Searcharxiv⌕ Search

arXiv subjects

Di Zhang

Publications and source records attributed to Di Zhang.

At least 37 records · Page 2Linked to original sources
arXiv2026

LLMODE: Aligning ODEs with LLMs via Gated Token Injection for Irregular Spatio-Temporal Forecasting

Large language models (LLMs) have shown promise for spatio-temporal forecasting, but existing approaches often rely on regularly sampled token sequences and struggle with irregular observations because of temporal asynchrony, representation-space misalignment, and limited context windows. We propose LLMODE, a token-efficient framework for irregular spatio-temporal forecasting with a frozen LLM backbone. LLMODE first uses a graph-aware ODE encoder to reconstruct irregular graph observations as a continuous-time latent trajectory. A Fixed-Budget Perceiver Resampler then compresses this variable-length trajectory into a fixed number of dynamic memory tokens. In parallel, compact statistical descriptors are encoded and resampled into context memory tokens. A dual-source gated cross-attention module injects both memories into the frozen LLM, enabling controlled utilization of external spatio-temporal evidence. Experiments on three real-world urban datasets and two physical-dynamics benchmarks show competitive overall performance, with clearer advantages under sparse or dynamically complex irregular sampling. Additional evaluations on unseen urban regions further demonstrate strong zero-shot generalization without adaptation.

cs.LG↗
arXiv2026

AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning

Test-time scaling strategies for Large Language Models predominantly rely on either reinforcement learning with sparse outcome rewards or search-based methods guided by static Process Reward Models. However, outcome-based RL often suffers from training instability and sample inefficiency, while static PRMs require expensive step-wise supervision and are susceptible to reward hacking due to distributional shifts. In this paper, we introduce AIRL-S, a unified framework that integrates Adversarial Inverse Reinforcement Learning with Group Relative Policy Optimization. By inferring a dense, step-wise reward model directly from reference trajectories, AIRL-S eliminates the dependency on labeled process data and uses the same learned PRM as both a training signal and a verifier for search-based TTS. Extensive evaluations across eight benchmarks in mathematics, science, and code generation demonstrate that our policy model improves average performance by 9\% over the base model, matching GPT-4o. We further analyze how the AIRL and GRPO objectives complement each other and how the learned PRM transfers across generators and search algorithms, establishing a robust and cost-effective methodology for scaling test-time computation in complex reasoning tasks.

cs.LG↗
arXiv2026

Macaron-V1: Towards Open Continual Learning with Self-Improvement and Mixture-of-LoRA

Macaron-V1 is an open agent-model family for experiential intelligence: learning from experience in real environments and continuing to learn after deployment. It is organized around two system goals. Adaptation is pursued through recursive improvement of versioned model-harness pairs, where experience from one configuration is evaluated under an external contract and used to construct its successor. Collaboration is pursued via the Mixture-of-LoRA (MoL) architecture that freezes a base model, composes specialist LoRA adapters, and selects one LoRA per user turn. The flagship Macaron-V1-Venti (748B) combines a 744B GLM-5.2 base with four LoRAs for chat, agent, coding, and GenUI; the Qwen3.6-35B-based Macaron-V1-Tall (50B) uses the same design for local deployment. This report presents Macaron-V1 as a co-designed system spanning architecture, algorithms, and infrastructure. The MoL architecture supports continual learning through extensible LoRA specialists. The algorithm combines Model-Harness Co-design and recursive self-improvement loop, including the UI4A component-native GenUI harness, a stateful action substrate, versioned Harness Context Protocol contract, and the agentic RL framework MindForge. The supporting infrastructure includes the post-training platform MinT, the long-context RL method LongStraw, and stability techniques for sparse MoE and DSA base models. We evaluate Macaron-V1 on Personal Intelligence, GenUI, and general capability benchmarks against frontier baselines. Our results validate the current system, while compounding gains from continual learning and collective intelligence remain open questions.

cs.LG↗
arXiv2026

Empowering Polymeric Materials Discovery by Artificial Intelligence

Polymeric materials underpin modern technologies spanning energy storage, microelectronics, healthcare and sustainable manufacturing. Yet their rational design remains exceptionally challenging because material performance emerges from complex interactions among molecular composition, chain architecture, processing history and hierarchical structural evolution across multiple length and time scales. Consequently, polymer research has long relied on labor-intensive experimentation and fragmented modeling approaches, limiting both mechanistic understanding and innovation efficiency. Recent advances in data infrastructure, machine learning, large artificial intelligence (AI) models and laboratory automation are beginning to reshape this landscape. Rather than functioning as isolated tools, polymer databases, predictive models, AI agents and automated laboratories are increasingly converging into interconnected discovery ecosystems. As a result, the central challenge is shifting from improving predictive accuracy alone to enabling reliable decision-making, adaptive learning and seamless integration across computation, experimentation and scientific reasoning. We argue that polymer science is entering an era of autonomous discovery, in which data, simulation, reasoning and experimentation operate within self-improving feedback loops that continuously generate hypotheses, design materials, execute experiments and refine predictive models. By unifying molecular design, process optimization, experimental validation and industrial translation, such autonomous ecosystems establish a more predictive, reproducible and scalable paradigm for polymer innovation, fundamentally transforming how polymer research is conducted.

physics.chem-ph↗
arXiv2026

Multidimensional Design of Metal-Nitrogen-Carbon Electrocatalysts for Direct Propylene Epoxidation

Propylene oxide is a major industrial chemical whose production currently relies on hazardous chlorine- or peroxide-based oxidants. Direct electrochemical epoxidation using water as the oxygen source offers a sustainable alternative, but controlling oxygen-atom transfer against the competing oxygen evolution reaction remains a fundamental challenge. Here, we show that propylene epoxidation selectivity cannot be described by oxygen binding energy alone, but is jointly governed by oxygen adsorption, the potential of zero charge, and applied potential. By combining theoretical calculations with pH-field-coupled microkinetic modeling across 41 metal-nitrogen-carbon single-atom catalysts, we first identified an optimal oxygen-binding window and Co as the most favorable metal center. We then found that peripheral substituents can tune the PZC while largely preserving the optimal oxygen adsorption energetics, thereby providing an independent design dimension to further optimize the already favorable Co active site. This sequential, multidimensional design strategy identified CoPc-NH2-CNT as the optimal catalyst, delivering a record PO Faradaic efficiency of 70-80 percent for direct propylene epoxidation in aqueous electrolyte under ambient conditions. These results establish interfacial electrostatics as an independently tunable design dimension for controlling selective oxygen-atom transfer in electrocatalysis.

physics.chem-ph↗
arXiv2026

Listen, See and Track: Spatio-Temporal Audio-Visual Sound Event Reasoning for Omni-Modal Language Models

Understanding dynamic sound sources requires jointly determining what produces a sound, where the source is located, and how it moves over time. Yet existing audio-language models often represent clips as global acoustic events, while vision-language models lack the spatial audio cues needed to localize and track individual sources. To evaluate this missing capability, we introduce ST-OmniQA, a spatio-temporal audio-visual question-answering benchmark built from panoramic videos paired with synchronized first-order Ambisonics (FOA) audio of moving sound sources. It contains 40K videos and 400K question-answer pairs organized into four capability levels covering sound-event recognition, direction of arrival, source distance, motion trajectories, and temporally grounded audio-visual reasoning. Building on this benchmark, we propose ST-Omni-R1, which integrates FOA-derived semantic and trajectory representations with panoramic visual context and is trained through progressive curriculum learning and reasoning-tree reinforcement learning. ST-Omni-R1 achieves 77.83\% average semantic accuracy across the four levels, compared with 37.28\% for the best evaluated baseline. Results on three public spatial-audio benchmarks further indicate that its learned spatial and motion representations transfer beyond ST-OmniQA.

cs.AI↗
arXiv2026

RFCheck: Synthetic RF Sensing Data Can Fail Measurement Consistency

Synthetic radio-frequency (RF) sensing data are widely used to augment wireless sensing tasks, yet their measurement consistency with real data is rarely evaluated under matched acquisition conditions. This paper identifies a measurement-consistency failure mode: synthetic samples may pass task-facing checks while deviating from the measurement behavior of real samples collected and processed by the same sensing pipeline, potentially introducing synthetic shortcuts and biasing downstream model selection. We propose RFCheck, a calibrated measurement audit that uses held-out real data from the same acquisition and preprocessing pipeline as the reference. RFCheck calibrates representation-specific tests on real samples and flags synthetic samples whose responses exceed the calibrated real-data range. We use the audit for candidate screening and residual repair. We validate RFCheck primarily on Wi-Fi channel state information (CSI), where the audit examines delay-domain and local frequency-domain structures. Experiments show that aggregate statistics and label-based screening can miss measurement failures detected by RFCheck. Under the same label acceptance rule, low-risk and high-risk synthetic candidates exhibit different downstream behavior. A repair reference reduces the flagged ratio to 10.83% while preserving mean task performance. In a held-out proposal study, correction followed by calibrated selection produces a class-balanced set with no flagged samples under a fixed training budget. We further apply the same calibration principle to frequency-modulated continuous-wave (FMCW) millimeter-wave radar gesture sensing. The results show that synthetic RF sensing data can violate measurement consistency even when conventional task checks are satisfied, motivating measurement-aware diagnosis and mitigation before augmentation.

eess.SP↗
arXiv2026

Thinking in Scales: Accelerating Gigapixel Pathology Image Analysis via Adaptive Continuous Reasoning

Traditional whole slide image (WSI) analysis methods typically rely on the multiple instance learning (MIL) paradigm, which extracts patch-level features at high magnification and aggregates them for slide-level prediction. However, such exhaustive patch-level processing is computationally expensive, severely limiting the efficiency and scalability of WSI analysis. To address this challenge, we propose PathCTM (a Pathology-oriented Continuous Thought Model) that enables token-efficient scale-space continuous reasoning for gigapixel WSIs. PathCTM formulates diagnostic inference as a dynamic sequential information pursuit. It progressively transitions from low-magnification global to high-magnification local inspection, and adaptively terminates inference when sufficient evidence is gathered to effectively bound decision uncertainty. Specifically, it uses conditional computation for dynamic scale switching with attention-guided region pruning, coupled with confidence-aware early stopping. Extensive experiments demonstrate that, compared with standard MIL-based methods, PathCTM reduces the number of required image patches by 95.95% and shortens inference time by approximately 95.62%, while maintaining AUC without degradation. Code is available at https://github.com/JSGe-AI/PathCTM.

cs.CV↗
arXiv2026

GWM-VLA: Geometry-Aware Latent World Modeling for Vision-Language-Action Learning

Vision-Language-Action (VLA) models achieve strong robotic manipulation performance but often degrade under visual and environmental shifts. Latent world modeling offers a promising approach to improving robustness, yet existing methods commonly encode camera views independently and predict holistic scene dynamics without explicitly modeling their geometric relationships. We propose GWM-VLA, a geometry-aware latent world modeling framework for VLA learning. GWM-VLA combines geometry-aware multi-view state encoding, global context-conditioned target-view prediction, and shared latent-action representations grounded by robot-action supervision. Specifically, VGGT-$Ω$ jointly aggregates multi-view observations at each timestep to construct geometry-aware multi-view states. The latent world model predicts the next-step patch tokens of a selected target view using patch and register tokens obtained after multi-view aggregation, thereby retaining multi-view geometric information without predicting the complete multi-view state. We use the wrist view as the target in our experiments, placing greater emphasis on end-effector motion and local gripper-object interactions. Finally, the shared latent-action representations condition both the latent world model and the flow-matching action head, allowing latent-prediction supervision and ground-truth robot-action supervision to jointly shape the same latent-action representations. Experiments across both simulation and real-world environments demonstrate the effectiveness and robustness of GWM-VLA.

cs.RO↗
arXiv2026

From Patches to Evidence Balls: Class-Conditioned Evidence Retrieval for Few-Shot Whole Slide Image Classification

Whole slide image (WSI) classification is an evidence-driven task, where diagnostic cues are often sparse, spatially organized, and class-dependent. Existing MIL and vision-language methods aggregate a large pool of patch features into a single global slide representation. Under few-shot supervision, limited slide-level labels make it difficult to learn a reliable aggregation mechanism that organizes sparse local cues into compact and coherent diagnostic evidence. Moreover, a shared slide representation compresses evidence supporting a candidate class and its alternatives into the same feature, limiting class-specific reasoning and interpretability. To address these issues, we propose EviBall, a class-conditioned evidence retrieval framework for few-shot WSI classification. EviBall organizes local patches into Evidence Balls through semantic-spatial assignment and center refinement, yielding compact and spatially coherent evidence units under weak supervision. It then uses task-specific class queries, including language-guided queries for morphology-oriented tasks and molecular-guided queries for molecular endpoint prediction, to retrieve supporting evidence balls and produce class-conditioned evidence representations for direct class-wise prediction. By introducing structured evidence units and task-relevant semantic guidance, EviBall reduces the reliance on learning an unconstrained global aggregation mechanism from scarce slide-level labels. It therefore reformulates few-shot WSI classification as structured evidence retrieval and competition among candidate classes. Extensive experiments across four morphology-oriented and molecular endpoint WSI tasks demonstrate that EviBall consistently outperforms conventional and vision-language MIL baselines under diverse few-shot settings, while providing spatially localized and class-specific evidence for each prediction.

cs.CV↗
arXiv2026

HiFi-UMI: Learning Deployable Manipulation Policies from High-Fidelity UMI Data Alone

Learning deployable manipulation policies is bottlenecked by the scarcity of data that is both high-fidelity and scalable. Real-robot teleoperation is accurate but costly to scale; robot-free UMI capture scales readily, and current practice uses the resulting data mainly for pre-training, adding a small real-robot "anchor" at post-training. We ask whether raising the fidelity of robot-free UMI data, rather than shrinking the real-robot fraction, can remove that anchor. We present HiFi-UMI, a portable UMI data-production system co-designed for trajectory accuracy, inter-gripper relative pose, synchronization, and field of view: head-mounted offline stereo-inertial SLAM, native rather than reconstructed relative pose, a shared microsecond GPIO trigger, and two wide-angle cameras per hand covering ~200 degrees. It reaches 3 mm workspace-local end-effector accuracy without external tracking infrastructure. Using this corpus, we demonstrate zero-robot post-training: a policy post-trained solely on HiFi-UMI demonstrations deploys directly on a real robot and matches in-domain teleoperation across three backbones spanning the vision-language-action and world-action-model families, with success-rate differences of -2.5, +3.1, and -0.6 percentage points on StarVLA-QwenPI, OpenPI-pi_0.5, and LingBot-VA; the strongest policy reaches 85% on a precision insertion task, even though the teleoperation baseline is collected in the evaluation scene and no HiFi-UMI trajectory is. Pre-training on 4,000 hours from the same corpus lowers action error on ten unseen tasks by 41% and, on StarVLA-QwenPI, raises real-robot success by a further 18.1 percentage points. We open-source HiFi-UMI-2K, 2,000 hours of microsecond-synchronized, ultra-wide-FoV demonstrations, each automatically reconstructed and validated through simulation replay, as a large-scale, high-fidelity resource for the robot-learning community.

cs.RO↗
arXiv2026

LongStraw: Long-Context RL Beyond 2M Tokens under a Fixed GPU Budget

Long-context RL post-training is constrained by the lifetime of state and gradients, not attention cost alone. In GRPO, one multi-million-token prompt must serve old-policy and reference scoring plus multiple policy responses, while conventional autograd keeps the prompt graph and all response graphs live alongside model weights, caches, and distributed communication buffers. We present LongStraw, an objective-aware, architecture-aware system for resident-state virtualization, response replay, and distributed-gradient execution. Its transaction captures the shared prompt without autograd, retains only the architecture-required state on explicitly owned pages, restores that state for each group member, scores old/reference branches without a graph, replays one policy response at a time with autograd, and accumulates the resulting gradients before one distributed finalization and optimizer step. This schedule bounds the live training graph by the response suffix while reusing the expensive prompt computation across the complete GRPO group. We instantiate this design for two incompatible model structures. Qwen3.6-27B combines 48 recurrent GDN layers with 16 full-attention layers; LongStraw keeps the compact recurrent state and physically CP8-sharded KV pages, composes global attention through cross-rank LSE/output merging, and performs blockwise response replay. GLM-5.2 combines a 78-layer MLA/DSA attention stack with a 256-expert, top-8 MoE tail. Its implementation keeps CP-sharded MLA latent pages and DSA indexer-key pages in CPU memory, stages one layer at a time, reconstructs IndexShare-aware global sparse selection over CP32, and dispatches routed response tokens over EP32. The two paths share one transaction contract while specializing the retained state, replay operator, and collective communication to the architecture...

cs.LG↗
arXiv2026

From Materials Database to Materials Bank: Assetizing Data for AI Driven Materials Innovation

Driven by high-throughput experimentation, computational modeling, and artificial intelligence (AI), materials data has expanded at an unprecedented rate. Conventional materials databases function only as passive repositories, archiving raw experimental records indiscriminately including both successful and failed data, without systematic value filtering or asset management. This creates a critical gap between massive data accumulation and actionable innovation, hindering the identification of high-potential materials and industrial translation. To address this bottleneck, we propose an industrialization-oriented Materials Bank, a dedicated valuefiltering and assetization layer that operates beyond traditional databases. It does not merely curate high-quality data but systematically elevates qualified candidates into standardized, upgradable materials assets via a multi-dimensional BankCard framework covering scientific validity, synthesis feasibility, application readiness, and industrial value. By unifying databases, AI models, automated experimentation, and multi-criteria assessment into a cohesive closed-loop ecosystem, the Materials Bank establishes a clear trajectory from data to knowledge, candidate, asset, and product. It serves not as an enhanced database or screening tool, but as a decision infrastructure bridging academic discovery and industrial demand, offering a scalable paradigm to accelerate AI-driven materials innovation and deliver tangible real-world impact.

cond-mat.mtrl-sci↗
arXiv2026

Bigger Is Safer: Provable Robustness in In-Context Learning Scales with Capacity

In-context learning (ICL) allows large language models to adapt to new tasks from a few examples without updating their parameters. Existing theories explain ICL by assuming the test task distribution matches pretraining -- an assumption that breaks down under adversarial distribution shifts. We introduce a distributionally robust meta-learning framework that provides worst-case guarantees for ICL under Wasserstein-based distribution shifts. Focusing on linear self-attention Transformers, we derive a non-asymptotic bound connecting adversarial perturbation strength ($ρ$), model capacity ($m$), and the number of in-context examples ($N$). The analysis reveals that the maximum safe perturbation radius scales as $ρ_{\max} \propto \sqrt{m}$, while maintaining performance under adversarial shift requires additional in-context examples with $N_ρ- N_0 \propto ρ^2$. Experiments on synthetic tasks confirm these scaling laws, and experiments on 21 real pretrained models (0.1B--7B parameters, 5 families) provide qualitative evidence consistent with the theory's predictions, while revealing that ICL capability is a prerequisite for robustness. These findings advance the theoretical understanding of ICL under adversarial conditions and formalize the sense in which larger models are safer under distributional shift.

cs.LG↗
arXiv2026

Positional Attention-based Graph Neural Network for Learning Permutation Non-equivariant Wireless Policies

Graph neural networks (GNNs) have emerged as a promising approach to learning wireless policies efficiently by leveraging topology prior and incorporating relational inductive biases. However, when the optimal policy is not permutation equivariant (PE), conventional GNNs suffer from mismatched inductive biases, leading to degraded performance or poor generalizability. This issue arises in wireless tasks with expected objectives, such as channel estimation and end-to-end (E2E) precoding, where the PE property of the optimal policy depends on the underlying channel distribution. In this paper, we propose a novel positional attention-based GNN to learn permutation nonequivariant policies efficiently. The core idea is to incorporate relative positions of vertices into the attention mechanism via an embedding function, enabling the GNNs to capture asymmetric relationships. Consequently, the proposed GNN can represent permutation non-equivariant functions, while retaining high learning efficiency and size generalizability through parameter sharing. We consider channel estimation and E2E precoding as case studies, and prove that their policies are PE to users but not to antennas under spatially correlated channels. We employ the proposed GNN to learn the policies, where the embedding function is designed based on the channel covariance matrix. Simulation results demonstrate that the proposed GNN outperforms existing channel estimation and E2E precoding methods, requires fewer samples for training, and can be generalized to systems with different numbers of antennas and users.

eess.SP↗
arXiv2026

MasFACT: Continual Multi-Agent Topology Learning via Geometry-Aware Posterior Transfer

Multi-agent systems (MAS) powered by large language models (LLMs) have emerged as a powerful paradigm for complex problem solving, where performance critically depends on the underlying inter-agent communication topology. However, existing topology generation methods mainly optimize for isolated tasks, while real-world deployments involve streams of evolving tasks, requiring previously effective collaboration patterns to be retained and reused rather than rediscovered or overwritten. We identify a previously underexplored failure mode, \emph{topology forgetting}, in which adapting to new tasks shifts the topology generator away from communication structures required by earlier tasks. This issue stems from cross-task misalignment in both agent-level functional semantics and relational communication structures. To address this challenge, we propose \textbf{\textsc{MasFACT}}, a geometry-aware posterior transfer framework that preserves and reuses historical collaboration knowledge as transferable topology priors. We transfer these priors across task-specific agent spaces through Fused Gromov-Wasserstein optimal transport and perform PAC-Bayes-guided conservative posterior adaptation to balance task-specific plasticity with structural stability. Experiments across class-, domain-, and task-level continual settings demonstrate that \textsc{MasFACT} consistently improves average accuracy while reducing topology forgetting compared to strong topology generation and replay-based baselines, and can be seamlessly integrated with different MAS topology generators.

cs.LG↗
arXiv2026

Selenium direct doping obtained high-performance-n-type Bi2Te3-based thermoelectric materials with a wide temperature range

The article reports on a series of n-type Bi2Te3-based thermoelectric materials prepared via a high-temperature melting combined with annealing process. The effects of Se doping content and annealing process on the carrier concentration, suppression of the bipolar effect, and thermoelectric performance of the materials were systematically investigated. The experimental results provide valuable reference for researchers in this field.

cond-mat.mtrl-sci↗
arXiv2026

Breaking Bottlenecks in Solid Electrolyte Discovery with Large Artificial Intelligence Models

Solid electrolytes (SEs) are central to next-generation metal batteries, yet their discovery remains constrained by fragmented data, limited transferability of simulations, and slow experimental iteration. Unlike catalysis, where surface reactivity dominates, SEs require simultaneous optimization of bulk ion transport, defect chemistry, mechanical integrity, and interfacial stability. Here, we outline a framework for autonomous SE discovery enabled by large artificial intelligence (AI) models, including machine learning interatomic potentials (MLIPs) and large language models (LLMs). We discuss the evolution from static materials databases to dynamic, self-updating knowledge systems, the role of MLIPs in bridging density functional theory (DFT) and long-timescale ion migration, and the emergence of LLMs as engines for literature mining, hypothesis generation, and scientific reasoning. We further describe a closed-loop architecture integrating AI-driven candidate design, multiscale simulation, uncertainty-aware selection, and experimental validation. Such systems shift SE research from intuition-guided exploration to data-informed, self-improving cycles. We conclude by highlighting challenges in data standardization, interfacial complexity, and reproducibility, and we propose design principles for building autonomous laboratories for solid-state battery materials.

cond-mat.mtrl-sci↗