Searcharxiv⌕ Search

arXiv subjects

Hao Ding

Publications and source records attributed to Hao Ding.

At least 19 recordsLinked to original sources
arXiv2026

Discovering Subtypes of Neurodegenerative Progression with a Scalable Connectome-Constrained Dynamic Model

Parkinson's disease is clinically and biologically heterogeneous, yet its spatiotemporal progression remains poorly characterized. We present a connectome-constrained disease progression model that jointly estimates subject-specific disease time and data-driven subtypes from longitudinal morphometry. Applied to 85 imaging and clinical biomarkers from the Parkinson's Progressive Markers Initiative (PPMI) cohort, the model recovers four morphologically distinct progression subtypes. We validate the model on a hold-out cross-sectional dataset and benchmark it against SuStaIn under a matched training and validation protocol. Only our method recovers subtypes that correspond significantly to clinical motor subtypes and genetic variants of Parkinson's Disease.

q-bio.QM↗
arXiv2026

V-Rubrics: Visual Faithfulness via Rubric-Based Reinforcement Learning

Vision-language models can produce fluent answers that are insufficiently grounded in the visual evidence: a single unsupported object, chart value, or intermediate inference can undermine an otherwise plausible response. We argue that this is a credit-assignment failure in multimodal post-training. Scalar outcome rewards indicate whether an answer is acceptable, but do not identify which visual facts are grounded, which reasoning steps are valid, or which instruction constraints are missed. We introduce Visual Rubrics-Based Reinforcement Learning, which decomposes reference responses into atomic propositions and scores generated answers along Visual Faithfulness (VF), Reasoning Consistency (RC), and Instruction Following (IF). The resulting rubric items provide structured partial credit and localize rubric credit when supporting evidence spans are available. We first obtain an SFT checkpoint by fine-tuning Qwen3-VL-8B-Instruct on the public OpenMMReasoner-SFT-874K corpus, adapting OpenMMReasoner's cold-start data recipe. We construct V-Rubrics 50K, a 50,248-example training set from 17 visually grounded sources, by applying rule-based filters before deriving example difficulty from rejection-sampling scores and then annotating every example with Gemini-3-Pro under the same structured prompt and protocol. We train our model based on the same SFT checkpoint using component-wise, prefix-localized rubric credit. Experiments show that our rubricbased GRPO improves over both the shared SFT baseline and answer-only GRPO, with the largest gains on knowledge-oriented and visually grounded reasoning benchmarks. The results show rubrics as a useful reward abstraction for visual post-training.

cs.CV↗
arXiv2026

Predicting Brain Morphometry with MT-GNN: Mesh Evolution in Continuous Time with Graph-Based Metric Tensor Embeddings

Predicting how a subcortical structure's shape will evolve from a few prior scans could support prognosis and clinical-trial enrichment. Existing longitudinal mesh predictors either extrapolate shape trajectories via high-dimensional embeddings or regress vertex deformations directly. We instead predict the surface's intrinsic geometry in continuous time: a single per-structure graph network predicts the future per-vertex first fundamental form (metric tensor) for an arbitrary causal multiple-visit history and an arbitrary prediction horizon, conditioned on a Fourier encoding of the lead time. The predicted metric is decoded into a surface by a differentiable As-Rigid-As-Possible solver, and the model is trained end-to-end on the rigid-aligned vertex error. Training through the reconstruction keeps the decoded prediction a valid surface and consistently improves it. On 14 subcortical structures from the ADNI dataset, the proposed mesh evolution model (MT-GNN) predicts best among the evaluated methods at every horizon ($-2.29\%$ mean vertex error vs. the temporal mean, $p{=}6.1{\times}10^{-5}$, beating it on 14/14 structures), ahead of geodesic shape regression (DCM, $-0.19\%$) and a mesh transformer (TransforMesh, $-0.45\%$; $p{=}1.2{\times}10^{-4}$), with the lead widening as the horizon grows.

cs.CV↗
arXiv2026

From Simple QA to Deep Research: A Verifiable Benchmark Constructed through Iterative Task Evolution

Deep research benchmarks require expert-level tasks and reliable evaluation grounded in task-specific knowledge. Existing benchmarks rely heavily on expert authoring or pre-existing human-authored materials, while fully automatic construction struggles to ensure consistent and traceable verification. To address this gap, we introduce a verifiable benchmark of 500 deep research tasks spanning 31 topics and 10 major categories, with three query forms designed to probe complementary capabilities required for deep research. The benchmark is constructed automatically using an iterative Explorer-Formalizer-Challenger pipeline that progressively transforms simple questions into deep research tasks. Each task is represented as a directed acyclic graph (DAG) of atomic steps and associated checkpoints, enabling the query, DAG, and rubrics to evolve together in a controlled manner. Experiments demonstrate that the benchmark clearly discriminates among models and query types, while its fact-grounded pointwise rubrics enable fine-grained, human-aligned, and stable evaluation. Our data, implementation, and results are publicly available.

cs.AI↗
arXiv2026

Kimi K3: Open Frontier Intelligence

We introduce Kimi K3, a 2.8T parameter Mixture-of-Experts model with 104 billion activated parameters, native vision capabilities, and a 1-million-token context window. Kimi K3 is built on Kimi Delta Attention and Attention Residuals, which improve information flow across sequence length and model depth. Together with Stable LatentMoE, which effectively activates 16 of 896 routed experts per token, and refined training and data recipes, these advances yield an approximately 2.5x improvement in overall scaling efficiency over Kimi K2. Post-training highlights reinforcement learning across general, agentic, and coding domains and multiple reasoning-effort levels, enabling compositional generalization and robust long-horizon execution. At 2.8T scale, Kimi K3 is supported by infrastructure advances in multiple areas: algorithm-system co-design for KDA, perfectly balanced expert-parallel training with efficient memory management, million-token agentic RL with persistent rollout and sandbox states, and deployment innovations. Extensive evaluations show that Kimi K3 achieves frontier-level performance across long-horizon coding, agentic, knowledge, reasoning, and vision tasks. While its overall performance still trails the most powerful proprietary models, namely Claude Fable 5 and GPT-5.6 Sol, Kimi K3 consistently outperforms other open and proprietary models evaluated in our suite. We release the full Kimi K3 model weights to facilitate future research and accelerate the broader deployment and adoption of frontier intelligence.

cs.CL↗
arXiv2026

HEMERA: A Heterogeneous Memory-Centric Accelerator with Recursive Dataflow for Edge-Constrained State-Space-Duality Models Inference

Structured State Space Models (SSMs), such as Mamba, enable efficient long-sequence modeling with linear time complexity. Recent implementations realize this capability through Structured State Space Duality (SSD), which transforms recursive state evolution into matrix-form computations. However, SSD introduces substantial system-level overheads, including quadratic intermediate materialization, irregular data movement, and prefix-dependent execution, leading to excessive memory traffic and bandwidth demand on conventional architectures. Although prior accelerators mitigate these overheads through optimized dataflows or compute-in-memory techniques, they largely retain matrix-oriented SSD execution and cannot simultaneously avoid quadratic intermediate storage and efficiently map dependency-bound state propagation. This paper presents HEMERA, a heterogeneous memory-centric accelerator for efficient Mamba-2 inference. Rather than directly executing the matrix-form SSD computation, HEMERA reformulates it into an algebraically equivalent streaming-recursive dataflow that avoids quadratic intermediate storage while preserving the original computation. The resulting heterogeneous execution paradigm maps dense linear operations onto in-memory computing units and recursive state updates onto a dedicated streaming engine. Across Mamba-2 models ranging from 130M to 2.8B, HEMERA achieves average latency speedups of 1.4x-3.6x and energy-efficiency improvements of 12.2x-27.0x over the official optimized fused Mamba-2 kernel on NVIDIA A100. It further reduces the average SSD-related execution-time ratio across model scales to 14.12% during long-sequence inference, demonstrating its potential for efficient deployment under edge constraints.

cs.AR↗
arXiv2026

Dense Structural Priors for Sparse Functional Landmark Localization in Surgical Videos

Vision foundation models such as SAM 3 can provide transferable object-level structure across diverse surgical video conditions, but segmentation outputs do not explicitly encode the action-conditioned semantics that define functional surgical landmarks. Estimating instrument extent and geometry differs from localizing the tip or anchor relevant to clipping, grasping, or dissecting. We investigate vision foundation model-enabled sparse action-aware landmark localization, using zero-shot, point-prompted structural masks to provide dense instrument-level context without manual pixel-level mask annotations. We propose a lightweight refinement framework that uses SAM 3 as a structural prior. A coarse multi-frame network predicts tip and anchor prompts, generating non-oracle masks that are fused with visual and heatmap features to refine functional landmark predictions. We compare direct mask-augmented supervision, prediction-derived mask-prior refinement, and auxiliary mask supervision to examine how vision foundation model-derived structure should enter a precision-oriented localization system. Experiments on 7,867 clips from 60 surgical videos spanning YouTube, Cholec80, HeiChole, SurgVU, and CRCD evaluate the approach under heterogeneous conditions. Without manual pixel-level mask annotations for training, the proposed model achieves overall F1 scores of 72.4% for tip and 58.0% for anchor localization. Ablations show that coarse-to-fine refinement provides a substantial performance gain, while prediction-derived structural priors provide additional improvement when incorporated as intermediate guidance rather than direct localization targets.

cs.CV↗
arXiv2026

MGDiff: Multi-Interest Sequence Recommendation with Masking GNN-Guided Diffusion

We propose a novel Multi-Interest Sequence Recommendation Framework with \underline{M}asking \underline{G}NN-Guided \underline{Diff}usion Model (MGDiff), designed to generate accurate, bias-free user interest information during the diffusion process. First, we propose a semantics-enhanced Dual-layer Semantic Guidance (DSG) framework, which decomposes guidance into two synergistic stages: extracting latent item semantics and decoupling multidimensional user intent. We design a Weight-adaptive Masking Graph Neural Network reconstructs missing links to uncover deep item relationships beyond superficial co-occurrence, while a Dynamic Multi-Expert Network projects user preferences into distinct semantic subspaces to suppress irrelevant interference. This hierarchical design yields structured guidance that significantly improves the generation accuracy of diffusion models. Second, We propose a Popularity-Aware Guidance (PAG) mechanism that performs spatial geometric adjustments on the outputs of diffusion models: by using item popularity as a differentiable adjustment signal to recalibrate similarity metrics, we enable DMs to generate diverse recommendations free from popularity bias. Finally, we compare MGDiff with multiple baseline models across four widely used datasets, demonstrating its superior performance and validating its effectiveness.

cs.IR↗
arXiv2026

VLBI Astrometry of Magnetars

The origin of the strongest magnetic fields in the Universe, i.e., the origin of magnetars, is a longstanding question. An enhanced dynamo effect in an irregular supernova explosion is a possible origin, which implies a stronger kick velocity and a higher proper motion of the magnetar compared to those of ordinary pulsars as well as an irregular morphology of the host supernova remnant (SNR). However, this hypothesis is not well studied yet, because there is a lack of precise measurement of the proper motion of magnetar and of identification of the host SNR. VLBI astrometry of magnetars is a unique tool to examine the hypothesis. In this chapter, we introduce the MONSTER (Monitoring Observations of the Neutron Stars That Evolve Rapidly) Project. SKA-VLBI's unprecedented sensitivity and the highest angular resolution will allow us to dramatically expand the survey volume in which we can measure the proper motion of magnetars within a radio outburst period of a few months.

astro-ph.HE↗
arXiv2026

Reasoning Text-to-Video Retrieval for Operating Room Clips via Action-Driven Digital Twins

Text-to-video retrieval in operating rooms (OR) is an enabling technology for OR safety, as it allows stakeholders to retrieve and inspect recordings of specific events. However, because the most safety-critical events may not follow the common structure, to unlock its full potential text-to-video retrieval must be able to handle implicit queries that require reasoning to identify the right video (e.g., the step right before clipping). However, existing methods rely on global embeddings that cannot reason over such queries. We propose OR3, a text-to-video retrieval method that converts clips into action-driven digital twins (ActDTs), grouping concurrent subject-action-object triplets under non-overlapping temporal intervals. Moreover, rather than cross-modal matching through paired encoders, OR3 performs imagination-based retrieval where an LLM generates hypothetical ActDTs from queries. This enables intra-modal matching via a single encoder trained with ActDT-tailored hard negatives. Finally, evidence-grounded refinement revises imagined ActDTs based on discrepancies with top candidates to capture procedure-specific patterns. We construct a benchmark from MM-OR with 276 implicit queries across four reasoning categories over 386 clips from robotic knee procedures. OR3 achieves 57.6 R@1 and 77.3 R@5, outperforming the strongest baseline. These results demonstrate that OR3 enables fine-grained discrimination between visually similar OR video clips through temporal action reasoning.

cs.CV↗
arXiv2026

Geometry-Consistent Endoscopic Representations for Image-Guided Navigation via Structured Foundation Model Adaptation

Accurate vision-based navigation in monocular endoscopy is difficult due to limited depth cues, weak tissue texture, non-rigid deformation, and substantial appearance variation across domains, all of which complicate pose estimation, depth prediction, and image-to-anatomy alignment. Although recent vision foundation models have shown promise, their learned representations often remain insufficiently geometry-consistent, hindering stable feature correspondence and limiting their reliability for downstream navigation tasks. We propose a unified framework for learning geometry-consistent and domain-robust image representations for monocular endoscopy. The framework combines a synthetic data pipeline that provides accurate geometric supervision with Hierarchy-Aware Geometry-Semantic Adaptation, a structured alternative to standard LoRA that inserts low-rank adapters selectively across the transformer hierarchy and couples them with layer-wise training objectives to encourage geometric correspondence in intermediate features and semantic consistency in deeper features. Experiments on public and proprietary datasets show improved geometric and semantic representation quality, leading to better performance on downstream navigation tasks including pose estimation and monocular depth estimation. The learned representations show favorable synthetic-to-real transfer on clinical bronchoscopy and provide a useful initialization for adaptation to sinus endoscopy and colonoscopy under limited supervision. The framework also shows favorable scaling with model size and training data. These results support hierarchy-aware, geometry-guided adaptation as a practical approach for endoscopic representation learning.

cs.CV↗
arXiv2026

Satellite gravity constraints on inner core viscosity and LLVPs density anomalies

Constraining the physical properties of Earth's deep interior, particularly the viscosity of the solid inner core and the density structure of large low-velocity provinces (LLVPs), remains a major challenge in geophysics. Here we develop a unified dynamical framework that combines mantle-inner core gravitational coupling (MICG) with torsional oscillations in the fluid outer core and show that their interaction can produce a distinct and testable geodetic signature. Guided by this prediction, we analyze satellite gravity observations together with independent corrections for surface mass variability. We identify a robust approximately 6-year signal in the Stokes coefficient Delta S22, while no corresponding stationary signal is detected in Delta C22. A signal with the same periodicity is independently detected in length-of-day variations (Delta LOD), and the two signals exhibit a near anti-phase relationship. Interpreting this coupled signature within the proposed framework allows us to constrain the inner core viscosity to approximately 4.6 (+/- 1.8) x 10^16 Pa s and the equatorial relief of the inner core boundary to a semi-axis difference of about 200 +/- 70 m. The inversion further indicates mean density anomalies of +5.5 (+/- 0.6) per mil at the base of LLVPs. These results indicate that satellite gravimetry provides a direct observational window into deep-Earth dynamics and the physical properties of Earth's deep interior.

physics.geo-ph↗
arXiv2026

Sparse Contextual Coupling Reshapes Diffusion Geometry in Multilayer Hypergraphs

Many complex systems combine dense background structure with sparse contextual information. We introduce a diffusion-based framework for analyzing how sparse condition-specific layers reshape diffusion geometry in multilayer hypergraphs. Each layer is represented as a weighted hypergraph, layers are coupled through shared entities, and random walks on the coupled system induce multiscale diffusion distances between nodes. We apply the framework to disease-conditioned gene networks by coupling a dense MSigDB functional gene-set layer to sparse disease-specific DGIdb drug-gene hypergraphs, with disease-associated drugs selected from DDDB and HumanNet-GSP used to define external gene weights. Across Bipolar Disorder, Schizophrenia, Leukemia, and Breast Cancer, the disease-specific layer contains less than 2 percent of genes in the coupled system, yet substantially changes diffusion distances and community structure. Centrality analysis suggests that this disproportionate effect arises because DGIdb-associated genes occupy influential positions in the MSigDB-derived functional network. The resulting diffusion-derived communities are stable under subsampling and show coherent post hoc functional enrichment, including signaling and neurotransmission categories in neuropsychiatric diseases and immune, translational, and metabolic categories in cancer-associated diseases. Community-level comparisons further reveal disease similarities not reducible to direct DGIdb gene overlap, including a Breast Cancer-Schizophrenia relationship consistent with recent biomedical evidence. These results show that sparse contextual layers can induce interpretable nonlocal changes in higher-order network geometry.

physics.soc-ph↗
arXiv2026

Open-H-Embodiment: A Large-Scale Dataset for Enabling Foundation Models in Medical Robotics

Autonomous medical robots hold promise to improve patient outcomes, reduce provider workload, democratize access to care, and enable superhuman precision. However, autonomous medical robotics has been limited by a fundamental data problem: existing medical robotic datasets are small, single-embodiment, and rarely shared openly, restricting the development of foundation models that the field needs to advance. We introduce Open-H-Embodiment, the largest open dataset of medical robotic video with synchronized kinematics to date, spanning more than 50 institutions and multiple robotic platforms including the CMR Versius, Intuitive Surgical's da Vinci, da Vinci Research Kit (dVRK), Rob Surgical BiTrack, Virtual Incision's MIRA, Moon Surgical Maestro, and a variety of custom systems, spanning surgical manipulation, robotic ultrasound, and endoscopy procedures. We demonstrate the research enabled by this dataset through two foundation models. GR00T-H is the first open foundation vision-language-action model for medical robotics, which is the only evaluated model to achieve full end-to-end task completion on a structured suturing benchmark (25% of trials vs. 0% for all others) and achieves 64% average success across a 29-step ex vivo suturing sequence. We also train Cosmos-H-Surgical-Simulator, the first action-conditioned world model to enable multi-embodiment surgical simulation from a single checkpoint, spanning nine robotic platforms and supporting in silico policy evaluation and synthetic data generation for the medical domain. These results suggest that open, large-scale medical robot data collection can serve as critical infrastructure for the research community, enabling advances in robot learning, world modeling, and beyond.

cs.RO↗
arXiv2026

Compressing Sequences in the Latent Embedding Space: $K$-Token Merging for Large Language Models

Large Language Models (LLMs) incur significant computational and memory costs when processing long prompts, as full self-attention scales quadratically with input length. Token compression aims to address this challenge by reducing the number of tokens representing inputs. However, existing prompt-compression approaches primarily operate in token space and overlook inefficiencies in the latent embedding space. In this paper, we propose K-Token Merging, a latent-space compression framework that merges each contiguous block of K token embeddings into a single embedding via a lightweight encoder. The compressed sequence is processed by a LoRA-adapted LLM, while generation remains in the original vocabulary. Experiments on structural reasoning (Textualized Tree), sentiment classification (Amazon Reviews), and code editing (CommitPackFT) show that K-Token Merging lies on the Pareto frontier of performance vs. compression, achieving up to 75% input length reduction with minimal performance degradation. Code is available at https://github.com/shsjxzh/K-Token-Merging.

cs.CL↗
arXiv2026

AffordTissue: Dense Affordance Prediction for Tool-Action Specific Tissue Interaction

Surgical action automation has progressed rapidly toward achieving surgeon-like dexterous control, driven primarily by advances in learning from demonstration and vision-language-action models. While these have demonstrated success in table-top experiments, translating them to clinical deployment remains challenging: current methods offer limited predictability on where instruments will interact on tissue surfaces and lack explicit conditioning inputs to enforce tool-action-specific safe interaction regions. Addressing this gap, we introduce AffordTissue, a multimodal framework for predicting tool-action specific tissue affordance regions as dense heatmaps during cholecystectomy. Our approach combines a temporal vision encoder capturing tool motion and tissue dynamics across multiple viewpoints, language conditioning enabling generalization across diverse instrument-action pairs, and a DiT-style decoder for dense affordance prediction. We establish the first tissue affordance benchmark by curating and annotating 15,638 video clips across 103 cholecystectomy procedures, covering six unique tool-action pairs involving four instruments (hook, grasper, scissors, clipper) and their associated tasks: dissection, grasping, clipping, and cutting. Experiments demonstrate substantial improvement over vision-language model baselines (20.6 px ASSD vs. 60.2 px for Molmo-VLM), showing that our task-specific architecture outperforms large-scale foundation models for dense surgical affordance prediction. By predicting tool-action specific tissue affordance regions, AffordTissue provides explicit spatial reasoning for safe surgical automation, potentially unlocking explicit policy guidance toward appropriate tissue regions and early safe stop when instruments deviate outside predicted safe zones.

cs.CV↗
arXiv2026

From Detection to Host Galaxy Identification: Precision Continuous Gravitational Wave Localization with a Few Anchor Pulsars

Pulsar Timing Arrays (PTAs) are rapidly advancing toward the detection of continuous gravitational waves from individual supermassive binary black holes. While it is well established that coherently utilizing the ``pulsar term" requires astrometric distance uncertainties to be smaller than the gravitational wavelength, achieving this precision across an entire array is observationally prohibitive. Here, we demonstrate that achieving sub-wavelength precision for a few ``anchor" pulsars is sufficient to phase-lock the array and drastically shrink the sky-localization error. Using 20 years of realistically simulated data, we systematically evaluate the localization performance of a 25-pulsar array containing three to six high-precision anchors. We show that while introducing three sub-wavelength anchors can reduce the 90\% credible sky area by a factor of 30 in certain directions, expanding this high-precision subset to six anchor pulsars ensures high-precision localizations across diverse source directions. Evaluating a representative set of sky directions, including local galaxy clusters and the locations of maximum and minimum array sensitivity, this six-anchor configuration yields 90\% credible localization areas ranging from $\sim 0.1$ to $9.2 \text{ deg}^2$ at a signal-to-noise ratio of 20. Furthermore, once this minimal subset crosses the sub-wavelength threshold, further reductions in distance uncertainty yield diminishing returns. This establishes a highly efficient near-term observational strategy: prioritizing intensive parallax campaigns for a small core of stable millisecond pulsars provides a cost-effective pathway to precision multi-messenger astronomy.

astro-ph.GA↗
arXiv2026

Spinon Singlet Pairing: Microscopic nature of plaquettes in stripy LDOS

Scanning tunneling microscopy (STM) is a powerful tool for visualizing the local density of states (LDOS) of individual stripes in cuprates. However, the microscopic nature of the observed exotic LDOS patterns and their connection to high-$T_c$ superconductivity remain open questions. Within the framework of the quantum colored string model, we reveal that the ubiquitous $4a_0\times4a_0$ plaquettes originate from either the breaking of local spinon singlet pairs through hole insertion, or the formation of an unpaired spinon upon electron addition in a stripe. Moreover, by comparing our data with LDOS of cuprates, we identify an effect of particle-hole symmetry breaking (PHSB): a $2a_0$ shift, which is predicted and confirmed in a longer stripe ($L=18$). At last, we establish and verify a general relation between hole density and plaquette size across multiple fillings. Our work offers a fresh wavefunction-based perspective on interpreting STM signals in cuprate experiments and demonstrates that their origin may arise from spinon singlet pairing in the ground state of fluctuating stripes, the same mechanism underlying the $d$-wave sign structure [Phys. Rev. Lett. \textbf{137}, 086702 (2026)].

cond-mat.str-el↗