Searcharxiv⌕ Search

arXiv subjects

Wei Zhao

Publications and source records attributed to Wei Zhao.

At least 55 records · Page 3Linked to original sources
arXiv2026

Yuvion LLM: An Adversarially-Aware Large Language Model for Content And AI Safety

As large language models are increasingly deployed in real-world systems, safety failures can still lead to harmful outputs and dangerous misuse. We argue that the essence of safety is adversarial: many failures arise not from natural inputs alone, but from strategic attempts to evade model policies and safeguards. However, existing general-purpose model development largely overlook this adversarial nature, and often remain insufficient for realistic safety scenarios involving planning, tool use, and multi-step reasoning, causing measured safety performance to overestimate real deployment robustness. To address this gap, we present Yuvion LLM, a large language model built for adversarially robust content safety and broader AI safety. Yuvion LLM treats adversarial robustness and agentic capability as first-class objectives. Its pipeline combines adversarially aware data construction, knowledge-enhanced continued pretraining, and policy-grounded multi-task safety post-training, including risk-aware supervised fine-tuning and reinforcement learning-based policy optimization, together with safety-aware agentic reinforcement learning for tool use and multi-step reasoning in complex safety scenarios. We further introduce the Yuvion LLM RiskEval (YLRE), a collection of 93 benchmarks across four evaluation categories, covering diverse open and internal evaluations with a focus on safety, adversarial robustness, and real-world capability requirements. Across these evaluations, Yuvion LLM demonstrates clear advantages on safety-focused benchmarks and particularly strong robustness under adversarial conditions, while maintaining solid overall capability. Notably, Yuvion-8B outperforms most state-of-the-art baselines, including substantially larger models such as GPT-5.4 and Qwen3-MAX, on several safety tasks.

cs.CL↗
arXiv2026

RISE: Relay Inference and Online Scheduling for Efficient Edge-Device Collaborative Diffusion Model Services

Text-to-image diffusion models are increasingly deployed at the network edge to serve heterogeneous workloads with diverse quality and latency requirements. However, existing deployment strategies choose either large edge-side models with high fidelity but high latency or lightweight device-side models that offer speed at the cost of semantic coherence. Moreover, these approaches rarely split the denoising workload between models of different sizes across edge servers and user devices. To bridge this gap, we propose RISE, a method for edge-device diffusion model services that combines relay inference with online scheduling. Driven by the finding that the latent intensity exhibits minimal deviation after a model handoff, RISE uses a training-free relay mechanism that exploits the shared latent space within a model family: the large model on the edge handles the early denoising steps that shape semantic structure, then passes the intermediate latent to a small device-side model for detail refinement. To deploy this mechanism as a practical service, a contextual bandit scheduler selects the best relay configuration based on prompt complexity, user preferences, network quality and real-time node loads. Experiments on two benchmarks show that RISE's relay mechanism achieves up to 2.1$\times$ speedup while preserving full-model quality, and its context-aware scheduler effectively balances quality and latency under mixed workloads.

cs.DC↗
arXiv2026

Looped World Models

Current world models face a fundamental tension: faithful long-horizon simulation demands deep computation, but deeper models are expensive to deploy and prone to compounding errors. We resolve this by introducing Looped World Models (LoopWM), which are the first looped architectures for world modelling. Our method iteratively refines latent environment states through a parameter-shared transformer block. This yield up to 100x parameter efficiency over conventional approaches with adaptive computation that automatically scales depth to match the complexity of each prediction step. Orthogonal to scaling model size and training data, LoopWM establishes iterative latent depth as a new scaling axis for world simulation, which might significantly push the community forward.

cs.LG↗
arXiv2026

CogGuard: Cognitive and Operational Profiling for Proactive Warning in Edge Intelligent Services

Proactive warning is an important capability for edge intelligent services, where the system predicts whether a subject will successfully complete an incoming task under strict latency and privacy constraints. Such prediction depends on both long-term static attributes and short-term dynamic states derived from historical interaction logs. Recent Large Language Models (LLMs) offer strong long-context reasoning for constructing structured profiles from these logs, but existing solutions face two challenges for edge deployment: (1) profiling methods are typically domain-specific and lack a reusable abstraction across service scenarios, and (2) fine-tuning alignment models on heterogeneous edge clusters incurs high synchronization overhead due to the variance in input sequence lengths. To address these challenges, we propose CogGuard, a proactive-warning framework for edge intelligent services. CogGuard decouples offline LLM-based profile construction from online Small Language Model (SLM)-based score prediction through a shared static-dynamic profile-to-score pipeline, and instantiates it in two representative scenarios: educational performance warning and operational task outcome warning. For efficient profile construction, we design scenario-specific profiling methods with prefix-aligned KV-cache reuse to reduce repeated encoding overhead. For edge-side model alignment, we propose a length-aware distributed fine-tuning strategy with contrastive regularization to mitigate workload imbalance on heterogeneous clusters. Experiments on education and operation datasets show that CogGuard reduces profile construction time by up to 48% and distributed fine-tuning time by 19%, while achieving MAEs of 13.4 and 5.9, respectively, on 100-point-scale warning tasks. In the largest educational setting, CogGuard reduces prediction error by 15.4% compared with the strongest baseline.

cs.AI↗
arXiv2026

Intelligent Perception Assisted Light Modulation for real-time and program-free nanofabrication and particle manipulation

In this paper, we present an Intelligent Perception-Assisted Light Modulation (IPALM) technique, for femtosecond laser nanofabrication and particle manipulation. IPALM technique integrates real-time hand-motion recognition with dynamic spatial light modulation to achieve programming-free laser beam control. In contrast to the conventional programmed laser fabrication techniques which need setup a project prior to fabrication, IPALM offers a direct "mind-to-matter" pathway for laser nanostructuring and biological cells handling with high flexibility and multiple degree of freedom. In nanofabrication, IPALM provides high resolution with feature dimensions down to 280 nm. By mind-driven hand gesture control, the laser beam is regulated to enable direct writing of micro/nanostructures with a minimum feature size down to 280 nm via IPALM. In precise particle manipulation, multiple cells can be simultaneously moved to achieve cell coalescence. With these examples, IPALM showcases its potential for applications in photonics, biomedicine, and microfluidics, for high-dimension and flexible laser applications.

physics.optics↗
arXiv2026

A symmetric Gauss-Seidel based alternating proximal ALM for generalized Nash Equilibrium problems in Banach spaces

In this paper, we study a class of monotone generalized Nash equilibrium problems (GNEPs) with jointly linear constraints. The players' strategy spaces are real Hilbert spaces, while the joint constraint is formulated in a Banach space. To solve such problems, we propose a novel symmetric Gauss-Seidel (sGS) based alternating proximal augmented Lagrangian method (sGS-APALM) which incorporates newly designed quadratic surrogates. In contrast to existing regularization and ALM-type methods, the proposed method avoids solving coupled Nash equilibrium subproblems at each iteration and instead updates the players' strategies alternately by solving a sequence of unconstrained quadratic programs. Moreover, unlike many existing splitting-based methods, our global convergence analysis and convergence rate estimation require only monotonicity and Lipschitz continuity of the pseudo-gradient mapping, without imposing stronger assumptions such as strong monotonicity or co-coercivity. Finally, we apply the method to a class of risk-neutral PDE-constrained GNEPs with joint state constraints, and preliminary numerical results demonstrate its efficiency and effectiveness.

math.OC↗
arXiv2026

EviRank: Evidence-Based Confidence Estimation for LLM-Based Ranking

Large Language Models show promise for recommendation, but they raise reliability concerns due to limited domain coverage and inherent stochasticity. Existing uncertainty quantification methods persist two fundamental challenges: (1) the global confidence score designed for question answering fails to reveal which positions are unreliable in ranking list; (2) fine-grained confidence extracted from model internals exhibits uniformly low values across all positions, making it impossible to filter unreliable predictions. To tackle the challenges, we propose an evidence-based confidence estimation for LLM-based ranking (EviRank). We extract three complementary evidences from a single forward pass and aggregate them via reliable opinion aggregation. Furthermore, we recognize that ranking positions are inherently unequal, and introduce a position-aware calibration. Lastly, the calibrated confidence guides ranking optimization. Experiments on three datasets demonstrate that our method achieves state-of-the-art performance on both recommendation and uncertainty quantification.

cs.IR↗
arXiv2026

On the Scaling of PEFT: Towards Million Personal Models of Trillion Parameters

Parameter-efficient fine-tuning (PEFT) is usually treated as a cheaper alternative to full fine-tuning. We study a broader role: small trainable adapters as persistent local state on top of strong shared foundation models. In this framing, the base model provides shared competence while adapters carry instance-specific behavior such as preferences, skills, tool habits, and memory-like updates. We organize the problem around three scaling axes: Scale Up, where stronger shared priors make small local updates more useful; Scale Down, where we study how small adapters can be while remaining reliable; and Scale Out, where many persistent adapted instances coexist. MinT provides one infrastructure example for managing adapter identity, revision, provenance, evaluation, and serving residency. Together, the results suggest that PEFT can be a compact substrate for persistent personal models rather than only a budget substitute for full fine-tuning.

cs.LG↗
arXiv2026

Target-Agnostic Calibration under Distribution Shift with Frequency-Aware Gradient Rectification

Real-world model deployments inevitably encounter distribution shifts, rendering the confidence estimates of deep neural networks highly unreliable, posing severe risks in safety-critical applications. Existing methods improve calibration via training-time regularization or post-hoc adjustment, but often rely on access to (or simulation of) target domains, limiting practicality. We propose Frequency-aware Gradient Rectification (FGR), a target-agnostic training framework for robust calibration. From a frequency perspective, FGR applies low-pass filtering to a subset of training images to diminish spurious high-frequency cues and encourage the learning of domain-invariant features. However, the associated information loss can degrade In-Distribution (ID) calibration. To resolve this trade-off, FGR treats ID calibration as a hard constraint and rectifies conflicting parameter updates via geometric projection. This ensures a first-order non-increase in the ID calibration objective without introducing an additional loss-balancing coefficient. Extensive experiments on synthetic, real-world, and semantic shift datasets demonstrate that FGR significantly improves calibration under diverse shifts while preserving ID performance, and it remains compatible with post-hoc calibration methods. Our code is available at https://github.com/YilinZhang107/FGR-Calib.

cs.CV↗
arXiv2026

Graph-Enhanced Policy Optimization in LLM Agent Training

Multi-step LLM agents in interactive environments represent a crucial step toward long-horizon decision-making. To train such agents, group-based reinforcement learning is widely adopted, which reinforces trajectories with higher relative performance within the group. However, in most existing methods, every step within a trajectory and every trajectory with the same terminal reward receive identical credit, regardless of their actual contributions. Since different states play different structural roles in an online state-transition graph built from sampled trajectories, their impacts should be differentiated and converted into task-aware credit at both the step and trajectory levels. We therefore present Graph-Enhanced Policy Optimization (GEPO), a framework for dual-level structural credit assignment in multi-step LLM agent training. Specifically, GEPO derives a state-level Task-Conditioned Criticality score that combines topological betweenness on the state-transition graph with semantic similarity to the task prompt. Based on this score, trajectory-level credit is reshaped through a state-adaptive discount, while step-level credit is scaled by the criticality of its successor state. Experimental results show that GEPO outperforms the strongest baselines by 1.1\% in success rate on ALFWorld, 3.2\% on WebShop, and 3.8\% on average across search-augmented QA tasks at the 7B scale. Compared with flat group-based methods, GEPO reduces across-seed variance and concentrates gradient signals on the most critical steps.

cs.AI↗
arXiv2026

MinT: Managed Infrastructure for Training and Serving Millions of LLMs

We present MindLab Toolkit (MinT), a managed infrastructure system for Low-Rank Adaptation (LoRA) post-training and online serving. MinT targets a setting where many trained policies are produced over a small number of expensive base-model deployments. Instead of materializing each policy as a merged full checkpoint, MinT keeps the base model resident and moves exported LoRA adapter revisions through rollout, update, export, evaluation, serving, and rollback, hiding distributed training, serving, scheduling, and data movement behind a service interface. MinT scales this path along three axes. Scale Up extends LoRA RL to frontier-scale dense and MoE architectures, including MLA and DSA attention paths, with training and serving validated beyond 1T total parameters. Scale Down moves only the exported LoRA adapter, which can be under 1% of base-model size in rank-1 settings; adapter-only handoff reduces the measured step by 18.3x on a 4B dense model and 2.85x on a 30B MoE, while concurrent multi-policy GRPO shortens wall time by 1.77x and 1.45x without raising peak memory. Scale Out separates durable policy addressability from CPU/GPU working sets: a tensor-parallel deployment supports 10^6-scale addressable catalogs (measured single-engine sweeps through 100K) and thousand-adapter active waves at cluster scale, with cold loading treated as scheduled service work and packed MoE LoRA tensors improving live engine loading by 8.5-8.7x. MinT thus manages million-scale LoRA policy catalogs while training and serving selected adapter revisions over shared 1T-class base models.

cs.LG↗
arXiv2026

When Medical Safety Alignment Fails: A Benchmark for Evaluating LLMs on High-Risk Medical Queries

Large language models (LLMs) are increasingly used for medical and health-related questions, yet their safety in high-risk medical scenarios remains poorly understood. We introduce \textsc{MedHarm}\footnote{Code and data will be released upon acceptance. Due to the sensitive nature of high-risk medical queries, data access will be available to qualified researchers upon request.}, a high-risk medical safety benchmark with 1,100 medically grounded queries across 10 safety-critical categories, including toxicology, pharmacology, covert poisoning, anesthesia, and fetal harm. Unlike broad medical QA benchmarks, \textsc{MedHarm} targets realistic clinical, educational, and technical prompts that require refusal, caution, or safe redirection rather than direct helpfulness. We evaluate 15 LLMs spanning general-purpose, medical-purpose, closed-source, and downstream SFT models, together with 4 representative guardrail models. Results reveal a substantial gap between apparent alignment and medical safety: aligned models can still produce unsafe or actionable responses, medical fine-tuning can amplify harmful specificity, and external guardrails reduce some failures while introducing brittle blocking and weak safe helpfulness. These findings show that medical safety cannot be inferred from general alignment or medical capability alone, highlighting the need for domain-specific stress testing before deploying LLMs in safety-critical medical applications.

cs.CY↗
arXiv2026

MDGMIX: Boundary-Aware Subgraph Mixing for Multi-Domain Graph Pre-Training

Multi-domain graph pre-training is a crucial step in constructing foundational graph models with cross-domain generalization capabilities. However, existing methods predominantly rely on jointly training all source domain graphs, resulting in high computational costs. Furthermore, it remains unclear whether all source domain graph data contribute equally to effective transfer. This paper empirically reveals significant data redundancy in multi-domain graph pre-training. Based on this finding, we propose the Multi-domain Graph Pre-training Framework, MDGMIX, which combines boundary-aware subgraph mixing with hierarchical discrimination. By selecting boundary nodes to construct challenging mixed-domain subgraphs, MDGMIX employs coarse-grained domain discrimination and fine-grained domain decomposition losses to decouple shared patterns from domain-specific patterns. During adaptation, MDGMIX employs a lightweight prompt weighting mechanism to transfer source domain knowledge. Extensive experiments demonstrate that MDGMIX consistently outperforms strong baselines in few-shot classification tasks while exhibiting superior time and memory efficiency. The code is available at: https://github.com/zhengziyu77/MDGMIX.

cs.LG↗
arXiv2026

Design and Initial Test of the Weighing Unit for Tsinghua Tabletop Kibble Balance

This paper presents a customized weighing unit developed for the Tsinghua tabletop Kibble balance. The system is based on a flexure hinge mechanism sourced from a commercial weighing cell, with a major modification to the feedback control loop. The redesigned loop incorporates a capacitive displacement sensor for high-resolution position detection and a novel PID control strategy that ensures both fast dynamic response and high static stability. Initial characterization results demonstrate a repeatability better than 0.1 mg in air for 1 kg mass exchanges, validating the system's potential for high-accuracy mass metrology in Kibble balances.

physics.ins-det↗
arXiv2026

Design and Verification of a New Current Source for Tsinghua Tabletop Kibble Balance

This paper presents an ultra-stable current source tailored for the one-mode, two-phase (OMTP) measurement scheme in Tsinghua tabletop Kibble balances. To achieve simultaneous high resolution and nA/A-level stability, a composite 'coarse-fine' control topology is proposed, utilizing a dual-DAC architecture and an active digital feedback loop. Experimental results show that the Allan deviation reaches 1 nA/A at an integration time of approximately 3 minutes, representing a tenfold improvement in measurement speed compared to commercial-source-based setups. Furthermore, the design offers a significant cost advantage, providing a satisfying option for high-precision, cost-effective mass realization.

physics.ins-det↗
arXiv2026

An Optical System for Monitoring Coil Parasitic Motion and Mass Position for Tsinghua Tabletop Kibble Balance

This paper presents a novel seven-channel optical measurement system for monitoring coil parasitic motion and mass position in the Tsinghua Tabletop Kibble balance. The system employs seven spectrally-confocal displacement sensors arranged in a distributed configuration to simultaneously measure the coil's translational ($x_{\rm{c}}, y_{\rm{c}}$), rotational ($θ_x,θ_y$) degrees of freedom, and the mass position offset ($x_{\rm{m}}, y_{\rm{m}}$) due to corner errors. Three vertically oriented sensors target an equilateral triangle target rigidly connected to the coil, enabling real-time calculation of tilt angles through geometric relationships. Two horizontally oriented sensors measure the translational displacement of a frame target on the coil assembly. Two additional horizontal sensors monitor the mass position to quantify corner errors. The initial experimental setup has been completed, featuring sufficient resolution and minimal signal loss, providing a new approach for alignment adjustment and corner error compensation in high-precision Kibble balances.

physics.ins-det↗
arXiv2026

ClawGuard: A Runtime Security Framework for Tool-Augmented LLM Agents Against Indirect Prompt Injection

Tool-augmented Large Language Model (LLM) agents have demonstrated impressive capabilities in automating complex, multi-step real-world tasks, yet remain vulnerable to indirect prompt injection. Adversaries exploit this weakness by embedding malicious instructions within tool-returned content, which agents directly incorporate into their conversation history as trusted observations. To address these vulnerabilities, we introduce \textsc{ClawGuard}, a novel runtime security framework that enforces a user-confirmed rule set at every tool-call boundary, transforming unreliable alignment-dependent defense into a deterministic, auditable mechanism that intercepts adversarial tool calls before any real-world effect is produced. By automatically deriving task-specific access constraints from the user's stated objective prior to any external tool invocation, \textsc{ClawGuard} blocks all three injection pathways without model modification or infrastructure change. Experiments across five state-of-the-art language models on six injection benchmarks covering web, local, MCP, and skill channels, as well as three utility benchmarks covering OS, web, and code tasks, demonstrate that \textsc{ClawGuard} achieves robust protection against indirect prompt injection without compromising agent utility or introducing significant token overhead. This work establishes deterministic tool-call boundary enforcement as an effective defense mechanism for secure agentic AI systems. Code is publicly available at github.com/Claw-Guard/ClawGuard/.

cs.CR↗
arXiv2026

Empowering Heterogeneous Graph Foundation Models via Decoupled Relation Alignment

While Graph Foundation Models (GFMs) have achieved remarkable success in homogeneous graphs, extending them to multi-domain heterogeneous graphs (MDHGs) remains a formidable challenge due to cross-type feature shifts and intra-domain relation gaps. Existing global feature alignment methods (PCA or SVD) enforce a shared feature space blindly, which distorts type-specific semantics and disrupts original topologies, inevitably leading to "Type Collapse" and "Relation Confusion". To address these fundamental limitations, we propose Decoupled relation Subspace Alignment (DRSA), a novel, plug-and-play relation-driven alignment framework. DRSA fundamentally shifts the paradigm by decoupling feature semantics from relation structures. Specifically, it introduces a dual-relation subspace projection mechanism to coordinate cross-type interactions within a shared low-rank relation subspace explicitly. Furthermore, a feature-structure decoupled representation is designed to decompose aligned features into a semantic projection component and a structural residual term, adaptively absorbing intra-domain variations. Optimized via a stable alternating minimization strategy based on Block Coordinate Descent, DRSA constructs a well-calibrated, structure-aware latent space. Extensive experiments on multiple real-world benchmark datasets demonstrate that DRSA can be seamlessly integrated as a universal preprocessing module, significantly and consistently enhancing the cross-domain and few-shot knowledge transfer capabilities of state-of-the-art GFMs. The code is available at: https://github.com/zhengziyu77/DSRA.

cs.SI↗