arXiv subjects

Lan Zhang

Publications and source records attributed to Lan Zhang.

At least 19 recordsLinked to original sources

arXiv2026

Impact of Gaia-Sausage-Enceladus Debris on Milky Way Escape-velocity Estimates in TNG50

The escape velocity is a key probe of the Milky Way (MW) potential, but standard estimates can be biased by unrelaxed substructures. Using MW analogs from the TNG50 cosmological simulation, we investigate the impact of the dominant, highly radial Gaia-Sausage-Enceladus (GSE) debris on escape-velocity measurements. We introduce an adaptive selection strategy that balances statistical reliability with sample size to identify the power-law decline regime of the local halo velocity distribution. Applying this strategy to 16 MW analogs with prominent GSE-like structures and continuous high-velocity tails, we find that retaining GSE-like stars yields a systematic underestimate, with a recovery fraction of $97.5^{+2.1}_{-2.9}\%$, while removing GSE-like stars alone increases the recovery fraction to $99.4^{+2.1}_{-2.5}\%$. The $\sim 2\%$ underestimation of the escape velocity corresponds to an $\sim 6\%$ underestimation of the total mass. This improvement persists in LAMOST-like magnitude and footprint selection. GSE-like debris is therefore a key contributor to the systematic underestimation of escape-velocity measurements, with direct implications for the inferred MW mass.

astro-ph.GA↗

arXiv2026

STEAM: Stable Self-Training with Elastic Matching and Adaptive Purification

Cross-view geo-localization (CVGL) aims to perform visual localization by matching drone-view images with corresponding satellite images. Existing supervised methods rely on large-scale, manually annotated cross-view image pairs, which are costly and hard to scale. Many unsupervised approaches obtain pseudo-supervision through image-generation-based warm-up or clustering. These procedures may introduce distribution bias or propagate pseudo-label errors. To address these limitations, we propose STEAM (Stable Self-Training with Elastic Matching and Adaptive Purification), a direct self-training framework for unsupervised cross-view geo-localization. It combines end-to-end representation learning with iterative pseudo-label refinement without image-generation-based warm-up or clustering. Specifically, the Stable Spatial-Aware Module constructs image descriptors through bounded nonlinear spatial weighting. Elastic Matching combines Bidirectional Top-K Soft Matching with Dynamic Threshold Filtering to accommodate many-to-one cross-view correspondences while filtering ambiguous candidates. Adaptive Purification maintains the pseudo-label repository through Confidence-Aware Updates, Age-Aware Updates, and Expired Label Removal. These operations respectively update assignments through confidence comparisons, track consecutive epochs without agreement or successful replacement, and discard entries whose age reaches a preset limit. On University-1652, STEAM achieves the highest R@1 and AP among the compared unsupervised methods in both retrieval directions. It also achieves competitive performance on SUES-200 and DenseUAV.

cs.CV↗

arXiv2026

Scratched Lenses, Shifted Depth: Passive Camera-Side Optical Attacks

Physical adversarial attacks on vision systems are typically studied via scene manipulation, such as adversarial patches or projections. A smaller body of work has examined camera-side manipulations, such as camera stickers and auxiliary optics, which introduce consistent artifacts in the captured image. However, these approaches treat the attack as an image-space perturbation produced by designed transmissive patterns, without explicitly modeling how physical imperfections interact with scene-dependent lighting and optics. In contrast, we identify a distinct and underexplored threat: passive lens-side damage that is persistent yet trigger-conditioned, producing scene-dependent optical artifacts that bias geometric inference. We instantiate this threat through Scratch-induced Lens Adversarial Streak Hijacking (SLASH), a physical-world attack caused by small scratches on a camera lens or protective cover. These scratches interact with bright point-like light sources and specular reflections to generate structured optical artifacts that systematically distort depth cues. We formulate the attack in optical space rather than pixel space, modeling the scratch pattern as a trigger-conditioned optical channel, and using the same scratch pattern across diverse viewing conditions. We evaluate SLASH in both digital and real-world settings on monocular depth estimation and monocular 3D object detection. On the nuScenes dataset, SLASH causes monocular depth models to estimate target vehicles as up to 32% closer to the camera or 36% farther away, with consistent effects extending to monocular 3D object detection. Physical experiments with both acrylic and glass covers confirm that the attack produces substantial depth errors on real camera inputs. These findings reveal a new attack surface in which benign-looking hardware imperfections act as latent, scene-triggered adversarial mechanisms.

cs.CV↗

arXiv2026

Disk tilts in Milky Way-like galaxies and their link to mergers

We investigated disk reorientation and its relationship to mergers in a sample of 93 Milky Way-mass galaxies drawn from the TNG50 simulations. From this sample, we identified 117 individual disk-tilt events with clearly defined start and end times, and we divided them into two classes based on whether the disk evolves smoothly or undergoes rapid disruption and regrowth during the tilt duration. For each tilt event, we linked the most massive merger that occurs in close temporal association with it. We focus on those tilt events occurring during phases of smooth disk evolution, which typically show tilt rates below 10 degrees per Gyr, persist for 2-9 Gyr, and are generally connected to minor mergers. Roughly 40% of these events reach tilt angles exceeding 30 degrees, most often triggered by satellites whose masses are comparable to or less massive than the galaxy's total stellar mass, merging on tangential orbits with orbital angular momentum that is retrograde or nearly perpendicular to the disk spin. The satellite consistently drives the disk tilt toward alignment with the satellite's orbital spin; satellites that came in on retrograde orbits relative to the disk spin can thus induce large tilt angles and can produce an almost orthogonal configuration between the stellar disk and the satellite's orbital plane as an intermediate phase. Owing to the comparatively slow tilt rates, the stellar disk can keep stable and continue to grow throughout the reorientation process. Consequently, galaxies that experience strong disk tilts can still maintain stellar disks that are as thin as those in systems without significant reorientation.

astro-ph.GA↗

arXiv2026

Decoupled Safety Control: A Safety-Control Algorithm for Training-Free Safety Guidance

Diffusion models achieve strong performance in text-to-image generation, but they can still produce unsafe content. Training-free safety guidance offers a practical inference-time alternative that intervenes during sampling without modifying model weights. However, many existing safeguards still rely on a coupled safety-control chain. In this chain, the same unsafe condition simultaneously specifies what should be avoided, indicates the current unsafe signal, constructs the intervention, and determines when that intervention remains active. This coupling can work when unsafe factors are explicit, but becomes less reliable once harmful and benign semantics are entangled in the current generation state, often leaving residual unsafe content or unnecessarily degrading benign-generation quality or prompt fidelity. We address this problem with Decoupled Safety Control (DSC), a plug-in safety-control algorithm for training-free safety guidance. DSC reformulates training-free safety guidance as an explicit state-dependent control process, so the executed intervention is determined from the current generation state rather than inherited directly from a single unsafe condition. Experiments on standard safety benchmarks show that DSC improves representative latent- and text-space safeguards, achieving stronger unsafe-content suppression while preserving benign-generation quality and maintaining efficient inference.

cs.CV↗

arXiv2026

A GHOST in Long-Horizon Agents: Governance Hazard from Overlooked Safety Constraints across Turns

Long-horizon agents are now playing an increasingly significant role in assisting humans with complex problem-solving. However, it is exactly their extended interaction history that introduces an underexplored execution-safety concern. Under benign interaction conditions, an agent may execute an action that violates a safety constraint specified many turns earlier. We term this failure mode Governance Hazard from Overlooked Safety Constraints across Turns (GHOST), which may cause irreversible damage. Our experiments reveal that GHOST events are not isolated cases: this failure mode, occurring precisely under benign interaction conditions, yields an occurrence rate of 11.5% on GPT-5.5. Furthermore, we theoretically show that if the residual conditional violation hazard along each safe prefix is bounded below by a non-summable sequence, the execution enters the hazard region almost surely. Leveraging this theoretical insight, we further propose STAR-Guard, a two-layer defense coupling historical semantic safety constraint restoration with pre-execution audit. STAR-Guard restores applicable safety constraints to reduce unsafe proposals, while its deterministic audit layer prevents residual violations from reaching the environment. Consistent with this two-layer design, we observe no GHOST events in our experiments under the GPT-5.5 setup.

cs.AI↗

arXiv2026

Nice Fold or Hero Call: Learning Budget-Efficient Thinking under Policy-Dependent Solvability

Large reasoning models (LRMs) improve problem solving through extended reasoning, but often misallocate test-time compute. Existing efficiency methods reduce cost by compressing reasoning traces or conditioning budget on perceived difficulty, yet read the resulting pass rate as a difficulty score, leaving the zero-return regime unmodeled. As a result, they overspend on queries beyond the model's capability while compressing hard-but-solvable ones that need deeper reasoning. In this work, we formulate adaptive reasoning as a computational investment under uncertainty, where budget follows the expected return of reasoning rather than perceived difficulty. To instantiate this principle, we propose Budget-Efficient Thinking (BET), a two-stage framework that combines behavioral cold-start with GRPO under an investment-cost-aware reward. By aligning solve-or-fold decisions with rollout-derived solvability, BET learns three behaviors: (1) short solve, answering easy queries concisely; (2) nice fold, abstaining early when continued reasoning has near-zero expected return; and (3) hero call, preserving sufficient compute for hard-but-solvable queries. Across seven benchmarks and three base models, BET reduces reasoning tokens by 54% while improving up to 3.2% accuracy, and transfers zero-shot to scientific QA and logical reasoning with comparable efficiency gains. Code is available at https://github.com/houqiii/BET

cs.AI↗

arXiv2026

The More It Says, the More You Pay: A Black-Box Audit of Provider-Side Token Inflation in LLM Services

In pay-per-token LLM services, the more a model says, the more users pay. Dishonest providers can covertly manipulate generation to inflate output tokens while largely preserving task utility. We define such manipulation as a Provider-Side Token Inflation Attack (PTIA) and instantiate five representative attacks at the query, prompt, representation, and model levels of the provider-controlled pipeline. Our experiments show that each attack increases mean output length to more than 10.2x the clean baseline, demonstrating PTIA's financial appeal and feasibility at multiple stages of generation. Yet auditing PTIA from black-box responses is difficult for users. Our key observation is PTIA saturation: an initial attack sharply lengthens output, but further strengthening or composition has much less effect. We trace this saturation to stopping behavior: an initial PTIA sharply lowers the end-of-sequence token probability, whereas further intervention lowers it only marginally. Building on this insight, we design a lightweight single-probe audit that applies a controlled lengthening intervention. Under PTIA, the probe induces far fewer additional tokens than under normal service. The audit requires neither a trusted local reference model nor historical clean responses, and its separately issued original and probed requests resemble ordinary traffic, making evasion difficult. Across four open-weight models, it achieves an average detection rate of 85.1% with false-positive rates below 2%. Across 15 real LLM API services, the audit flags 7 for PTIA-consistent behavior.

cs.CR↗

arXiv2026

Beyond Gold Standards: Epistemic Ensemble of LLM Judges for Formal Mathematical Reasoning

Statement autoformalization plays a crucial role in formal mathematical reasoning by enabling the automatic translation of natural language statements into formal languages. While recent advances using large language models (LLMs) have shown promising capability of autoformalization, methods for automatically evaluating autoformalization remain underexplored. LLM-as-a-judge presents a promising approach for automating such evaluation, however, existing methods typically employ coarse-grained and generic evaluation criteria, which limit their effectiveness for advanced formal mathematical reasoning, where quality hinges on nuanced, multi-granular dimensions. In this work, we take a step toward addressing this gap by introducing a systematic, automatic method to evaluate autoformalization tasks. The proposed method is based on an epistemically and formally grounded ensemble (EFG) of LLM judges, defined on criteria encompassing logical preservation (LP), mathematical consistency (MC), formal quality (FQ), and formal validity (FV), resulting in a transparent assessment that accounts for different contributing factors. We validate the proposed framework to serve as a proxy for autoformalization assessment within the domain of formal mathematics. Overall, our experiments demonstrate that the EFG ensemble of LLM judges is a more suitable emerging proxy for evaluation than a coarse-grained model. These findings suggest that LLM-as-judges, especially when guided by a well-defined set of atomic properties, could offer a scalable, interpretable, and reliable support for evaluating formal mathematical reasoning.

cs.CL↗

arXiv2026

Multimodal Wearable-Based Olfactory-Induced Emotion Recognition in Arousal-Valence Dimensions

Olfaction is important for emotion regulation because it acts as a non-intrusive and cognitively lightweight pathway that directly engages the brain s affective circuitry and achieves unobtrusive emotional modulation. This trait is essential for advancing practical affective computing in daily and attention-critical scenarios. However, current olfactory emotion research has two key limitations. First, it overemphasises the valence dimension while neglecting arousal. Second, it lacks multimodal datasets that synchronously capture central and peripheral physiological responses to olfactory stimuli. To address these issues, we construct a large-scale multimodal olfactory emotion dataset based on 111 subjects, in which odors are labeled in the 2D arousal-valence space and electroencephalogram (EEG), electrocardiogram (ECG), and photoplethysmography (PPG) signals synchronously recorded. Nevertheless, multimodal signals present challenges such as non-stationarity, differences in latency, and cross-modal heterogeneity. Thus, we propose a spatiotemporal-frequency hybrid fusion network (STF-HFNet), which integrates three core modules. Frequency aggregation processing learns adaptive frequency aggregation in order to model non-stationary dynamics. Reciprocal guided attention enables reciprocal bidirectional calibration for cross-modal temporal alignment without synchronisation priors. Hybrid collaborative fusion combines spatial and channel attention mechanisms to enhance cross-modal complementarity while suppressing redundant information. Extensive experiments show that STF-HFNet achieves state-of-the-art (SOTA) recognition accuracies of 88.34% on the AMIGOS dataset and 92.40% on our self-constructed dataset, and outperform the SOTA methods by 8.27% and 5.07%, respectively.

eess.SP↗

arXiv2026

ANet Patu-1: The Value of Connection in the Agent Network

The Internet taught us that the value of a network depends on \emph{how} its nodes connect: broadcast stars scale as $V\!\propto\!N$ (Sarnoff), fully-connected meshes as $N^2$ (Metcalfe), and group-forming networks as $2^{N}$ (Reed). We ask the analogous question for networks of AI agents. We model the net value of connection as a function of coordination-group size, derive from it the properties an optimal collaboration protocol must have, and introduce ANet Patu-1 -- a self-organizing consensus protocol in which the network continuously re-forms its own coalitions, adaptively riding the upper envelope of all three regimes at $O(1)$ parallel consensus rounds. To measure value without opinion-grading, we score an emergent protocol by formally specifying it and deriving its complexity, the way distributed algorithms are analyzed. Two results follow. (i)~Emergence -- a crowd of the \emph{cheapest} model, when heterogeneous, starts weak but its collective value compounds with $N$ and \emph{overtakes} a crowd of a far \emph{stronger} model that is homogeneous: a crossover that marks a scaling law for collaboration rather than for scale. (ii)~Reflexivity -- a heterogeneous network, given only its own problem and no design hints, converges on ANet Patu-1 itself, reconstructing the high-dimensional law that governs its own connective value.

cs.NI↗

arXiv2026

DiffRadar: Differentiable Physics-Aware Radar SLAM with Gaussian Fields

Radar sensing is increasingly used in mobile systems because it operates reliably under poor lighting, adverse weather, and privacy-sensitive settings where cameras and LiDAR often fail. However, most existing radar SLAM systems estimate motion through scan matching on discretized radar heatmaps, which breaks geometric continuity and fails to capture key radar sensing properties, often leading to unstable pose estimation and degraded mapping in regenerate or dynamically changing environments. We present DiffRadar, a real-time radar SLAM system that models radar observations as a differentiable, physics-aware Gaussian field rather than discrete scans. DiffRadar represents the scene as anisotropic Gaussian primitives and renders radar measurements in range-azimuth and Doppler-azimuth spaces through a differentiable radar forward model, enabling joint optimization of robot pose and scene structure directly from radar measurements. We implement DiffRadar on commodity FMCW radar hardware and evaluate it on both the public Radarize benchmark and a controlled stress-test suite that targets common radar SLAM failure modes, including corridor degeneracy, motion regime transitions, dynamic clutter, and long-horizon loop closures. DiffRadar achieves substantial reductions in trajectory error on the benchmark, with especially large gains under feature-poor corridor motion, while more than doubling map consistency and maintaining real-time performance at 70 FPS. These results show that modeling radar observations directly in the signal domain enables substantially more robust and consistent radar-only SLAM for mobile platforms.

cs.RO↗

arXiv2026

Valley-locked Optical Spin Merons in Valley Photonic Crystal Waveguides

Optical skyrmionic textures, including Néel-type skyrmions and merons, have attracted significant attention across diverse physical systems for their promising applications in ultra-precise metrology, optical information processing, and quantum technologies. However, the lack of an effective approach for their on-chip directional transport and manipulation impedes their applications in photonic integrated devices. Here, we numerically demonstrate a photonic platform that utilizes topologically protected valley edge state to achieve robust on-chip directional transport of optical spin merons. These merons originate from spin-orbit coupling within the evanescent field at the valley photonic crystal surface and exist as eigenstates of the topologically protected edge state, ensuring their robust propagation even in the interface with defects. Leveraging the valley degree of freedom of topological edge states, we further achieve valley-locked spin merons, enabling flexible control over the polarity of spin merons. By endowing spin merons with topological protection in momentum space, our work provides an approach for robust on-chip transport and manipulation of spin merons, thereby paving the way for expanding their application in photonic systems.

physics.optics↗

arXiv2026

MVP-Shapley: Feature-based Modeling for Evaluating the Most Valuable Player in Basketball

The burgeoning growth of the esports and multiplayer online gaming community has highlighted the critical importance of evaluating the Most Valuable Player (MVP). The establishment of an explainable and practical MVP evaluation method is very challenging. In our study, we specifically focus on play-by-play data, which records related events during the game, such as assists and points. We aim to address the challenges by introducing a new MVP evaluation framework, denoted as \oursys, which leverages Shapley values. This approach encompasses feature processing, win-loss model training, Shapley value allocation, and MVP ranking determination based on players' contributions. Additionally, we optimize our algorithm to align with expert voting results from the perspective of causality. Finally, we substantiated the efficacy of our method through validation using the NBA dataset and the Dunk City Dynasty dataset and implemented online deployment in the industry.

cs.GT↗

arXiv2026

Metis: Learning to Jailbreak LLMs via Self-Evolving Metacognitive Policy Optimization

Red teaming is critical for uncovering vulnerabilities in Large Language Models (LLMs). While automated methods have improved scalability, existing approaches often rely on static heuristics or stochastic search, rendering them brittle against advanced safety alignment. To address this, we introduce Metis, a framework that reformulates jailbreaking as inference-time policy optimization within an adversarial Partially Observable Markov Decision Process (POMDP). Metis employs a self-evolving metacognitive loop to perform causal diagnosis of a target's defense logic and leverages structured feedback as a semantic gradient to refine its policy, offering enhanced interpretability through transparent reasoning traces. Extensive evaluations across 10 diverse models demonstrate that Metis achieves the strongest average Attack Success Rate (ASR) among compared methods at 89.2%, maintaining high efficacy on resilient frontier models (e.g., 76.0% on O1 and 78.0% on GPT-5-chat) where traditional baselines exhibit substantial performance degradation. By replacing redundant exploration with directed optimization, Metis reduces token costs by an average of 8.2x and up to 11.4x. Our analysis reveals that current defenses remain vulnerable to internally-steered, closed-loop reasoning trajectories under the tested settings, highlighting a critical need for next-generation defenses capable of reasoning about safety dynamically during inference.

cs.LG↗

arXiv2026

When AI Persuades: Adversarial Explanation Attacks on Human Trust in AI-Assisted Decision Making

Most adversarial threats in artificial intelligence (AI) target the computational behavior of models rather than the humans who rely on them. Yet modern AI systems increasingly operate within human decision loops, where users interpret and act on model recommendations. Large Language Models (LLMs) generate fluent natural-language explanations that shape how users perceive and trust AI outputs, revealing a new attack surface at the cognitive layer: the communication channel between AI and its users. We introduce adversarial explanation attacks (AEAs), where an attacker manipulates the framing of LLM-generated explanations to modulate human trust in incorrect outputs. We formalize this behavioral threat through the trust miscalibration gap, a metric that captures the difference in human trust between benign and adversarial explanations. Using this metric as a lens, we highlight a behavioral risk where persuasive explanation framing can preserve user trust even when the underlying AI prediction is wrong. To characterize this threat, we conducted a human study with over 200 participants, systematically varying four dimensions of explanation framing: reasoning mode, evidence type, communication style, and presentation format. Our findings show that users report nearly identical trust for adversarial and benign explanations, with adversarial explanations preserving the vast majority of benign trust despite being incorrect. The most vulnerable cases arise when AEAs closely resemble expert communication, combining authoritative evidence, neutral tone, and domain-appropriate reasoning. Vulnerability is highest on hard tasks, in fact-driven domains, and among participants who are less formally educated, younger, or highly trusting of AI.

cs.AI↗

arXiv2026

Polymer-inspired mechanical metamaterials

Metamaterials benefit from unique architected patterns to achieve lightweight with exceptional mechanical properties inaccessible to conventional materials. Typical mechanical metamaterials are inspired by crystal-like lattice structures, whose closely packed frameworks often exhibit a rigid mechanical nature. Here, we present polymer-inspired metamaterials (PIMs) by programming deformation and strengthening mechanisms that mimic the mechanical roles of key constituent elements in polymer networks. By combining metamaterial programmability with polymer-inspired structures, we design crosslinking, proto-crystalline order, and entanglement in PIMs to enable macroscale strengthening mechanisms inspired by crosslink, molecular-density, and pre-stretch strengthening in polymers, expanding the metamaterial structure-property design space. This macroscale polymer-inspired programmability also suggests that PIMs could serve as a design platform incorporating the programmability strategies to achieve desired deformation and strengthening responses, holding a potential for applications in soft robotic joints and compliant connectors.

physics.app-ph↗

arXiv2026

Milky Way Dynamics Favor Dark Matter over Modified Gravity Models

Modified gravity theories such as Modified Newtonian Dynamics (MOND) and Scalar-Tensor-Vector Gravity (STVG) have been proposed as alternatives to dark matter, but decisive tests have been hindered by degeneracies between baryonic structure and gravitational laws. Here we break this degeneracy using independent, high-precision constraints: the Milky Way radial rotation curve, vertical phase-space spirals from Gaia, and a broken-exponential stellar disk. A joint reconstruction of the radial and vertical gravitational fields reveals a structural inconsistency in modified gravity -- no model can simultaneously reproduce both observations. Our results strongly disfavor MOND at $>13σ$ and STVG at $>4σ$. In contrast, dark matter halo models naturally explain the observations, providing a self-consistent test of gravity on galactic scales.

astro-ph.GA↗