arXiv subjects

Zhe Sun

Publications and source records attributed to Zhe Sun.

At least 19 recordsLinked to original sources

arXiv2026

Polished but Unresolved: Identifying Late-Stage Pressure States in Long-Horizon Tool-Use Agents

Long-horizon tool-use agents need not only to search and plan, but also to decide when to finalize. We study late-stage pressure states, in which an agent is biased toward submitting a final answer that appears complete and polished while key constraints remain unresolved. We first train a linear probe to show that this pressure state is identifiable from the agent's hidden states. Then, we use activation interventions along this pressure direction and find that shifting the hidden states changes both the pressure score and whether the agent continues tool use or submits early. Through controlled context manipulations, we further see that the pressure is mitigated by constraint clarity and action mapping. Based on these findings, we propose Probe-Sensed Pressure Relief (PSPR), a plugin that applies lightweight pressure relief direction under moderate pressure and moves to structured organization under high pressure risk. Experiments on multiple long-horizon benchmarks show that our method consistently strengthens existing agent methods.

cs.AI↗

arXiv2026

Contrastive Branch Policy Optimization

Reinforcement learning with verifiable rewards (RLVR) enables language models to learn multi-turn interaction with external tools, yet its sparse outcome rewards provide no signal for identifying which intermediate decisions are responsible for success. Branch sampling induces local comparisons among alternative continuations, but existing methods tend to conflate two distinct problems: allocating a fixed rollout budget and translating branch outcomes into token-level credit. We introduce Contrastive Branch Policy Optimization (CBPO), which disentangles these two problems and assigns a dedicated mechanism to each. Generation entropy screens candidate branch positions across the entire response, while path-level and node-level decay distribute a fixed budget across trajectories and positions to prevent exploration from collapsing onto a few paths or adjacent tokens. A parent trajectory together with the branches that share an identical token prefix forms an exact-prefix group, and the reward variation within this controlled group defines the Contrastive Branch Value (CBV), an outcome-based estimate of local decision sensitivity that rescales continuation advantages without altering their sign. When multiple nodes are selected along the same trajectory, CBPO partitions it into non-overlapping credit segments, thereby avoiding duplicated gradients on shared tokens. Requiring only outcome rewards and no process-level annotation, CBPO provides a practical solution for fine-grained credit assignment in tool-integrated agent training. Extensive experiments on ten benchmarks, including five for mathematical reasoning and five for knowledge-intensive search, show that CBPO consistently outperforms state-of-the-art policy-optimization and branch-based methods, attaining the highest macro-average accuracy in both domains and across two model scales.

cs.LG↗

arXiv2026

Two-Stage Teacher-Student Reliable Prior Learning for Robust Underwater Image Enhancement

Underwater image enhancement (UIE) aims to recover clear images from observations affected by wavelength-dependent absorption, scattering, and spatially nonuniform degradation. Although existing generative methods can handle complex degradations, severe information loss may lead to semantic drift in the restored results. To address this issue, we propose RPL-UIE, a two-stage teacher--student framework for reliable prior learning. In the teacher stage, the network learns reliable and complementary spatial priors characterizing appearance and photometric properties from paired degraded and reference images. In the student stage, the network takes only degraded images as input and learns to emulate the teacher's prior extraction capability, thereby providing more reliable restoration guidance for the enhancement process without requiring reference images at inference. To reduce the prior-learning discrepancy between the teacher and student models, we further develop Residual Prior Refinement Diffusion (RPRD) and Frequency-Aware Prior Residual Calibration (FPRC). RPRD uses the coarse priors as anchors and progressively predicts the necessary corrections in the residual space. FPRC retains stable low-frequency residual components and selectively modulates high-frequency detail residuals, producing calibrated priors to support high-quality reconstruction. Experiments on multiple UIE benchmarks demonstrate competitive restoration performance. Downstream underwater object detection and instance segmentation experiments further demonstrate the improved utility of enhanced images for visual perception, while tests on real-world data captured by a remotely operated vehicle (ROV) support the robustness and practical applicability of RPL-UIE.

eess.IV↗

arXiv2026

RRTrack: Robust and Recoverable Object 6D Pose Tracking for Dynamic Scenes

Robust object 6D pose tracking is critical for robotic systems operating in dynamic and occluded scenes. Per-frame estimators are accurate but computationally expensive, while current trackers struggle with fast motion and complete occlusion due to their reliance on continuous visibility. To address these challenges, we present RRTrack, an efficient, recoverable object 6D pose tracker that enables robust tracking through fast motion and target disappearance--reappearance. RRTrack introduces a 2D--6D closed-loop tracking strategy that integrates memory-based video object segmentation (VOS) with 6D pose refinement. The 2D branch maintains target localization, and the 6D branch verifies geometric consistency before memory updates. In addition, a DINOv2-based dual-bank template matching module is developed to recover lost targets by jointly exploiting offline synthetic templates and online observation anchors while maintaining real-time efficiency. We also introduce a synthetic RGB-D benchmark comprising three robotic scenarios with fast motion and full occlusion. Experimental results on the synthetic benchmark demonstrate that RRTrack improves equal-subset mean ADD-S AR by 66.3\% and ADD-S AUC by 65.7\% over FoundationPose while achieving 55.2 FPS. Real-world experiments further validate the robustness of RRTrack under noisy sensing conditions. Project page: https://github.com/7kevin24/RRTrack

cs.CV↗

arXiv2026

Multimodal Reward Hacking in Reinforcement Learning

Reinforcement learning (RL) is increasingly used to align multimodal large language models (MLLMs), but higher rewards do not always imply better task performance. This risk is amplified when visual evidence is evaluated by text-only or weakly grounded rewards. We study reward hacking in MLLM RL across safety VQA, chart VQA, and stress-test settings, varying reward design, data ambiguity, model scale (2B-32B), and RL algorithm (GRPO, RLOO, DAPO). We introduce Newly Rewarded Failure Rate (NRFR), which measures failures among samples whose proxy reward improves over the SFT baseline. Outcome-only rewards cause severe hacking, reaching 48.1% Reward Hacking Rate (RHR), while NRFR exceeding RHR shows that RL creates new failures rather than merely inheriting them. Scaling reduces but does not eliminate hacking: even the 32B model retains a 54.9% worse rate under outcome-only rewards, whereas answer-aware rewards improve the oracle trend at every scale. Robustness is also algorithm- and scale-dependent: GRPO is consistently most resistant, RLOO remains vulnerable, and DAPO improves substantially from 2B to 8B. Visual-evidence rewards help only with reliable verification: keyword-based checks increase hacking, while VLM-as-judge semantic verification reduces it. Overall, multimodal reward hacking is a systematic result of optimizing imperfect rewards, and robust alignment requires rewards and verifiers that remain reliable under optimization pressure.

cs.AI↗

arXiv2026

UAV-OVVIS: Unmanned Aerial Vehicles Also Need Open-Vocabulary Video Instance Segmentation

Unmanned Aerial Vehicle (UAV) videos are widely used in traffic monitoring, urban management, and emergency rescue. However, existing UAV video perception is largely limited to box-level detection and tracking over predefined categories, making it difficult to jointly support flexible queries and fine-grained instance-level understanding of temporal dynamics in open scenarios. To this end, we introduce a new task, UAV Open-Vocabulary Video Instance Segmentation (UAV-OVVIS), which aims to discover targets in UAV videos according to open-vocabulary queries and output instance segmentation trajectories with globally consistent identities. Considering the scarcity of instance-level annotations in UAV scenarios, we propose AeroTrack, a training-free framework that coordinates existing visual foundation models to realize UAV-OVVIS. AeroTrack performs target discovery and segmentation through periodic open-vocabulary detection and short-segment mask propagation, and introduces Lifecycle-aware ID Association (LIA) to recover global identities under segment-wise inference. Based on this framework, we instantiate five feasible variants and construct AeroVIS, a UAV-OVVIS evaluation benchmark containing 9 UAV object categories and 8,279 trajectories. Experiments show that AeroTrack achieves better overall performance than the evaluated OV-VIS methods transferred to AeroVIS, while demonstrating good open-vocabulary transferability and dense-target handling capability in long UAV videos. The AeroTrack framework and the AeroVIS dataset will be open-sourced upon acceptance.

cs.CV↗

arXiv2026

GUIDE: Goal-Initialized Directional Understanding for End-to-End Visual Navigation

Learning-based visual navigation for legged robots typically relies on continuous goal updates from hierarchical state estimation to provide a persistent directional reference. This reliance incurs additional sensory and computational overhead and deviates from fully end-to-end mobile autonomy. Furthermore, under partial observability, policies are prone to learn myopic behaviors, easily becoming trapped in dead ends and complex structural layouts. To address these limitations, we investigate a goal-initialized navigation setting, where the target is provided only once at the beginning of an episode, requiring the robot to operate based on intrinsic spatial memory without subsequent goal updates from external modules. In this work, we propose GUIDE, a fully end-to-end reinforcement learning framework designed to cultivate internal directional awareness. Specifically, GUIDE incorporates a spatial anchor predictor that leverages multi-frequency proprioceptive history to extract egomotion representations, thereby maintaining a persistent long-horizon spatial context for navigation. Concurrently, it utilizes raw depth streams to perceive local environmental geometry. We evaluate the proposed framework across both simulation and real-world scenarios on a quadruped robot. Experiments show that GUIDE learns reliable egomotion and directional awareness, enabling a fully end-to-end deployed policy to safely navigate through dense clutter and structured mazes without subsequent goal guidance or prior maps.

cs.RO↗

arXiv2026

BotDirector: Robot Storytelling Across the Symmetrical Reality with Multi-modal Interactions

Robot storytelling offers a unique blend of technological innovation and creative expression that engages children in unprecedented ways. However, the technical aspects are often too complicated for children. We propose an interactive system that facilitates robot storytelling with tangible and natural language interactions. Children arrange the playground with their own stuff and create narratives with an LLM agent. The created narratives are transformed into a motion sequence based on the map and characters, and the motions are executed by self-navigating swarm robots. This system enhances robot storytelling with flexible scenarios, enabling young children to create robot dramas with everyday objects.

cs.RO↗

arXiv2026

Cleavage-History-Dependent Low-Temperature ARPES Spectra of Charge-Ordered EuAl$_4$

Charge ordering in EuAl$_4$ has been widely discussed in connection with band reconstruction, magnetism, and topological electronic states, yet the microscopic origin of the complex low-temperature ARPES spectra remains unresolved. Here we combine photon-energy-, temperature-, and cleavage-history-dependent ARPES with first-principles calculations to distinguish intrinsic bulk bands from surface-preparation-dependent spectral weight. Spectra measured on high-temperature-cleaved surfaces, both at 160 K and after cooling to 10 K, are broadly consistent with the calculated three-dimensional bulk electronic structure, whereas low-temperature-cleaved surfaces exhibit additional electron-like bands, replica-like Fermi-surface contours, and a pronounced $\delta$ band near -0.57 eV that is absent from the calculated bulk bands. The additional features are observed at multiple photon energies and on multiple independently cleaved surfaces and are selectively suppressed upon warming, while the bulk-derived bands remain comparatively stable. The $\delta$ band does not emerge when the same high-temperature-cleaved surface is cooled through $T_{\rm CDW}$. Comparison with the projected bulk bands and the calculated spectral function of an ideal Eu-terminated surface further associates the additional bands with the surface electronic structure. These results establish a strong cleavage-history dependence of the low-temperature ARPES spectra and provide spectroscopic criteria for separating surface-reconstruction and bulk charge-order contributions in EuAl$_4$.

cond-mat.str-el↗

arXiv2026

Termination-Dependent Surface States and Magnetic Fingerprints of Chiral Helimagnet Cr1/3TaS2

Chiral helimagnets based on intercalated transition-metal dichalcogenides, characterized by nano-scale spin ordering, provide a powerful route to engineer chiral spin textures (e.g. the topologically protected magnetic solitons) and emergent electronic functionality at reduced dimensions, where surface and interface states often dominate device operation. However, despite growing interest, direct experimental studies of termination-dependent surface electronic structures and their temperature-driven magnetic evolution remain largely unexplored, hindering a microscopic understanding of the electronic states that is crucial for the development of low-dimensional spintronic devices. Here, for the first time, taking Cr1/3TaS2 as a representative example, we systematically investigate the termination-dependent surface electronic states of the chiral helimagnets and uncover their distinct temperature evolution across the magnetic transition (TC~142K) by combining high-resolution ARPES with a micro-focused beam and surface-state-resolved first-principles calculations. The TaS2-terminated surface hosts folded monolayer-like TaS2 bands under the $\sqrt3\times\sqrt3$ superlattice potential and a shallow triangular electron pocket at the superlattice $\bar K$ point arising from Cr-Ta orbital hybridization. In contrast, the Cr-terminated surface exhibits reconstructed hole pockets with pronounced magnetic band splitting. This splitting disappears above TC and closely follows the chiral helimagnetic order parameter, providing a direct spectroscopic fingerprint of chiral helimagnetic order. In addition, multiple ultranarrow Cr-d-derived surface flat bands are resolved. These findings establish Cr1/3TaS2 as a model system in which surface electronic states are strongly coupled to chiral magnetism, opening new opportunities for chiral spintronic and valleytronic micro/nanodevices.

cond-mat.mtrl-sci↗

arXiv2026

Compressive sensing inspired self-supervised single-pixel imaging

Single-pixel imaging (SPI) is a promising imaging modality with distinctive advantages in strongly perturbed environments. Existing SPI methods lack physical sparsity constraints and overlook the integration of local and global features, leading to severe noise vulnerability, structural distortions and blurred details. To address these limitations, we propose SISTA-Net, a compressive sensing-inspired self-supervised method for single-pixel imaging. SISTA-Net unfolds the Iterative Shrinkage-Thresholding Algorithm (ISTA) into an interpretable network consisting of a data fidelity module and a proximal mapping module. The fidelity module adopts a hybrid CNN-Visual State Space Model (VSSM) architecture to integrate local and global feature modeling, enhancing reconstruction integrity and fidelity. We leverage deep nonlinear networks as adaptive sparse transforms combined with a learnable soft-thresholding operator to impose explicit physical sparsity in the latent domain, enabling noise suppression and robustness to interference even at extremely low sampling rates. Extensive experiments on multiple simulation scenarios demonstrate that SISTA-Net outperforms state-of-the-art methods by 2.6 dB in PSNR. Real-world far-field underwater tests yield a 3.4 dB average PSNR improvement, validating its robust anti-interference capability.

cs.CV↗

arXiv2026

Multiple Topological States in LaAgAs2, a Failed Square-Net Semimetal

The rational design of new materials emerges as an important direction to explore new topological materials, which is based on the understanding of the correlation between crystal and electronic structures. In this paper, we perform a comprehensive study on the crystal and electronic structures in LaAgAs2 through a combination of single-crystal x-ray diffraction (XRD), quantum oscillation, and angle-resolved photoemission spectroscopy (ARPES) experimental measurements, and density functional theory (DFT) calculations. Single-crystal XRD measurements reveal that LaAgAs2 crystallizes into a HfCuSi2-derived structure with the square net distorted into cis-trans chains. Quantum oscillation measurements reveal two frequencies with small effective masses and quasi-two-dimensional (2D) characters. ARPES measurements reveal an electronic structure strikingly different from the square-net-based semimetals, such as LaAgAs2. The Fermi surface is quasi-two-dimensional (2D), with Dirac-like hole pockets at the zone center and a quasi-1D elliptical electron pocket at the zone boundary. Based on the DFT calculations, the measured electronic structure can be well understood regarding the cis-trans distortion, which transforms the two-dimensional square net-derived Dirac bands into quasi-1D trivial bands. Intriguingly, multiple topological states can be identified around the zone center, including a nontrivial Z2 topological surface state and a bulk Dirac state. Our study clarifies the impact of cis-trans distortion and identifies LaAgAs2 as a topological material with multiple topological states near the Fermi level, providing a guideline for intentionally designing new topological materials.

cond-mat.mtrl-sci↗

arXiv2026

Programmable, Spontaneous Superlattice Memory in a Monolayer Topological Insulator

Memory is a foundational concept across disciplines, from neurobiology and electronics to artificial intelligence and quantum gravity. In materials, memory effects typically arise from ferroic orders, such as ferroelectricity and ferromagnetism, where information is stored in charge or spin degrees of freedom. Here, we report a surprising discovery of a nonvolatile superlattice memory effect in monolayer TaIrTe4, a dual quantum spin Hall insulator, where information is encoded through sharply contrasting lattice periodicities. In particular, in a pristine monolayer, we observe the spontaneous emergence of a long-period superlattice that can be programmed ON and OFF in a nonvolatile manner by electrostatic tuning of low-energy electronic states. This switching toggles the system between two structural configurations with unit cell areas differing by nearly two orders of magnitude. Mechanistically, our results reveal two independent and distinct instabilities, one in the lattice and the other in the QSH electrons, which are coupled, leading to electrostatic control of lattice configurations with nonvolatile memory. This finding is enabled by combining linear and nonlinear transport measurements, Raman spectroscopy, and scanning tunneling microscopy, which probe complementary aspects of the underlying orders. Remarkably, this nonvolatile memory effect stabilizes a spontaneous superlattice with a periodicity on the few-nanometer scale that remains robust across a wide doping range, persists over days, and survives above 70 K. Combined with the QSH topology, this stability offers a promising route to nonvolatile memory control of topological flat bands and their filling enabled quantum states. Our preliminary data indeed show the emergence of new insulating states at fractional superlattice fillings, which can be clearly switched ON and OFF together with the superlattice.

cond-mat.mes-hall↗

arXiv2026

Advancing Visual Reliability: Color-Accurate Underwater Image Enhancement for Real-Time Underwater Missions

Underwater image enhancement plays a crucial role in providing reliable visual information for underwater platforms, since strong absorption and scattering in water-related environments generally lead to image quality degradation. Existing high-performance methods often rely on complex architectures, which hinder deployment on underwater devices. Lightweight methods often sacrifice quality for speed and struggle to handle severely degraded underwater images. To address this limitation, we present a real-time underwater image enhancement framework with accurate color restoration. First, an Adaptive Weighted Channel Compensation module is introduced to achieve dynamic color recovery of the red and blue channels using the green channel as a reference anchor. Second, we design a Multi-branch Re-parameterized Dilated Convolution that employs multi-branch fusion during training and structural re-parameterization during inference, enabling large receptive field representation with low computational overhead. Finally, a Statistical Global Color Adjustment module is employed to optimize overall color performance based on statistical priors. Extensive experiments on eight datasets demonstrate that the proposed method achieves state-of-the-art performance across seven evaluation metrics. The model contains only 3,880 inference parameters and achieves an inference speed of 409 FPS. Our method improves the UCIQE score by 29.7% under diverse environmental conditions, and the deployment on ROV platforms and performance gains in downstream tasks further validate its superiority for real-time underwater missions.

cs.CV↗

arXiv2026

Quantum criticality in sub-Ohmic systems with three competing terms: beyond conventional spin-boson physics

Quantum phase transitions (QPTs) in the spin-boson model with/without the rotating-wave approximation (RWA) are systematically investigated through variational calculations using a sub-Ohmic bath with high spectral density. Four cases involving different system-environment interactions are examined, where transition points and critical exponents are accurately determined across varying tunneling strengths. Contrary to prior work, a rich phase diagram is revealed in the tunneling-coupling plane even at the low spectral exponent $s<1/2$, with a novel U(1)-symmetric phase being identified. As coupling increases, a multi-stage QPT sequence arises for the tunneling $0<\Delta < \Delta^*=0.074(1)$, whereas a single transition occurs beyond this range. Furthermore, an odd-parity phase is found to emerge even under the positive tunneling, exhibiting distinct characteristics relative to the prototype model.

quant-ph↗

arXiv2026

Robust Single-message Shuffle Differential Privacy Protocol for Accurate Distribution Estimation

Shuffler-based differential privacy (shuffle-DP) is a privacy paradigm providing high utility by involving a shuffler to permute noisy report from users. Existing shuffle-DP protocols mainly focus on the design of shuffler-based categorical frequency oracle (SCFO) for frequency estimation on categorical data. However, numerical data is a more prevalent type and many real-world applications depend on the estimation of data distribution with ordinal nature. In this paper, we study the distribution estimation under pure shuffle model, which is a prevalent shuffle-DP framework without strong security assumptions. We initially attempt to transplant existing SCFOs and the na\"ive distribution recovery technique to this task, and demonstrate that these baseline protocols cannot simultaneously achieve outstanding performance in three metrics: 1) utility, 2) message complexity; and 3) robustness to data poisoning attacks. Therefore, we further propose a novel single-message \textit{adaptive shuffler-based piecewise} (ASP) protocol with high utility and robustness. In ASP, we first develop a randomizer by parameter optimization using our proposed tighter bound of mutual information. We also design an \textit{Expectation Maximization with Adaptive Smoothing} (EMAS) algorithm to accurately recover distribution with enhanced robustness. To quantify robustness, we propose a new evaluation framework to examine robustness under different attack targets, enabling us to comprehensively understand the protocol resilience under various adversarial scenarios. Extensive experiments demonstrate that ASP outperforms baseline protocols in all three metrics. Especially under small $\epsilon$ values, ASP achieves an order of magnitude improvement in utility with minimal message complexity, and exhibits over threefold robustness compared to baseline methods.

cs.CR↗

arXiv2026

Experimental realization and self-testing of semisymmetric informationally complete measurements via a one-dimensional photonic quantum walk

Generalized quantum measurements play a crucial role in quantum mechanics, and symmetric informationally complete positive operator-valued measurements (SIC POVMs) provide a powerful and flexible framework for extracting information from quantum systems. However, the existence of SIC-POVMs in every finite dimension remains an open question, which has stimulated extensive research into alternative classes of POVMs. Recently, Geng $et$ $al$. [Phys. Rev. Lett. 126, 100401 (2021)] proposed a broader class of SIC POVM, called semisymmetric informationally complete POVM (semi-SIC POVM), which extends beyond SIC POVM. In this work, we focus on the four-outcome POVMs and experimentally realize the semi-SIC POVMs using a one-dimensional discrete-time quantum walk. Additionally, employing single photons and linear optics, we perform an experimental self-testing of semi-SIC POVMs in the semi-device-independent manner. Our results pave the way for exploring quantum certification with generalized quantum measurements.

quant-ph↗

arXiv2026

Coordinates for ${\rm SL}_3$-web basis elements in closed surfaces

The ${\rm SL}_3$-skein algebra of a closed surface $\Sigma_g$ is a quantization of the ${\rm SL}_3$ character variety of $\Sigma_g$, where $g$ denotes the genus of the surface. This algebra admits a basis consisting of non-elliptic web diagrams in $\Sigma_g$. In this paper, we introduce explicit coordinates for non-elliptic web diagrams on $\Sigma_g$, yielding a parametrization by a submonoid of $\mathbb Z^{d}$. Here $d = 16g - 16$ for $g \ge 2$ and $d = 4$ in the torus case $g = 1$, coinciding with the dimension of the corresponding character variety.

math.GT↗