arXiv subjects

Tao Shen

Publications and source records attributed to Tao Shen.

At least 19 recordsLinked to original sources

arXiv2026

NetPTR: Optimal Differentially Private Spectral Community Detection on Sparse Networks

Spectral community detection estimates latent labels from the leading eigenspace of a network adjacency matrix, but releasing the resulting labels can disclose sensitive relational information. We consider this problem under differential privacy for both ordinary and bipartite networks. For ordinary networks, the protected unit is a single edge, leading to edge differential privacy (edge-DP). For bipartite networks, the inferential target is the community structure of the left-side nodes, while the protected unit is an entire right-side incidence profile, leading to column-node-DP. We propose NetPTR, a private spectral clustering procedure that releases a noisy empirical spectral embedding after a stability test. The algorithm requires perturbation bounds for empirical eigenspaces under neighboring-network changes, which yield computable stability certificates and local sensitivity bounds. For ordinary networks, we establish edge-DP and the error bound under the degree-corrected stochastic blockmodel, which separates the non-private spectral clustering error from the additional privacy-induced error. It therefore guarantees weak consistency in sparse networks and exact recovery in moderate sparse networks. A matching lower bound shows that the required privacy budget is sharp up to logarithmic factors. We further develop a column-node-DP algorithm for bipartite networks and prove consistency under a bipartite degree-corrected block model. Simulations and real-data examples illustrate the resulting privacy--accuracy tradeoff.

cs.SI↗

arXiv2026

Efficient Propose-Test-Release for Optimal Differentially Private Estimation

Differential privacy (DP) is a rigorous framework that protects the participation of individuals in a dataset by controlling information leakage through released estimators. It brings a challenge for statisticians: DP uniformly considers all possible datasets, whereas statistical practice often downweights atypical or rare outcomes. The conceptual challenge is especially pronounced in sensitivity analysis, where atypical datasets introduces markedly high sensitivity, even for a basic estimator such as ordinary least square. Standard DP recipe adds a noise governed by this large overall sensitivity, which causes excessive loss in accuracy. We introduce an efficient Propose-Test Release (ePTR) pipeline, which tests the dataset via a user-designed Safety Lower Bound, and then probabilistically releases the estimator based on local sensitivity level. This flexible pipeline enables substantially simple DP mechanisms for many problems. To illustrate, we study basic estimators for Bayes classification, linear regression, and kernel regression. Each estimator can be highly sensitive to atypical datasets, yet admits simple ePTR-based algorithms that achieve minimax optimality. In numerical studies, these ePTR estimators demonstrate improved accuracy against popular DP baselines under privacy guarantees.

stat.ME↗

arXiv2026

Mamoda2.5: Enhancing Unified Multimodal Model with DiT-MoE

We present Mamoda2.5, a unified AR-Diffusion framework that seamlessly integrates multimodal understanding and generation within a single architecture. To efficiently enhance the model's generation capability, we equip the Diffusion Transformer backbone with a fine-grained Mixture-of-Experts (MoE) design (128 experts, Top-8 routing), yielding a 25B-parameter model that activates only 3B parameters, significantly reducing training costs while scaling up the model capacity. Mamoda2.5 achieves top-tier generation performance on VBench 2.0 and sets a new record in video editing quality, surpassing evaluated open-source models and matching the performance of current top-tier proprietary models, including the Kling O1 on OpenVE-Bench. Furthermore, we introduce a joint few-step distillation and reinforcement learning framework that compresses the 30-step editing model into a 4-step model and greatly accelerates model inference. Compared to open-source baselines, Mamoda2.5 achieves up to $95.9\times$ faster video editing inference. In real-world applications, Mamoda2.5 has been successfully deployed for content moderation and creative restoration tasks in advertising scenarios, achieving a 98% success rate in internal advertising video editing scenario.

cs.CV↗

arXiv2026

Construction of Knowledge Graph based on Language Model

Knowledge Graph (KG) can effectively integrate valuable information from massive data, and thus has been rapidly developed and widely used in many fields. Traditional KG construction methods rely on manual annotation, which often consumes a lot of time and manpower. And KG construction schemes based on deep learning tend to have weak generalization capabilities. With the rapid development of Pre-trained Language Models (PLM), PLM has shown great potential in the field of KG construction. This paper provides a comprehensive review of recent research advances in the field of construction of KGs using PLM. In this paper, we explain how PLM can utilize its language understanding and generation capabilities to automatically extract key information for KGs, such as entities and relations, from textual data. In addition, We also propose a new Hyper-Relarional Knowledge Graph construction framework based on lightweight Large Language Model (LLM) named LLHKG and compares it with previous methods. Under our framework, the KG construction capability of lightweight LLM is comparable to GPT3.5.

cs.CL↗

arXiv2026

Community Detection on Inhomogeneous Multilayer Networks with Extreme Sparsity

We study layer-specific community detection in an $L$-layer network $\{A^{(l)}\}_{l\in[L]}$ on a common set of $n$ nodes. Because modern networks are constructed from multi-modal data or with different contexts, the community labels $\pi^{(l)}\in[K]^n$ are layer-dependent and the degree heterogeneity parameters $\theta_i^{(l)}$ vary widely across nodes and layers. The inhomogeneity and extreme sparsity raise a challenge for classical community detection methods. We propose a multilayer-assisted regularized spectral method (MARS-CD) to address this challenge. For layer $l$, MARS-CD first constructs $X^{(l)}$ from the remaining layers, so that the problem is transformed into a network-with-covariates clustering problem on $(A^{(l)}, X^{(l)})$. Then we recover $\pi^{(l)}$ by NAC in Hu and Wang (2024) that allows misalignment. The key component is to construct $X^{(l)}$, where we stack regularized embeddings. Building upon this, we establish the first theoretical guarantees for the quality of $X^{(l)}$ under multilayer networks with extreme sparsity. These further lead to weak and strong consistency for recovering $\pi^{(l)}$. We further develop an optional label alignment step to interpret the shared community structure across layers. Simulations demonstrate the superior performance of our MARS-CD method. Applying MARS-CD to international food trading networks provides an interpretable product-specific community structure.

stat.ME↗

arXiv2026

Capacitive Pixelated CMOS Electronic Nose

Although some of the human senses can nowadays be replaced by low-cost electronic sensors such as microphones and image sensors, a compact low-cost electronic nose (E-nose) remains elusive. In this work, an E-nose is presented that can capacitively detect volatile organic compounds (VOCs). The E-nose consists of an array of 1024 capacitive microelectrodes on a complementary metal-oxide-semiconductor (CMOS) chip, functionalized by inkjet printing. The pixels are coated with a UV-curable ink and metal-organic frameworks (MOFs: ZIF-8, MIL-101(Cr), MIL-140A) to create chemically diverse microdomains that generate gas-specific response patterns through adsorption-driven dielectric loading. ZIF-8 exhibits the highest response to 2-butanone, whereas the UV-curable layer responds most strongly to toluene; both show low cross-sensitivity to water vapor, enabling operation under humid conditions. After calibration in pure gases, reproducible responses to controlled binary mixtures of toluene and 2-butanone are observed. The device operates at low power, combines a large 1024-pixel array with CMOS integration, and offers application-specific functionalization by inkjet printing, providing both low cost and versatility. By further extending the range of functionalization materials, the E-nose can be applied to analyze a wide variety of gases, with potential applications in safety monitoring, health, agriculture, and robotics.

physics.app-ph↗

arXiv2026

Seed2Scale: A Self-Evolving Data Engine for Embodied AI via Small to Large Model Synergy and Multimodal Evaluation

Existing data generation methods suffer from exploration limits, embodiment gaps, and low signal-to-noise ratios, leading to performance degradation during self-iteration. To address these challenges, we propose Seed2Scale, a self-evolving data engine that overcomes the data bottleneck through a heterogeneous synergy of "small-model collection, large-model evaluation, and target-model learning". Starting with as few as four seed demonstrations, the engine employs the lightweight Vision-Language-Action model, SuperTiny, as a dedicated collector, leveraging its strong inductive bias for robust exploration in parallel environments. Concurrently, a pre-trained Vision-Language Model is integrated as a Verifer to autonomously perform success/failure judgment and quality scoring for the massive generated trajectories. Seed2Scale effectively mitigates model collapse, ensuring the stability of the self-evolution process. Experimental results demonstrate that Seed2Scale exhibits signifcant scaling potential: as iterations progress, the success rate of the target model shows a robust upward trend, achieving a performance improvement of 131.2%. Furthermore, Seed2Scale signifcantly outperforms existing data augmentation methods, providing a scalable and cost-effective pathway for the large-scale development of Generalist Embodied AI. Project page: https://terminators2025.github.io/Seed2Scale.github.io

cs.RO↗

arXiv2026

Watch Wider and Think Deeper: Collaborative Cross-modal Chain-of-Thought for Complex Visual Reasoning

Multi-modal reasoning requires the seamless integration of visual and linguistic cues, yet existing Chain-of-Thought methods suffer from two critical limitations in cross-modal scenarios: (1) over-reliance on single coarse-grained image regions, and (2) semantic fragmentation between successive reasoning steps. To address these issues, we propose the CoCoT (Collaborative Coross-modal Thought) framework, built upon two key innovations: a) Dynamic Multi-Region Grounding to adaptively detect the most relevant image regions based on the question, and b) Relation-Aware Reasoning to enable multi-region collaboration by iteratively aligning visual cues to form a coherent and logical chain of thought. Through this approach, we construct the CoCoT-70K dataset, comprising 74,691 high-quality samples with multi-region annotations and structured reasoning chains. Extensive experiments demonstrate that CoCoT significantly enhances complex visual reasoning, achieving an average accuracy improvement of 15.4% on LLaVA-1.5 and 4.0% on Qwen2-VL across six challenging benchmarks. The data and code are available at: https://github.com/deer-echo/CoCoT.

cs.CV↗

arXiv2026

Multispectral UV Imaging on Capacitive CMOS Arrays Enabled by Solution-Processed Metal-Oxide Nanoparticles

Ultraviolet (UV) imagers are important for a variety of applications, such as quality inspection in the semiconductor industry, forensics and food quality inspection, but are often costly because they require dedicated semiconductor process flows. Here, an imaging chip is introduced that has been fabricated using standard 40 nm complementary metal-oxidesemiconductor (CMOS) technology. Instead of using a conventional charge-based photodetection principle, the imager uses a capacitive operation principle where UV-light causes capacitance changes via the photodielectric effect in a functionalization layer, which are measured by the underlying CMOS circuitry. This spin-coated or inkjet-printed functionalization layer consists of solution-processed, wide-bandgap, semiconducting metaloxide nanoparticles, and facilitates multispectral imaging. The sensors exhibit low noiseequivalent powers (17-138 fW Hz^-1/2) across the UV bands. Unlike conventional silicon CMOS imagers, the present capacitive-CMOS platform is inherently visible-blind, providing selective UV detection. This work positions late-functionalized capacitive-CMOS arrays as a route toward reducing the cost of UV imagers, which can lead to their more widespread implementation in consumer and low-volume application-specific products.

physics.ins-det↗

arXiv2025

You Never Know a Person, You Only Know Their Defenses: Detecting Levels of Psychological Defense Mechanisms in Supportive Conversations

Psychological defenses are strategies, often automatic, that people use to manage distress. Rigid or overuse of defenses is negatively linked to mental health and shapes what speakers disclose and how they accept or resist help. However, defenses are complex and difficult to reliably measure, particularly in clinical dialogues. We introduce PsyDefConv, a dialogue corpus with help seeker utterances labeled for defense level, and DMRS Co-Pilot, a four-stage pipeline that provides evidence-based pre-annotations. The corpus contains 200 dialogues and 4709 utterances, including 2336 help seeker turns, with labeling and Cohen's kappa 0.639. In a counterbalanced study, the co-pilot reduced average annotation time by 22.4%. In expert review, it averaged 4.62 for evidence, 4.44 for clinical plausibility, and 4.40 for insight on a seven-point scale. Benchmarks with strong language models in zero-shot and fine-tuning settings demonstrate clear headroom, with the best macro F1-score around 30% and a tendency to overpredict mature defenses. Corpus analyses confirm that mature defenses are most common and reveal emotion-specific deviations. We will release the corpus, annotations, code, and prompts to support research on defensive functioning in language.

cs.CL↗

arXiv2025

MammothModa2: A Unified AR-Diffusion Framework for Multimodal Understanding and Generation

Unified multimodal models aim to integrate understanding and generation within a single framework, yet bridging the gap between discrete semantic reasoning and high-fidelity visual synthesis remains challenging. We present MammothModa2 (Mammoth2), a unified autoregressive-diffusion (AR-Diffusion) framework designed to effectively couple autoregressive semantic planning with diffusion-based generation. Mammoth2 adopts a serial design: an AR path equipped with generation experts performs global semantic modeling over discrete tokens, while a single-stream Diffusion Transformer (DiT) decoder handles high-fidelity image synthesis. A carefully designed AR-Diffusion feature alignment module combines multi-layer feature aggregation, unified condition encoding, and in-context conditioning to stably align AR's representations with the diffusion decoder's continuous latents. Mammoth2 is trained end-to-end with joint Next-Token Prediction and Flow Matching objectives, followed by supervised fine-tuning and reinforcement learning over both generation and editing. With roughly 60M supervised generation samples and no reliance on pre-trained generators, Mammoth2 delivers strong text-to-image and instruction-based editing performance on public benchmarks, achieving 0.87 on GenEval, 87.2 on DPGBench, and 4.06 on ImgEdit, while remaining competitive with understanding-only backbones (e.g., Qwen3-VL-8B) on multimodal understanding tasks. These results suggest that a carefully coupled AR-Diffusion architecture can provide high-fidelity generation and editing while maintaining strong multimodal comprehension within a single, parameter- and data-efficient model.

cs.CV↗

arXiv2025

MirrorLimb: Implementing hand pose acquisition and robot teleoperation based on RealMirror

In this work, we present a PICO-based robot remote operating framework that enables low-cost, real-time acquisition of hand motion and pose data, outperforming mainstream visual tracking and motion capture solutions in terms of cost-effectiveness. The framework is natively compatible with the RealMirror ecosystem, offering ready-to-use functionality for stable and precise robotic trajectory recording within the Isaac simulation environment, thereby facilitating the construction of Vision-Language-Action (VLA) datasets. Additionally, the system supports real-time teleoperation of a variety of end-effector-equipped robots, including dexterous hands and robotic grippers. This work aims to lower the technical barriers in the study of upper-limb robotic manipulation, thereby accelerating advancements in VLA-related research.

cs.RO↗

arXiv2025

RealMirror: A Comprehensive, Open-Source Vision-Language-Action Platform for Embodied AI

The emerging field of Vision-Language-Action (VLA) for humanoid robots faces several fundamental challenges, including the high cost of data acquisition, the lack of a standardized benchmark, and the significant gap between simulation and the real world. To overcome these obstacles, we propose RealMirror, a comprehensive, open-source embodied AI VLA platform. RealMirror builds an efficient, low-cost data collection, model training, and inference system that enables end-to-end VLA research without requiring a real robot. To facilitate model evolution and fair comparison, we also introduce a dedicated VLA benchmark for humanoid robots, featuring multiple scenarios, extensive trajectories, and various VLA models. Furthermore, by integrating generative models and 3D Gaussian Splatting to reconstruct realistic environments and robot models, we successfully demonstrate zero-shot Sim2Real transfer, where models trained exclusively on simulation data can perform tasks on a real robot seamlessly, without any fine-tuning. In conclusion, with the unification of these critical components, RealMirror provides a robust framework that significantly accelerates the development of VLA models for humanoid robots. Project page: https://terminators2025.github.io/RealMirror.github.io

cs.RO↗

arXiv2025

FedEve: On Bridging the Client Drift and Period Drift for Cross-device Federated Learning

Federated learning (FL) is a machine learning paradigm that allows multiple clients to collaboratively train a shared model without exposing their private data. Data heterogeneity is a fundamental challenge in FL, which can result in poor convergence and performance degradation. Client drift has been recognized as one of the factors contributing to this issue resulting from the multiple local updates in FedAvg. However, in cross-device FL, a different form of drift arises due to the partial client participation, but it has not been studied well. This drift, we referred as period drift, occurs as participating clients at each communication round may exhibit distinct data distribution that deviates from that of all clients. It could be more harmful than client drift since the optimization objective shifts with every round. In this paper, we investigate the interaction between period drift and client drift, finding that period drift can have a particularly detrimental effect on cross-device FL as the degree of data heterogeneity increases. To tackle these issues, we propose a predict-observe framework and present an instantiated method, FedEve, where these two types of drift can compensate each other to mitigate their overall impact. We provide theoretical evidence that our approach can reduce the variance of model updates. Extensive experiments demonstrate that our method outperforms alternatives on non-iid data in cross-device settings.

cs.LG↗

arXiv2025

Lost in Pronunciation: Detecting Chinese Offensive Language Disguised by Phonetic Cloaking Replacement

Phonetic Cloaking Replacement (PCR), defined as the deliberate use of homophonic or near-homophonic variants to hide toxic intent, has become a major obstacle to Chinese content moderation. While this problem is well-recognized, existing evaluations predominantly rely on rule-based, synthetic perturbations that ignore the creativity of real users. We organize PCR into a four-way surface-form taxonomy and compile \ours, a dataset of 500 naturally occurring, phonetically cloaked offensive posts gathered from the RedNote platform. Benchmarking state-of-the-art LLMs on this dataset exposes a serious weakness: the best model reaches only an F1-score of 0.672, and zero-shot chain-of-thought prompting pushes performance even lower. Guided by error analysis, we revisit a Pinyin-based prompting strategy that earlier studies judged ineffective and show that it recovers much of the lost accuracy. This study offers the first comprehensive taxonomy of Chinese PCR, a realistic benchmark that reveals current detectors' limits, and a lightweight mitigation technique that advances research on robust toxicity detection.

cs.CL↗

arXiv2025

ThinkQE: Query Expansion via an Evolving Thinking Process

Effective query expansion for web search benefits from promoting both exploration and result diversity to capture multiple interpretations and facets of a query. While recent LLM-based methods have improved retrieval performance and demonstrate strong domain generalization without additional training, they often generate narrowly focused expansions that overlook these desiderata. We propose ThinkQE, a test-time query expansion framework addressing this limitation through two key components: a thinking-based expansion process that encourages deeper and comprehensive semantic exploration, and a corpus-interaction strategy that iteratively refines expansions using retrieval feedback from the corpus. Experiments on diverse web search benchmarks (DL19, DL20, and BRIGHT) show ThinkQE consistently outperforms prior approaches, including training-intensive dense retrievers and rerankers.

cs.IR↗

arXiv2025

FlowerTune: A Cross-Domain Benchmark for Federated Fine-Tuning of Large Language Models

Large Language Models (LLMs) have achieved state-of-the-art results across diverse domains, yet their development remains reliant on vast amounts of publicly available data, raising concerns about data scarcity and the lack of access to domain-specific, sensitive information. Federated Learning (FL) presents a compelling framework to address these challenges by enabling decentralized fine-tuning on pre-trained LLMs without sharing raw data. However, the compatibility and performance of pre-trained LLMs in FL settings remain largely under explored. We introduce the FlowerTune LLM Leaderboard, a first-of-its-kind benchmarking suite designed to evaluate federated fine-tuning of LLMs across four diverse domains: general NLP, finance, medical, and coding. Each domain includes federated instruction-tuning datasets and domain-specific evaluation metrics. Our results, obtained through a collaborative, open-source and community-driven approach, provide the first comprehensive comparison across 26 pre-trained LLMs with different aggregation and fine-tuning strategies under federated settings, offering actionable insights into model performance, resource constraints, and domain adaptation. This work lays the foundation for developing privacy-preserving, domain-specialized LLMs for real-world applications.

cs.CL↗

arXiv2025

Adaptive Bayesian Very Short-Term Wind Power Forecasting Based on the Generalised Logit Transformation

Wind power plays an increasingly significant role in achieving the 2050 Net Zero Strategy. Despite its rapid growth, its inherent variability presents challenges in forecasting. Accurately forecasting wind power generation is one key demand for the stable and controllable integration of renewable energy into existing grid operations. This paper proposes an adaptive method for very short-term forecasting that combines the generalised logit transformation with a Bayesian approach. The generalised logit transformation processes double-bounded wind power data to an unbounded domain, facilitating the application of Bayesian methods. A novel adaptive mechanism for updating the transformation shape parameter is introduced to leverage Bayesian updates by recovering a small sample of representative data. Four adaptive forecasting methods are investigated, evaluating their advantages and limitations through an extensive case study of over 100 wind farms ranging four years in the UK. The methods are evaluated using the Continuous Ranked Probability Score and we propose the use of functional reliability diagrams to assess calibration. Results indicate that the proposed Bayesian method with adaptive shape parameter updating outperforms benchmarks, yielding consistent improvements in CRPS and forecast reliability. The method effectively addresses uncertainty, ensuring robust and accurate probabilistic forecasting which is essential for grid integration and decision-making.

stat.AP↗