arXiv subjects

Chenyu Zhang

Publications and source records attributed to Chenyu Zhang.

At least 19 recordsLinked to original sources

arXiv2026

One Sentence, One Drama: Personalized Short-Form Drama Generation via Multi-Agent Systems

Existing approaches for digital short-drama production typically rely on one-shot LLM generated scripts and loosely coupled pipelines, which fail to satisfy three key requirements of short-drama generation: (1) narrative pacing, resulting in weak hooks, insufficient escalation, and unattractive endings; (2) spatial consistency, leading to drifting scene layouts and inconsistent character positions across clips; and (3) production-level quality control, requiring extensive manual review and correction across script and visual stages. We present One Sentence, One Drama, a hierarchical multi-agent framework that transforms a user's single-sentence idea into a fully produced short drama through structured intermediate modules and iterative refinement. Our approach is built upon three key components: (1) a multi-agent debate-based story generation module that enforces short-drama pacing and narrative coherence; (2) a 3D-grounded first-frame generation mechanism that establishes a shared spatial reference for consistent character positioning and scene layout across clips; and (3) multi-stage reviewer loops that perform comprehensive error detection and targeted revision across script, visual, and video generation stages. We also introduce scene-level BGM matching and scene transition planning to improve the audience's immersive experience. To systematically evaluate this task, we introduce Short-Drama-Bench, a benchmark that extends standard video quality metrics with short-drama-specific criteria. Experimental results demonstrate that our method significantly outperforms existing pipelines in narrative quality, cross-clip consistency, and overall viewing experience.

cs.CV↗

arXiv2026

Robotic Spatiotemporal Editing: A Unified Interface for Controllable Action Generation via Masked Action Reconstruction

Generative policies can synthesize robot behaviors, yet controlling and editing them through diverse spatiotemporal references remains challenging. In the real world, human intent requires robot behaviors that satisfy constraints on where, when, and how actions are executed. Our key insight is that spatiotemporal control references, such as keypoints, flows, orientations, and keyposes, can be viewed as partial specifications of the same temporal Cartesian action space. We introduce the MAP (Masked Action Space) interface to unify these action references and formulate controllable manipulation as an action inpainting problem. Our Inpainting Policy (IP) reconstructs missing action dimensions of the end-effector, combining progress-guided alignment with multi-mask training supporting unseen action-constraint combinations. We further adopt a training-free RePaint variant that iteratively inpaints the action during denoising to reduce control error. On ManiSkill and LIBERO-10, our policy achieves average task success rates of 92.67% and 64.67%, respectively.Real-world experiments demonstrate spatiotemporal control with average positional and rotational errors of 44.3 mm and 7°, respectively, showing that our policy can use diverse references to guide executable robot behaviors.

cs.RO↗

arXiv2026

Network Intervention by Polling Strategic Agents

A planner in a network of strategic agents faces three entangled challenges: the optimum depends on agents' private information, queried agents may misreport to steer the outcome, and exact computation does not scale. We study these challenges in multi-activity network games with heterogeneous private technologies, in which the planner sets non-discriminatory prices. We show that the optimal prices admit a centrality-based decomposition of the welfare kernel: each agent's contribution scales with its squared centrality in a network reweighted by agents' preferences across activities. This decomposition motivates Poll, a polling algorithm in which the planner samples one agent per round, walks briefly through the agent's neighborhood, and updates the price from a local report. From the same decomposition flow three forms of efficiency: computationally, Poll uses significantly fewer operations than exact computation and other distributed methods, requiring up to three orders of magnitude less communication on a real-world network with over 300,000 agents; statistically, its query complexity scales with topology and preference heterogeneity rather than explicitly with population size; and economically, it converges to welfare-maximizing prices while admitting behavior-specific implementations that induce truthful reports and detect adversarial deviations.

cs.GT↗

arXiv2026

Careful Judge: Safe and Efficient Human-AI Collaborative Decision Making

In human-AI collaborative decision making, human review can prevent unsafe AI decisions, but each human judgment is costly. Treating human intervention after AI abstention as a one-off fallback misses the opportunity to improve future AI decisions for greater automation, yet AI adaptively learning from selectively queried human feedback breaks safety guardrails calibrated for old models. We approach this challenge with CARE---calibrated adaptive rectification and escalation---an end-to-end pipeline that combines AI models and human reviewers to guarantee safe, human-aligned decisions, while continuously learning from human feedback to achieve greater automation with fewer human queries. CARE is principled, general, modular, and works with any black-box AI model. Our novel adaptive calibration module guarantees risk control at every time step for any rectification module. We further show how CARE improves query efficiency when the AI model is well trained and the human-AI misalignment has a clear structure. Experiments on four safety-critical real-world datasets spanning driving, language, and robotics demonstrate that CARE achieves human-aligned decisions while reducing human queries by 25-81% relative to baselines.

stat.ML↗

arXiv2026

Dex-X: Learning Visual-Tactile Dexterous Manipulation From Human Videos with Simulated Interaction

Human videos are an abundant source of dexterous manipulation behaviors, but they lack tactile information that is crucial for contact-rich interaction. This raises a fundamental question: can robots learn deployable visual-tactile dexterous manipulation policies from human video demonstrations without robot-side data collection? We present DEX-X, a framework for learning visual-tactile dexterous manipulation from human videos through simulation. Our key insight is that simulation can serve as a tactile completion engine. Given monocular human demonstrations, DEX-X reconstructs hand-object interactions in simulation, where physically grounded contact dynamics provide tactile supervision unavailable in the original videos. Leveraging this recovered tactile information, we train visual-tactile dexterous manipulation policies and distill them into deployable policies operating on point-cloud observations and tactile sensing. We demonstrate zero-shot sim-to-real transfer on a dexterous hand-arm platform across diverse grasping and contact-rich tool-use tasks. The teacher policy achieves 65.9% average success across six task categories in simulation, while the distilled visual-tactile policy achieves 93% success on real-world cube picking and 53% on the challenging table-cleaning task. Zero-shot generalization to unseen object geometries is also observed on object-picking tasks. Our results suggest that simulated interaction is a key bridge between human videos and deployable dexterous manipulation policies, providing the missing physical supervision needed for scalable robot skill learning from Internet-scale human video data.

cs.RO↗

arXiv2026

Rethinking Causal Action Tokenization with Conditional Annealing in Flow Matching

Autoregressive Vision-Language-Action (VLA) models offer a scalable path to robot learning, yet existing action tokenizers treat tokenization as a compression problem, producing representations that are semantically misaligned with the autoregressive backbone. We propose CATok, a causal action tokenizer that reframes tokenization as a causally structured generative process. CATok introduces a conditional annealing mechanism that extracts action tokens by progressively annealing a flow-matching process: each token is conditioned on all preceding tokens and encodes the residual reconstruction signal at a specific noise level, establishing a coarse-to-fine causal token space whose generative semantics are structurally aligned with autoregressive modeling. A token-conditioned flow-matching decoder built on Multimodal Diffusion Transformer (MMDiT) reconstructs continuous action chunks from these discrete tokens with the precision of hybrid diffusion-head architectures. This discrete bottleneck enforces knowledge insulation by design, cleanly separating high-level semantic reasoning from low-level motor execution without requiring explicit attention masking. Extensive evaluations across three simulation benchmarks and real-world robotic manipulation tasks demonstrate that CATok consistently surpasses existing tokenization methods in both reconstruction fidelity-compression tradeoff and inference efficiency, while improving VLA task success rate and training efficiency, establishing a high-performance, scalable foundation for purely autoregressive VLA systems.

cs.RO↗

arXiv2026

How Strongly Should Task State Influence an LLM Agent?

Long-horizon assigned work requires an LLM agent to track the state of a task: which steps are done, blocked, cancelled, or open to repetition. Agent systems either keep this state as text in the prompt and rely on the model to read that text, or move the state into a module that enforces it, and each system is evaluated as a whole, so no one knows how much reliability comes from the state being shown, told, or enforced. We fix the task rules, the model, and paired episodes and vary how strongly task state reaches the agent: a raw transcript, an exact checklist, per-turn directives from a state machine compiled from the brief and advanced only by execution receipts, or an enforcement gate on that machine that refuses state-violating actions; every episode is scored by exact payload matching against dynamic ground truth. Across three models, two reasoning regimes, and two domains, four findings hold without per-turn reasoning: displaying accurate state is unreliable, an unverified ledger the agent writes itself beats an accurate checklist it is shown, directives help in proportion to the model's obedience, and enforcement needs no obedience but is bounded by the correctness of its state and by the matcher that maps requests to steps; per-turn reasoning at a 235B agent compresses these separations without repairing the text rungs. The same gate, compiled from $τ^2$-bench's airline policy, raises a 235B agent's pass$^1$ from 0.39 to 0.54 and changes nothing for a 35B agent that rarely violates the policy; on PM-Bench, where acting turns on recognizing a cue rather than on state, showing the record is the best rung--matching or beating both gates and reversing the ledger-over-checklist finding--and enforcing the matcher's judgement drops a 35B agent below its raw transcript. Enforcement pays when failures are state-decidable and frequent, and hurts when the gate's judgement is wrong.

cs.AI↗

arXiv2026

Are Targeted Data Poisoning Attacks as Effective as We Think?

Targeted data poisoning attacks manipulate model predictions on specific test samples by injecting malicious data into training. Yet existing evaluations report average attack success rates over randomly selected targets, obscuring true worst-case effectiveness. We argue that the right evaluation focuses on the hardest samples to poison. The same reasoning applies to defense: since targeted attacks leave no footprint at the distribution level, defenders should proactively identify the most vulnerable samples and apply targeted countermeasures. Given a test dataset, this paper identifies both the easiest and hardest to poison examples based on only clean model information. Specifically, we offer coarse evaluations using clean training dynamics, and fine-grained classification on poison class using poison distances and budgets. Our experiments show these metrics reliably stratify samples by poisoning vulnerability, enabling both rigorous worst-case evaluation and proactive vulnerability-aware defense.

cs.LG↗

arXiv2026

Adaptive 3D-RoPE: Physics-Aligned Rotary Positional Encoding for Wireless Foundation Models

Wireless foundation models (WFMs) have emerged as a promising paradigm for unified channel state information (CSI) acquisition across diverse tasks in sixth-generation (6G) networks. Although WFMs significantly outperform task-specific small models, their zero-shot cross-scenario generalization still remains limited for real-world applications. Existing positional embeddings, the sole interface through which self-attention perceives the temporal-frequency-antenna 3D physical coordinates of CSI, fail to capture the highly dynamic and axis-dependent coherence inherent in wireless channels. This paper proposes Adaptive 3D-RoPE, a channel-driven 3D rotary positional embedding framework for WFMs to dynamically align the 3D positional embeddings with the instantaneous coherence state of heterogeneous CSI. The design proceeds in three stages: first, an axis-wise learnable rotary prior independently preserves the temporal, frequency, and antenna coordinate structures; second, a feature-guided rotary modulation module maps the feature-wise standard deviation of visible CSI tokens to compact, sample-adaptive scales; third, identical coordinate offsets induce dynamically adjusted query-key interactions tailored to the instantaneous channel state. Extensive experiments on both simulated and measured datasets validate the effectiveness of Adaptive 3D-RoPE across three complementary dimensions. It reduces NMSE by 10.14, 6.25, and 4.61 dB relative to baselines under antenna, temporal, and frequency scaling, respectively. It transfers effectively to real-world measured CSI and remains robust under imperfect CSI. Finally, it transfers to the independently designed LWM backbone and beam-prediction task, improving zero-shot Top-1 accuracy by 8.03 percentage points.

eess.SP↗

arXiv2026

Chiral superconductors and competing states across a Lifshitz transition in rhombohedral pentalayer graphene

Rhombohedral multilayer graphene hosts a distinctive low-energy electronic structure in which strong Coulomb interactions and nontrivial quantum geometry intertwine to generate exotic quantum states. Recent experiments reported signatures of chiral superconductivity in electron-doped rhombohedral multilayer graphene within the spin- and valley-polarized regime. Here we map the normal-state fermiology surrounding chiral superconductivity in rhombohedral pentalayer graphene. Quantum oscillation measurements reveal an electrically controlled Lifshitz transition between a simply-connected circular quarter-metal Fermi surface and an annular quarter-metal Fermi surface. The Lifshitz boundary itself shifts with perpendicular magnetic field, consistent with the strongly momentum-dependent orbital magnetic moment of the low-energy band. Approaching the transition from either side, the electron effective mass becomes strongly enhanced, implying the formation of a nearly dispersionless band bottom and a strongly reduced kinetic-energy scale. This singular electronic structure produces a regime of exceptionally strong instability in which chiral superconductivity competes with Wigner crystalline phases and reentrant quantum Hall states. In particular, two superconducting regions with signatures of orbital time-reversal-symmetry breaking lie on opposite sides of the Lifshitz boundary and have comparable transition temperatures, yet the annular-side state is suppressed by a substantially smaller perpendicular magnetic field. Our calculation finds comparable chiral pairing tendencies on the two parent Fermi surfaces while producing a much lower orbital-Zeeman pair-breaking scale and an additional finite-momentum pairing tendency for the annular state. These results identify Fermi-surface topology as a key control parameter for chiral superconductivity in rhombohedral graphene.

cond-mat.mes-hall↗

arXiv2026

Beyond geometric symmetry: Broadband linear relations in wave scattering

The design and control of wave scattering, that is, of the reflection and transmission parameters of a device, is of ubiquitous importance. These parameters generally change with varying frequency, though certain \emph{frequency-independent} linear relations may exist between them. Reciprocity and geometric symmetry (reflections, rotations, etc.) are classic and well-known examples that are present in many devices and significantly ease their design. In this work, we go beyond these and introduce a new class of relations that cannot be induced by reciprocity or geometric symmetry. Choosing networks of waveguides as our workhorse, we discuss the conditions and consequences of such novel behaviour and showcase suitable example setups. We further experimentally test our predictions using coaxial cables and find excellent agreement in the broad frequency range between 0 and 1 GHz. Our work not only deepens the theoretical understanding of waveguide network dynamics, but also opens new avenues for applications in broadband signal processing, quantum information, and integrated photonics.

physics.optics↗

arXiv2026

UI-Mate: Advancing Open-Weight Foundation GUI Agents with In-Context Demonstrations

Foundation GUI agents can automate complex digital tasks, but deployment is hindered by scarce and biased training data, ambiguous prompts, and unreliable execution. Routine workflows rely on user-specific tools and tacit conventions, so unstated instructions can produce arbitrary variations across runs. We present UI-Mate, a foundation GUI agent that integrates an environment-grounded training stack with in-context demonstration learning. UI-Mate makes three contributions: A Scalable Environment-Grounded Training Stack: A closed-loop data engine automates task generation, environment construction, rollout, filtering, capability balancing, SFT, and online RL across massively parallel environments via unified task-verifier bundles. In-Context Demonstration Learning: A mechanism that transforms multimodal demonstrations into flexible subtask-level workflows, follows relevant demonstrated steps, and re-plans from the live interface. OSWorkerBench Benchmark and Insights: A benchmark of 100 long-horizon office tasks across 41 applications that supports instruction-only and demonstration-guided evaluation. Its demonstration resources separate a 33-task self-demo setting, built from successful strong-agent rollouts of the same targets, from a 45-task variant-demo setting, built from human recordings of related but non-identical tasks. Experiments show that UI-Mate-27B sets a new open-weight state of the art on general computer-use benchmarks, scoring 77.0% on OSWorld-Verified and 66.2% on WindowsAgentArena. On OSWorkerBench, it reaches 41.0% strict success and 76.9% progress, outperforming its Qwen3.6-27B base by 17.7 and 24.5 points. On the 33-task self-demo subset, one demonstration raises strict success from 17.2% to 35.4% and progress from 67.9% to 81.1%, substantially improving long-horizon reliability. Project page: https://ui-mate.github.io.

cs.AI↗

arXiv2026

ToW3D: Consistency-aware Interactive Point-based Mesh Editing on GANs

In this paper, we propose ToW3D that enables precise and consistent control over 3D generative adversarial networks (GANs) with the Tug-of-War competition between shape deformation and appearance consistency. Existing point-based GAN editing methods such as DragGAN and GANWarping have yielded impressive performance for 2D image manipulation. However, as 3D generators present weaker generalization ability compared with 2D due to limited training data, they would suffer from drastic changes in global appearance when editing local areas of meshes. To address this, we design a pipeline of ``drag locally, shove globally'', which iteratively performs two optimization steps: 1) pull the point towards the target, and 2) push the structure and semantics back to the source. Specifically, we design a structure adaption module based on structure which guarantees the preservation of basic geometric properties, and a semantic preservation module that maintains semantic similarity across different views. Extensive qualitative and quantitative experiments demonstrate superiority of our ToW3D approach over prior methods in terms of appearance consistency and fidelity especially under large deformations.

cs.CV↗

arXiv2026

SalesLoop: Reinforcement Learning from Performance Feedback for Sales Lead Ranking

Lead ranking in Customer Relationship Management (CRM) systems faces a persistent challenge: models achieving high offline accuracy often underperform in production. We identify three fundamental gaps responsible for this disconnect: offline-online metric mismatch, pointwise-listwise objective misalignment, and temporal distribution drift. To address these gaps, we propose SalesLoop, a reinforcement learning framework that establishes a closed feedback loop between model predictions and real-world business outcomes. Our approach introduces (1) a performance-aware reward that encodes conversion outcomes weighted by ranking position and conversion velocity, and (2) Discriminative GRPO, a listwise optimization objective that adapts Group Relative Policy Optimization to discriminative ranking models. SalesLoop improves NDCG@K by +7.9\% and P@K by +15.8\% over the strongest static baseline. A 160-day production A/B test at a New Energy Vehicle manufacturer, spanning 16.5M leads and 280 sales specialists across two provincial markets, validates statistically significant cumulative lift of +4.7\% ($p=0.047$) and +8.7\% ($p=0.002$). In production, the ranking backbone achieves Top-10\% recall of 44.1\% and surfaces high-intent leads at $2.3\times$ the conversion rate of specialist baselines.

cs.LG↗

arXiv2026

Evaluating Newtonian Mechanics in Video Generative Models with Real Physical Systems

Recent advances in image and video generation raise hopes that these models possess world modeling capabilities-the ability to generate realistic, physically plausible videos. This could revolutionize applications in robotics, autonomous driving, and scientific simulation. However, before treating these models as world models, we must ask: Do they adhere to physical laws? Current evaluation methods rely on subjective judgments or trajectory matching, limiting their usage for physical reasoning estimation, where many generations could be physically plausible. Thus, we introduce Morpheus, one of the first physics-informed evaluation frameworks for measuring the ability of video generation models to comprehend Newtonian dynamics. Morpheus features 130 real-world videos capturing physical phenomena, guided by conservation laws. Using those as conditioning for video generation, we assess physical plausibility leveraging interpretable metrics evaluated with respect to infallible conservation laws known per physical setting, leveraging advances in physics-informed neural networks and vision-language foundation models. Importantly, Morpheus targets controlled Newtonian rigid-body settings to enable quantitative checks. Our findings reveal that even with advanced prompting and video conditioning, contemporary models struggle to encode physical principles despite generating aesthetically pleasing videos.

cs.CV↗

arXiv2026

Towards CSI-Native Foundation Models: A Channel-Adaptive Roadmap for 6G

Wireless foundation models offer a path toward reusable channel state information (CSI) intelligence for sixth-generation (6G) systems. However, existing generic-backbone adaptation and CSI pretraining methods often treat CSI as task tensors rather than propagation-conditioned channel responses, thereby failing to capture the intrinsic time-frequency-spatial geometry of wireless environments. This paper presents a channel-adaptive roadmap toward CSI-native foundation models, proposing a unified framework that aligns pretraining, positional modeling, and attention control with three channel requirements: scale-aware heterogeneous exposure, physical time-frequency-antenna coordinates, and correlation-bounded token interaction. Extensive experiments demonstrate the superiority of the proposed framework across three dimensions: zero-shot generalization, reducing NMSE by more than 4 dB across spatial-temporal-frequency tasks; scale extrapolation, yielding up to a 5.4 dB gain under 8 times unseen antenna scaling; and inference efficiency, accelerating mobility-aware processing by up to 18.8%. A system-level evaluation with Sionna SYS further shows that the proposed framework uses only 7.01% of dense-pilot overhead, reaches -18.64 dB average NMSE, and improves average net spectral efficiency by 36.6% over dense LMMSE and 15.5% over WiFo, indicating that CSI-native representation learning can support pilot-efficient radio access.

cs.LG↗

arXiv2026

Rethinking Sales Lead Scoring with LLM-based Hierarchical Preference Ranking

Sales lead conversion in high-stakes domains (e.g., automotive, real estate) differs fundamentally from e-commerce recommendation due to prolonged decision cycles and multi-stage funnels. Traditional lead scoring methods rule-based scorecards, machine learning, or pointwise CTR models face severe challenges: sparse supervision, a semantic gap in unstructured CRM logs, and inability to capture relative lead priority. While Large Language Models(LLMs) offer superior semantic understanding of customer interactions, general-purpose LLMs are ill-suited for lead ranking: they generate text rather than comparable scores, and lack alignment with the hierarchical priorities of sales funnels. We introduce an LLM-based discriminative framework for sales lead scoring, which supports joint modeling of structured CRM features and unstructured customer interactions. On top of this framework, we propose HPRO (Hierarchical Preference Ranking Optimization), which augments sales lead scoring with a hierarchical preference ranking objective. HPRO employs a margin-aware Bradley-Terry formulation to transform sparse binary labels into dense, funnel-aware preference pairs, enabling lead scoring to leverage both pointwise and pairwise supervision. Experiments on large-scale data from a leading NEV brand demonstrate state-of-the-art classification (AUC 0.8161) and ranking performance (+39.7% precision among top-ranked leads). A 132-day online A/B test validates 9.5% sales volume uplift, confirming real-world commercial impact.

cs.IR↗

arXiv2026

Self-supervised Monocular Depth and Pose Estimation for Endoscopy with Latent Priors

Accurate 3D mapping in endoscopy enables quantitative, holistic lesion characterization within the gastrointestinal (GI) tract, requiring reliable depth and pose estimation. However, endoscopy systems are monocular, and existing methods relying on synthetic datasets or complex models often lack generalizability in challenging endoscopic conditions. We propose a robust self-supervised monocular depth and pose estimation framework that incorporates a Generative Latent Bank and a Variational Autoencoder (VAE). The Generative Latent Bank leverages extensive depth scenes from natural images to condition the depth network, enhancing realism and robustness of depth predictions through latent feature priors. For pose estimation, we reformulate it within a VAE framework, treating pose transitions as latent variables to regularize scale, stabilize z-axis prominence, and improve x-y sensitivity. This dual refinement pipeline enables accurate depth and pose predictions, effectively addressing the GI tract's complex textures and lighting. Extensive evaluations on SimCol and EndoSLAM datasets confirm our framework's superior performance over published self-supervised methods in endoscopic depth and pose estimation.

cs.CV↗