coupled multi-orbital flat bands exhibit symmetry-indicated topology beyond the conventional $K$-valley setting; and nonsymmorphic moiré symmetries enforce semimetallic flat-band connectivity. Analogous quasi-one-dimensional flat-band structures are found in $M$-valley hexagonal systems and $X$-valley square or rectangular systems resulting from emergent momentum-space nonsymmorphic symmetries. Separately, coupled multi-valley manifolds with kagome-like connectivity are identified in several systems whose parent band edges lie at non-high-symmetry points. These results establish a valley-orbital-symmetry framework for connecting parent-material electronic structure to emergent moiré Hamiltonians relevant to correlated, topological and symmetry-enforced moiré phases.

cond-mat.mtrl-sci↗
arXiv2026

Breaking the Synthetic-Real Domain Shortcut for Training-Free Generative Replay-based Class Incremental Learning

Class-incremental learning (CIL) requires models to continuously acquire new knowledge while avoiding catastrophic forgetting. While exemplar replay is effective, it raises concerns regarding privacy and storage. Thus, generative replay has emerged as a viable alternative, synthesizing old data using frozen pretrained text-to-image (T2I) models without any extra training. However, we observe that directly mixing synthetic old-class data with real new-class data during incremental training leads to significant performance degradation. This issue stems from a "domain shortcut", where models rely on domain-discriminative features instead of semantic class cues. To address this, we propose DREAM ($\underline{\mathbf{D}}$omain-$\underline{\mathbf{R}}$egularized $\underline{\mathbf{E}}$xemplar-free $\underline{\mathbf{A}}$lignment $\underline{\mathbf{M}}$odel), which uses a training-free generator to synthesize old-class data and eliminates domain shortcut via subspace rectification and orthogonal projection, while reinforcing semantic alignment through real-anchored prototype regularization. Extensive experiments on 4 datasets demonstrate that DREAM outperforms existing exemplar-free CIL methods and achieves state-of-the-art performance. Our source code is available at https://github.com/Light-ZhangTao/DREAM.

cs.CV↗
arXiv2026

Learning to Sample in Variable Neighborhood Search Algorithm for Urban Cable Routing Optimization

Urban underground cable construction is essential for enhancing power grid reliability, yet the high construction costs demand systematic optimization. Constrained by road network infrastructure, this problem requires consideration of both connectivity relationships between substations and specific routing strategies along road networks, constituting a large-scale bilevel combinatorial optimization problem. Insufficient attention to routing subproblems in traditional research and simplistic algorithmic designs ill-equipped for large-scale optimization leave substantial room for advancement. To navigate the enormous combinatorial search space, we propose a learning-assisted variable neighborhood search (L-VNS) algorithm integrating four key components. First, an auxiliary task focusing on the upper-level connectivity subproblem generates high-quality initial solutions by employing hybrid genetic search for connection optimization and A* for detailed path routing. Subsequently, the algorithm iteratively refines the connectivity topology using variable neighborhood search with three complementary operators. A multi-agent deep reinforcement learning module adaptively guides probabilistic neighborhood sampling by jointly encoding upper-level connectivity patterns and lower-level routing structures, effectively exploiting problem structure. Finally, a modified A* operator re-plans lower-level paths affected by neighborhood modifications to ensure feasibility and solution completeness. Comprehensive experiments on 12 benchmark instances and 3 GIS-derived instances demonstrate the superiority of L-VNS, achieving total construction cost reductions of 0.92% to 73.72% compared to representative approaches. Ablation studies and sensitivity analyses further validate the effectiveness and robustness of the proposed algorithm.

cs.NE↗
arXiv2026

High-accuracy ultrasonic positioning of calibration sources in the Jiangmen Underground Neutrino Observatory

Precise source positioning is essential for detector calibration in large liquid scintillator detectors such as JUNO, particularly in regions where purely mechanical control is insufficient. An ultrasonic positioning system has been developed to reconstruct the three-dimensional coordinates of a calibration source without interfering with photon collection or contaminating the liquid scintillator. The method combines a sound-speed modeling based on dedicated laboratory measurements and in-detector temperature profiles, waveform-based arrival-time reconstruction, and an in-situ calibration of the effective receiver geometry using central-axis deployments. With six active receivers, central-axis positioning yields a mean error of 1.23 cm relative to the known deployment reference. For off-axis operation in the Cable Loop System calibration plane, a detector-realistic simulation that includes timing resolution, sound-speed variation, and receiver-coordinate smearing predicts a positioning uncertainty of 2.40 cm. These results demonstrate that ultrasonic positioning can provide centimetre-level source accuracy for large liquid scintillator detectors and can support off-axis calibration in JUNO-like experiments.

physics.ins-det↗
arXiv2026

Measuring and Mitigating Post-hoc Rationalization in Reverse Chain-of-Thought Generation

Reverse Chain-of-Thought Generation (RCG) synthesizes reasoning traces from query-answer pairs, but answer-visible generation can justify a pre-committed answer rather than derive it. This post-hoc rationalization creates a train-inference mismatch because student models are trained on answer-conditioned traces but must reason without answer access at inference time. We quantify this mismatch through lexical, trajectory, and probabilistic anchoring, measuring surface overlap, answer-conditioned generation dynamics, and answer recoverability from the trace, respectively. We find that semantic suppression, a seemingly intuitive mitigation, reduces lexical overlap but increases trajectory anchoring: avoiding the answer requires continually tracking it, thereby strengthening its influence on generation. We therefore propose Structural Skeleton-guided Reasoning (SSR), which replaces suppression with structural decoupling by first generating an abstract functional skeleton and then using it to guide the full reasoning trace. Anchoring analyses show that SSR reduces all three forms of answer dependence. Across in-domain and out-of-distribution benchmarks, its distilled variant, SSR-D, improves performance by up to 10 points over suppression baselines and better preserves out-of-distribution performance. Code is available at https://github.com/viniferagy/SSR.

cs.CL↗
arXiv2026

QIRF Quantum-Inspired Non-Orthogonal Function-Space Compression for 3D Gaussian Splatting

3D Gaussian Splatting (3DGS) achieves high-quality real-time rendering by representing a scene with a large collection of anisotropic Gaussian primitives. However, complex scenes often require millions of Gaussians, resulting in substantial storage and rendering costs. Existing compression methods mainly reduce redundancy through primitive-wise pruning, attribute quantization, clustering, or neural coding, while redundancy caused by strongly overlapping and non-orthogonal Gaussian basis functions remains largely unexplored. We present QIRF, a quantum-inspired non-orthogonal function-space compression method for 3D Gaussian Splatting. QIRF models neighboring Gaussian primitives as a local non-orthogonal basis and formulates primitive reduction as a subspace-aware selection problem. Specifically, an analytic Gaussian overlap matrix and a radiance-response density matrix are constructed to characterize functional redundancy and rendering relevance. Generalized eigendecomposition is then used to identify the dominant local subspace and select representative Gaussian primitives. An RRDM-based response model and detail-aware safeguarding further preserve visually important high-frequency structures under aggressive pruning. Experiments on 13 scenes from Mip-NeRF 360, Tanks and Temples, and Deep Blending show that QIRF reduces the Gaussian count and raw PLY storage by 71.7 percent on average, corresponding to approximately 3.54 times compression, while maintaining reconstruction quality comparable to 3DGS and achieving a marginal average PSNR improvement of 0.10 dB. QIRF also improves the average rendering speed over 3DGS by 34.3 percent. These results suggest that non-orthogonal function-space redundancy is an important yet underexplored source of representational redundancy in explicit Gaussian radiance fields.

cs.CV↗
arXiv2026

Both directions of Fuglede's conjecture fail in dimension two

Fuglede's conjecture asserts that a measurable set of positive and finite measure is spectral if and only if it tiles Euclidean space by translations. Counterexamples are known in every dimension $d\ge3$, whereas the one- and two-dimensional cases have remained unresolved. We construct two explicit $60$-point subsets of the rank-two finite Abelian group $\Z_{60}\times\Z_{12}$: one is a translational tile with no spectrum, and the other is spectral but does not tile. A finite-to-infinite transference principle lifts them to bounded subsets of $\R^2$ that are finite unions of unit squares. Consequently, both implications in Fuglede's conjecture fail in dimension two.

math.CO↗
arXiv2026

ShopX: A Foundation Model for Intent-to-Item Fulfillment in Agentic Shopping

The wave of AI-native applications is moving shopping beyond page- and feed-based browsing toward intent-driven experiences orchestrated by LLM agents. A common design wraps an LLM around existing search and recommendation pipelines, forcing complex intents through low-bandwidth retrieval or ranking interfaces and leaving a gap between language understanding and item-space fulfillment. Generative recommendation gives LLMs a direct item-space interface through semantic IDs (SIDs), but existing models mainly generate candidates for retrieval rather than translate flexible intents into item-space outcomes. We propose ShopX to address this bottleneck by unifying intent understanding, execution planning, and flexible SID-native item-space operations into a single foundation model. We deploy ShopX in agentic shopping workflows through a model-native item-fulfillment framework with a serving harness that defines a model-facing action protocol and exposes support surfaces for context access, catalog grounding, and state management. Within this framework, ShopX plans and composes SID-based item-space operations such as SID beam-search retrieval, listwise ranking, or product bundling. This model-centric design reduces lossy hand-offs between agent orchestration and item-space execution. To build ShopX, we design semantically recoverable, LLM-operable SIDs and a training recipe that equips a general LLM for flexible multi-turn item-space fulfillment while retaining the knowledge and instruction-following abilities needed by a shopping agent. We evaluate the ShopX framework against tool-mediated agentic systems on single- and multi-turn fulfillment tasks derived from anonymized Taobao production logs, showing that model-native fulfillment improves overall framework behavior, especially on complex or ambiguous requests.

cs.IR↗
arXiv2026

Low-Altitude ISAC With Spherical Directly-Connected Antenna Array: Performance Analysis and Beamforming Optimization

The safety development requirements of low-altitude economy (LAE) renders the robust low-altitude airspace monitoring critical important than ever before. Integrated sensing and communication (ISAC) as one of the key development trends of 6G provides potential solutions for the LAE. However, conventional antenna arrays suffer from limited three-dimensional (3D) sensing coverage and degraded angular resolution at high elevation angles. To address these challenges, this paper investigates low-altitude ISAC systems enabled by the recently proposed spherical directly-connected antenna array (DCAA). By carefully deploying multiple simple uniform planar arrays (sUPAs) over a spherical surface, without relying on any phase shifter, spherical DCAA enjoys advantages of full 3D coverage, superior and uniform angular resolution, enhanced energy-focusing and low hardware cost. In this paper, we first characterizes the sensing performance of the spherical DCAA, in terms of the sensing signal-to-noise ratio (SNR), area average probability of detection, and Cramér-Rao lower bounds (CRLBs) for both elevation and azimuth angle estimation. Then, a low-altitude ISAC optimization problem is formulated to maximize the worst-case sensing SNR over a prescribed aerial region while satisfying the communication quality-of-service requirements of ground users. To effective solve this mixed-integer non-convex problem, we develop a novel greed-based joint array selection and beamforming optimization framework. Simulation results demonstrate that spherical DCAA significantly outperforms conventional UPA in terms of sensing coverage, angle estimation accuracy, and communication-sensing SNR tradeoff, highlighting its potential for future low-altitude ISAC systems.

eess.SP↗
arXiv2026

MMAgent-R$^2$: Learning to Rerank and Reject for Agentic mRAG

Knowledge-based Visual Question Answering (KB-VQA) requires models to retrieve visual entities matching the query image from large-scale encyclopedic knowledge bases and answer related questions. Existing multimodal Retrieval Augmented Generation (mRAG) methods rely on global visual features to match candidate entities, yet when the knowledge base contains numerous visually similar entities, the retriever struggles to distinguish them, populating the candidate set with visually similar but factually mismatched distractors. Since subsequent processing steps such as noise filtering are also confined to this fixed candidate set, errors from failed retrieval inevitably propagate to the final answer. To address these challenges, we propose MMAgent-R$^2$, an agentic mRAG framework that integrates visual reranking and active rejection as its internal verification mechanism. Visual reranking directly compares query and candidate images, capturing discriminative details beyond textual descriptions to precisely identify the target entity among similar candidates; active rejection discards unreliable results and retrieves additional candidates when no confident match is found, moving beyond the fixed candidate pool. We design a composite reward function with step-level verification rewards and achieve joint optimization of external retrieval, internal verification, and answer generation via GRPO training. Experiments on InfoSeek, E-VQA, and MMhops demonstrate that \ours{} achieves state-of-the-art performance, with particularly notable advantages in challenging retrieval scenarios and complex multi-image multi-hop reasoning tasks.

cs.CV↗