arXiv subjects

Tianyu Yang

Publications and source records attributed to Tianyu Yang.

At least 19 recordsLinked to original sources

arXiv2026

Real-time volumetric MRI-guided MLC tracking: A feasibility study

Background and purpose: The MR SIGnature MAtching (MRSIGMA) was recently proposed as a real-time free-breathing 3D MR imaging technique with an imaging latency of 148 ms. It is a powerful tool for intrafractional motion monitoring for disease sites such as those in the abdominal region. This study presents a proof-of-concept MLC tracking system for dynamic adaptation of the radiation treatment based on the real-time target localization facilitated by MRSIGMA on a 1.5 T MR-Linac. Material and methods: An MLC tracking interface was developed for integration with MRSIGMA. The centroid position of a moving target was computed in real time by MRSIGMA, which was then processed by a predictive filter to account for imaging and MLC adjustment latencies. The MLC leaves were repositioned by rigid translation. To validate the MLC tracking process, a 10 cm circular field was delivered to a moving phantom for a series of sinusoidal motions and patient motion. A clinical pancreas SBRT plan was also delivered with the patient motion trace. Results: For the circular field, geometric errors of <1.8 mm and overall system latency of < 125 ms were achieved for all motion traces. Compared to no tracking, MLC tracking remarkably improved the dosimetric accuracy for the circular field and the clinical pancreas plan, with an average Gamma passing rate increased from 87.1% to 99.9% based on a 3%/2mm criteria. Interpretation: A 3D MRI-guided MLC tracking technique was developed and validated on a clinical MR-Linac. Preliminary results demonstrated great dosimetric improvements with MLC tracking for treatments with moving targets.

physics.med-ph↗

arXiv2026

Beyond Post-Hoc Rewards: Privileged Reasoning Exploration for Multimodal Embeddings

Advances in Multimodal Large Language Models (MLLMs) have driven substantial progress in multimodal embeddings. Recent methods exploit generative capabilities by augmenting embedded inputs with self-generated reasoning to produce more informative embeddings. However, conventional retrieval supervision for training embedding models specifies pairwise relevance relations without explicitly guiding the reasoning used for input augmentation. Such reasoning is neither directly observed nor uniquely determined by the input, allowing multiple plausible traces to introduce semantic cues of varying retrieval utility. Existing approaches largely optimize reasoning conditioned on the input alone, limiting the exploration of retrieval-relevant traces. We introduce PILOT, a framework that models reasoning as a latent variable and instantiates its prior and posterior distributions as distinct reasoning modes of a shared MLLM. The prior conditions only on the input, whereas the posterior uses additional training-time relational evidence to guide reasoning exploration. To translate the posterior's broader exploration into inference-time improvements, we derive an importance-weighted variational objective and instantiate it as a prior-posterior co-evolution framework that improves both modes under a shared inference-compatible reward while selectively transferring posterior trajectories to the prior according to their utility and prior reachability. On the 78-task MMEB-V2 benchmark, PILOT outperforms the primary baseline across all 12 task groups at both 2B and 7B parameters, improving overall scores by 3.2 and 3.3 points, respectively, with the largest gains in visual-document retrieval (+6.1 and +6.4 points). Our results show that relational supervision is most valuable not as a post-hoc reward but as a signal for shaping how reasoning is explored in training reasoning-enhanced embedding models.

cs.AI↗

arXiv2026

Seeing, Maintaining, and Learning: The Evolution of Multimodal Agent Memory

Multimodal models and agents have emerged to enable AI systems to learn from and interact with the real world more like humans by integrating information across multiple modalities. However, enabling multimodal agents to continually acquire useful information from complex, heterogeneous multimodal data remains a significant challenge as they move toward long-horizon tasks and lifelong settings. Addressing this challenge requires effective memory mechanisms that can maintain continuity across sessions, adapt to changing users and environments, and refine memory-management policies through accumulated experience. We introduce a two-axis analytical framework for multimodal agent memory structured around two key questions: how multimodal information is formed and represented (Axis I, Formation and Representation), and how its management adapts with accumulated experience (Axis II, Memory Management). Axis I, Formation and Representation, characterizes what information is retained and how it is encoded (e.g., modality-specific embeddings, textual and structured records, latent and parametric representations) to support subsequent reasoning and action. Axis II, Memory Management, distinguishes three tiers: T1 Static Memory reuses fixed memory; T2 Dynamic Memory updates memory through a stable policy; and T3 Self-Evolving Memory updates memory through an evolving policy. Based on this framework, we synthesize recurring design patterns and derive principles for balancing source fidelity with abstraction, maintaining temporal validity and provenance, grounding memory use in recoverable evidence, and evolving memory policies over time. Furthermore, we discuss the challenges and potential future directions for the memory mechanism of multimodal agents.

cs.AI↗

arXiv2026

Physics: Benchmarking Foundation Models on University-Level Physics Problem Solving

We introduce Physics, a benchmark for evaluating foundation models on advanced physics problem solving. It contains 1034 annotated, open-ended problems from PhD-qualifying exams across six core physics domains. Solving these problems requires applying domain-specific knowledge to physical setups and carrying out multi-step mathematical reasoning. Each problem includes a worked solution and independently checked reference answers. An LLM-based grading protocol evaluates individual answers and complete multi-part problems, supporting numeric, symbolic, and text responses. We evaluate 53 proprietary and open-weight models. On test problems without figures, high accuracy is concentrated in a small group of thinking models, with the strongest open-weight models approaching proprietary performance. Most open-weight models still solve fewer than half of these problems. Our analysis of Qwen3.5-9B reveals failures in applying physics knowledge and interpreting figures, with misread figures affecting several subquestions.

cs.AI↗

arXiv2026

Of Errors and Echoes: Responsibility-Aware Credit Assignment for Tool-Use Agents

Reinforcement learning improves tool-use agents, yet fine-grained credit assignment remains challenging even with step-level feedback. An observed error may be inherited from an earlier decision, while multiple responsible decisions may share the same downstream consequence or participate in later recovery. We introduce Error-Chain-Aware Optimization (ECHO), a responsibility-aware credit assignment framework that separates responsibility identification from credit allocation. A frozen judge combines retrospective trajectory analysis with decision-time assessment from the policy-visible prefix to distinguish newly introduced errors, inherited consequences, missed corrections, and successful corrections. ECHO then counts shared downstream consequences once across responsible decisions and constrains rewards for repairing self-induced errors through source-funded recovery, producing bounded action-level supervision alongside terminal supervision. Experiments on six diverse tool-use benchmarks at both 4B and 8B scales show strong and consistent improvements over the compared baselines, with ECHO achieving the highest score on five of six benchmarks at each scale. These results support decision responsibility as a useful basis for fine-grained credit assignment in tool-use agents. Our code is available at https://github.com/xid32/ECHO.

cs.CL↗

arXiv2026

ADAPT: Audio-Depth Action Planning with Transition-Aware Rewards for Lights-Out Industrial Navigation

Vision-language-action models for embodied navigation typically rely on RGB observations, limiting their applicability in lights-out industrial environments where visible illumination may be unavailable. We study illumination-independent navigation using spatial audio for target localization and depth for obstacle-aware geometry. To support this setting, we introduce DarkNav-Bench, a benchmark with 24 industrial scenes and four progressively challenging acoustic conditions. We further propose ADAPT, an audio-depth action-planning framework equipped with spatial-audio perception and optimized for closed-loop navigation through reinforcement learning (RL). Our key RL contribution, transition-aware reward shaping (TARS), defines a pose-aware shortest-action potential over a position-heading graph. Unlike conventional geodesic-progress rewards, TARS accounts for both feasible movements and in-place rotations, providing action-consistent feedback for policy optimization. Experiments show that ADAPT consistently outperforms training-based navigation baselines and zero-shot multimodal models on both seen and unseen scenes. On unseen environments, it achieves 90.6% success under Clean conditions and retains 60.0% success under the most challenging Localized Noise setting. These results demonstrate the potential of spatial audio and depth as complementary modalities for reliable RGB-free navigation in lights-out industrial environments.

cs.AI↗

arXiv2026

Data-Free On-Policy Distillation: How Far Can We Go Without External Data?

On-policy distillation (OPD) is increasingly applied to frontier foundation model post-training. Prior work in this area has largely focused on algorithmic advances, yet it remains unclear how much OPD depends on its training questions and, in particular, how far this dependence can be reduced. Across two representative single-teacher OPD settings, we find that training on 8 real prompts yields performance comparable to training on 17k problems, while datasets differing substantially in measured difficulty and initial distillation gap yield similar outcomes. Our analyses suggest two complementary explanations: repeated sampling could allow even a few prompts to expose substantial teacher supervision, while OPD transfers generalizable reasoning capabilities beyond dataset-specific knowledge. Building on these observations, we next propose a data-free on-policy distillation (DF-OPD) setting to investigate whether the system can supply the training questions itself, eliminating the need for external data. With 64 self-generated questions obtained without seed examples, DF-OPD yields performance comparable to full-data OPD in both single-teacher settings. This finding also holds in multi-teacher OPD: across mathematics, code, and instruction following, 1k generated questions achieve performance comparable to training on approximately 7k real post-training examples. We further explore whether OPD can operate even without explicit training questions. The experiments show that this is effective only in limited cases, where the student unexpectedly generates and answers its own questions, thereby reducing the process to an implicit form of DF-OPD. Together, these findings invite a reassessment of the role of training data in on-policy distillation. Code is available at https://github.com/Ryuki661/DF-OPD

cs.LG↗

arXiv2026

On-Policy Distillation with Curriculum Turn-level Guidance for Multi-turn Agents

Multi-turn agents that plan, invoke tools, and interact with environments offer a promising paradigm for solving complex tasks, yet their capabilities typically rely on very large models whose inference cost is prohibitive in practice. On-Policy Distillation (OPD) is a natural recipe for transferring such capabilities to smaller students, but we find that it suffers a characteristic failure mode in this setting: small student errors compound across turns and push the trajectory out of the teacher's familiar state distribution, so the teacher's supervision becomes least reliable precisely where the student needs it most. We propose Guided On-Policy Distillation (Guided-OPD), a simple yet effective algorithm that mixes teacher- and student-generated turns within each rollout and schedules the teacher's intervention probability along a curriculum that decays to zero. Strong guidance keeps early trajectories close to the teacher distribution and is then gradually withdrawn to recover the purely on-policy regime used at inference. On ALFWorld, ScienceWorld, and WebShop, distilling Qwen3 students from a Qwen3-30B-A3B teacher, Guided-OPD yields average relative gains of 21.1\% in Score and 25.5\% in Success Rate over vanilla OPD, with larger gains on smaller students.

cs.LG↗

arXiv2026

Ada3Drift: Adaptive Training-Time Drifting for One-Step 3D Visuomotor Robotic Manipulation

Diffusion-based visuomotor policies model complex action distributions through iterative denoising, but repeated inference adds latency to robotic control. One-step generators reduce this cost, motivating training objectives that retain useful action structure with few demonstrations. We present Ada3Drift, a point-cloud-conditioned policy that builds on Drifting Models to perform distribution refinement during training and generate action chunks in one forward pass. Our central design is a regression-to-drifting curriculum: paired action regression first emphasizes observation--action correspondence, while a sigmoid schedule progressively increases a batch-level action-distribution regularizer. The drifting term combines attraction to demonstrated actions and repulsion among generated samples using inherited multi-temperature aggregation. A timestep-free generator preserves single-step inference throughout. On Adroit, Meta-World, RoboTwin, and five real-world tasks, Ada3Drift achieves the highest reported average success rates among the evaluated baselines with 1 NFE, compared with 10 NFE for the diffusion baselines. Controlled ablations favor sigmoid scheduling at matched demonstration budgets. We will release our code and pretrained model weights.

cs.CV↗

arXiv2026

Prior-Aided Masked Vector Quantization CSI Feedback for FDD Massive MIMO Systems

Downlink channel state information (CSI) feedback is a key bottleneck in frequency-division duplex (FDD) massive MIMO systems, as the user equipment (UE) must convey its estimated channel to the base station (BS) over a limited uplink (UL) budget. To improve CSI reconstruction accuracy under tight feedback constraints, we propose prior-aided masked vector quantization (PM-VQ), a learning-based separate source--channel coding (SSCC) feedback scheme conditioned on the average angle--delay power map---a compact representation of the channel second-order statistics available at both the UE and the BS. In PM-VQ, a prior-aided encoder maps the CSI to latent tokens, and a spatially-adaptive masking module (SAMM) scores and selects the most informative tokens within the feedback budget. The selected tokens are vector-quantized and fed back together with their positions, while an adaptive de-masking module (ADM) completes the latent representation at the BS before prior-conditioned decoding. To support variable-rate compression, a single model is trained over a range of selected-token counts, enabling operation across multiple feedback dimensions without retraining. We evaluate PM-VQ against three representative baselines on a Sionna-generated 3GPP TR~38.901 UMa dataset, focusing on the most challenging diffuse regime where channel energy is spread across many angle--delay coefficients. Simulation results show that PM-VQ achieves the lowest NMSE across all tested SNR levels and feedback dimensions in this regime. Moreover, the angle--delay power-map prior remains beneficial even when estimated from only a few channel realizations.

cs.IT↗

arXiv2026

Low-Complexity Near-Field Channel Estimation and Subcarrier-Cooperative Hybrid Precoding for Wideband XL-MIMO OFDM Systems

This paper addresses both accurate near-field channel acquisition and scalable precoding for wideband extremely large aperture MIMO (XL-MIMO) OFDM systems, with particular focus on reducing complexity. In the considered system, each MIMO multipath component is characterized by five continuous angle-distance-delay parameters, making conventional sparse recovery methods computationally infeasible while suffering from grid mismatch. We propose a decoupled off-grid channel estimation algorithm based on sequential sparse Bayesian learning (SBL). By exploiting the separable structure of the OFDM near-field atom, the five-dimensional joint search is replaced by a sequence of low-dimensional operations and active-set inference, avoiding multiplicative scaling with the per-dimension grid sizes. A continuous-domain Newton refinement is incorporated based on the exact marginal likelihood to mitigate the grid mismatch. Based on the estimated multipath parameters, we then develop a subcarrier-cooperative hybrid precoding framework for multiuser wideband transmission. A large-scale-matrix-inversion-free alternating optimization algorithm is established to maximize the sum user rate, together with a non-iterative low-complexity design that exploits the parameterized channel structure and provides an effective initialization. Simulation results demonstrate the accuracy of the estimation methods compared to several benchmarks and the effectiveness of the precoding algorithm based on estimated channel parameters.

eess.SP↗

arXiv2026

Tool-Making and Self-Evolving LLM Agents in Low-Latency Systems

Production LLM agents often waste latency and reliability by regenerating code for the same procedural steps on every request. We replace this inference-time coding loop with an agentic tool-making pipeline that compiles repeated SOP steps into validated, versioned tools before deployment. The tool-maker grounds synthesis in the live environment as it collects execution traces, observes backend schemas and values, generates candidate tools, and repairs them against labeled cases. At runtime, the production agent calls these tools directly and falls back to code generation only when needed. We deploy the approach in a Fulfillment Center alarm-triage system, where an agent diagnoses alarms against a 44-node SOP over heterogeneous metric backends. In production, tool calls reduce p50 latency by 42%. On 1,500 historical alarms, they reduce end-to-end error rate by up to 53% by suppressing run-to-run variance in repeated steps. Because tools return compact structured verdicts, they also enable a simpler direct-call architecture, reducing p50 latency by a further 62% in a controlled ablation. Versioned tools also improve auditability and expose specification gaps and upstream data drift. Our results show that agents that build and maintain their own tool libraries, a key element of self-evolving agents, can make industrial LLM systems faster, more reliable, and easier to operate.

cs.CL↗

arXiv2026

Eluna: An Agentic LLM System for Automating Warehouse Operations with Reasoning and Task Execution

Warehouse operations are governed by Standard Operating Procedures (SOPs) that encode complex, multi-system decision logic, which must be executed reliably under strict time constraints, yet LLM agents lack mechanisms to enforce procedural compliance and degrade under the context overload full SOP specifications introduce. We present Eluna, a production-deployed agentic system for reliable SOP execution. Eluna is a graph-guided, multi-agent framework that encodes SOPs as directed acyclic graphs with progressive disclosure and delegates independent tasks to parallel sub-agents, each with persistent code execution and live data access. To meet production latency and accuracy needs, we use asymmetric episodic distillation where a strong teacher is improved through episodic error memories, then a smaller student is fine-tuned on the corrected trajectories with memory stripped, internalizing corrections without inference-time overhead. On a 13-task benchmark and two production applications, our fine-tuned models match or exceed their teacher, beat all larger off-the-shelf baselines, and reach 94% expert agreement on the ticket processing application.

cs.LG↗

arXiv2026

Is Visual Prompting All You Need? Studying VLM Spatial Reasoning under Progressive Visual Scaffolds

Vision-language models (VLMs) have advanced rapidly in multimodal reasoning, yet recent work shows that their failures often reflect an interaction between visual grounding and downstream reasoning. What remains less clear is how the visual presentation of a task shapes model performance and failure modes when the underlying reasoning problem is unchanged. We study this question in SPaRC, a benchmark for grid-based visual spatial planning, by introducing lightweight input-side scaffolds that preserve the visual modality while making spatial structure more accessible. Across multiple VLMs, these scaffolds improve task accuracy over the original visual setting by up to 34.0 percentage points and further complement GRPO-based training, yielding up to 4.6 additional accuracy points compared with near-zero gains on the original visual input. Analyses on both end-to-end task solving and object detection show that these gains are closely tied to reductions in grounding-related errors, while rule reasoning remains comparatively challenging. We find that visual presentation is a central factor that determines whether VLM benchmarks measure grounded perception, downstream reasoning, or a mixture of both.

cs.CV↗

arXiv2026

DesignAgent3D: Interactive 3D Scene Editing via Designer-like Multimodal Reasoning

Text guided 3D scene editing provides an intuitive interface for modifying reconstructed environments, but remains difficult because natural language design requests are often semantically underspecified and must be grounded in cluttered 3D scenes. Existing methods typically formulate the task as one-shot conditional generation from a single prompt, failing to resolve ambiguous user intents or achieve precise spatial grounding. Consequently, they suffer from severe object localization drift, tracking failure under occlusions, and the notorious multi-view "sticker effect." To overcome these limitations, we present DesignAgent3D, an interactive multimodal agentic framework that reformulates 3D scene editing as a designer-like Plan-Perceive-Act paradigm. The agent first plans by interacting with the user to clarify underspecified design goals, then perceives by grounding the intended edit to specific objects or regions in the 3D scene, and finally acts by applying controlled visual modifications while preserving scene consistency. The edits are further integrated into the underlying 3D representation, supporting persistent and multi-view consistent novel-view rendering. Extensive experiments across both NeRF and 3D Gaussian Splatting backbones demonstrate that DesignAgent3D significantly outperforms state-of-the-art baselines, delivering superior semantic intent alignment, impeccable spatial localization accuracy, and high-fidelity multi-view consistency.

cs.CV↗

arXiv2026

X-NavDP: Generalizing Navigation Diffusion Policy to Novel Behavior and Embodiments with Group Q-score Reweighted Matching

Pretraining navigation diffusion policies rely on large-scale expert demonstrations. These data are typically generated by a fully-informed oracle planner suited to a single nominal robot. This limits the policy's generalization to diverse embodiments and challenging scenarios (e.g., escaping dead ends or detouring long obstacles) that demand diverse local reactive behaviors with only onboard local observations. Post-training the policy with reinforcement learning (RL) offers a principled remedy. However, previous RL for diffusion approaches lead to only marginal improvements. This is because the intractable likelihood of diffusion policies renders policy gradients unstable in addition to inefficient policy exploration. To address these challenges, we propose a data-efficient diffusion RL post-training framework - GQRM (Group Q-score Reweighted Matching). Our framework introduces two complementary designs: (i) a self-bootstrapped exploration strategy with behavior perturbation that preserves the pretrained policy prior, and (ii) a group Q-score normalization mechanism that computes per-trajectory values on each state for efficient reweighted score matching. By conducting distributed online RL training across heterogeneous embodiments, the resulting fine-tuned policy, X-NavDP, achieves state-of-the-art cross-embodiment visual navigation performance, improving the overall success rate from 61.20% to 84.28% in simulation and 10% to 65% in real-world hard cases. The code and model are publicly available at https://yty-sky.github.io/x-navdp-project-page.

cs.RO↗

arXiv2026

Near-Field Communications with Different Array Geometries: Rayleigh Distance, Channel Estimation, and Transmission Design

This work establishes a framework of near-field communication under different array geometries of extremely large-scale multiple-input multiple-output (XL-MIMO). We first formulate the near-field spatial non-stationary channel model which is characterized by the distance between the user and each antenna on uniform and modular curved arrays. By fixing the total number of antennas while varying the degree of curvature, we investigate a fair case where the horizontal arc length of the curved array is the same as the planar array. We explicitly unveil the non-trivial impact of array curvature on extending the near-field region for cell edges. Then, for arbitrary array geometries and arbitrary-field channels, we estimate the spatial-domain channel by tackling a compressed sensing problem with a learned regularizer. Without relying on specific codebooks, we propose a denoising autoencoder (AE)-aided approximated message passing (AMP) algorithm and provide the corresponding theoretical replica bound. Finally, based on the estimated channel, we propose an optimization algorithm to maximize the sum user rate for sub-connected XL-MIMO systems by jointly designing the array geometry and hybrid precoding in the downlink. Numerical results demonstrate that the proposed AE-AMP algorithm can effectively estimate the spatial non-stationary near-field channels with robustness and generalities compared to several conventional and deep-learning-based benchmarks. The improvement of data rate by using modular curved arrays with the estimated channel is also validated.

eess.SP↗

arXiv2026

One-Step Epitaxial Access to Rhombohedral Graphene Flat-Band States on Step-Bunched SiC

Rhombohedral graphene multilayers provide a moiré-free platform for correlated and topological flat-band physics, but direct, transfer-free epitaxial access to thickness-tunable multilayers remains limited. Here we report a one-step graphitization route on 4$^\circ$ off-axis 4H-SiC, in which high-temperature flash annealing simultaneously drives self-organized step bunching and multilayer graphene formation. Atomic-resolution cross-sectional scanning transmission electron microscopy identify local ABC registry and distinguish rhombohedral from Bernal stacking. The thickness is tuned from bilayer to more than twenty layers by varying single parameter, the annealing temperature. Angle-resolved photoemission spectroscopy directly tracks the thickness-dependent evolution from interface-dominated low-energy states toward pronounced near-Fermi-level flat-band spectral weight in thick multilayers. Low-temperature scanning tunneling microscopy and spectroscopy on a 17-layer film further reveal a 13.4 meV low-energy spectral reconstruction and a $\sqrt{3} \times \sqrt{3}$ Kekulé-like modulation, providing microscopic signatures consistent with an intervalley-mixed electronic texture. This one-step, transfer-free approach establishes step-bunched SiC as an epitaxial platform that links stacking engineering with moiré-free correlated flat-band electronic states.

cond-mat.mtrl-sci↗