arXiv subjects

Jianwen Chen

Publications and source records attributed to Jianwen Chen.

15 recordsLinked to original sources

arXiv2026

A Vision-Language Foundation Model for Precise and Comprehensive Brain Tumor Diagnosis from Preoperative Multimodal Data

Background Non-invasive presurgical diagnosis of brain tumor types from Magnetic Resonance Imaging (MRI) is essential but challenging due to overlapping imaging features across tumor types, inter-observer variability, and the extensive training required for expertise. We aimed to develop an MRI-based Artificial Intelligence (AI) model for automatic and reliable brain tumor classification with diagnostic uncertainty quantification and radiology reports generation. Methods We developed BrainVLM to classify all 12 World Health Organization (WHO) 2021 brain tumor types. BrainVLM integrates an uncertainty quantification strategy to indicate prediction reliability and a module for generating radiology reports to elucidate the clinical rationale. BrainVLM was trained on multi-modal data (MRI scans, demographics, and radiology reports) from 40,043 individuals. It was validated on 5,211 patients with pathologically confirmed brain tumors, including 3,877 held-out patients from the primary hospital and 1,334 patients from 11 independent hospitals. We further conducted two proof-of-concept studies to validate its clinical utility in AI-clinician workflows: 1) a blinded multi-reader study where 12 neuroradiologists across varying experience levels interpreted 248 retrospective cases with or without AI assistance, and 2) a real-world prospective study in which 1,009 patients were independently and blindly assessed by BrainVLM and radiologists before surgery. Additionally, we demonstrated BrainVLM's utility in preoperative molecular subgroup prediction for adult-type diffuse gliomas, using a multi-center cohort of 632 patients.

cs.CV↗

arXiv2026

Square-Root Price Impact Is Necessary for Endogenous Manipulation Cycles in Learning-Agent Markets

We study a minimal agent-based market in which a single evolutionary-optimized institutional agent interacts with 20{,}000 herding retail traders. The agent spontaneously discovers a multi-cycle predatory strategy, producing 8--11 complete cycles over 2000 trading days with total portfolio return of $+51\%$ (best of 20 seeds; mean $+37.7\%$). Mean-field reduction maps the system onto a nonlinear oscillator that undergoes two distinct bifurcations: a continuous Hopf transition as institutional capital exceeds a critical threshold $C_c$, with oscillation amplitude $A \propto (C-C_c)^α$ where $α$ is consistent with the standard prediction of $1/2$; and a discontinuous fold transition in the herding-scale parameter space. The limit cycle persists even at $β= 0$: position-tracking feedback coupled with square-root price impact creates a self-sustained nonlinear oscillator requiring no retail herding. Square-root impact is shown to be necessary: linear impact eliminates the Hopf bifurcation entirely and renders the retail market unconditionally stable. Manipulation cycles thus emerge as the optimal-control solution of a nonlinear dynamical system, and a structural analogy to Maxwell's demon frames the agent as an information-processing controller that reduces the entropy rate of the price process.

q-fin.CP↗

arXiv2026

Order Splitting and Liquidity Replenishment Are Jointly Necessary for the Square-Root Law of Market Impact:

Three quantitative predictions have been advanced for the square-root law (SRL) of market impact, $I/σ_D = c\,(Q/V_D)^δ$ with $δ\approx 0.5$: GGPS ($δ=β-1$), FGLW ($δ=α-1$), and LOB walking ($δ=1/(1+γ)$). Using a minimal limit-order-book model populated by heterogeneous interacting agents and calibrated against the Tokyo Stock Exchange benchmark ($\langleδ\rangle = 0.489$~\citep{satoStrictUniversalitySquareRoot2025}), we test all three on identical simulated data and find that none matches the per-stock measured $δ$: GGPS and FGLW over-predict by factors of two and four respectively, while LOB walking under-predicts. The model reproduces $\langleδ\rangle = 0.539\pm 0.048$ across 2000 independently parameterised stocks. To identify which mechanisms are causally responsible, we perform counterfactual ablation by selectively suppressing each component. Removing order splitting collapses $δ$ from $0.549$ to $0.324$; removing liquidity replenishment by market makers drops it to $0.386$; perturbations that leave both intact (momentum trading, price limits, splitting rule, background liquidity) move $δ$ by less than $10\%$. Order splitting and liquidity replenishment are thus jointly identified as the necessary mechanisms for the SRL within this model, with the simulated SRL depending on neither the metaorder size tail nor the visible book shape in isolation.

q-fin.CP↗

arXiv2026

Non-Hermitian Skin Effect Enhances Pairing Correlations in Moiré Hubbard Systems

We show that the non-Hermitian skin effect (NHSE) can enhance pairing correlations in moiré Hubbard systems through a channel-selective mechanism: skin-induced localization amplifies the boundary density of states, strengthening local pairing tendencies within an intermediate ``golden window'' of non-reciprocity $γ\in[0.5,1.2]\,t$. Using exact diagonalization of the non-Hermitian Hubbard model on triangular lattices with open boundaries, we map the $(U,γ)$ phase diagram. The double occupancy $D(γ)$ exhibits non-monotonic behavior -- rising by up to 21\% then declining -- reflecting a competition between NHSE-enhanced boundary pairing and over-localization. A decomposition of the pairing susceptibility $χ_{\mathrm{SC}}$ on the $3\times3$ cluster reveals that the NHSE acts \emph{channel-selectively}: it enhances on-site pairing ($+21\%$) while simultaneously suppressing competing antiferromagnetic correlations (22\% reduction), so that the total pairing susceptibility, dominated by the on-site channel, grows by $+98\%$ on that cluster. These trends are corroborated by an independent non-Hermitian DMRG calculation and establish an enhancement of finite-cluster pairing correlations rather than trivial density redistribution. We do not claim long-range superconducting order. A BCS scaling estimate converts the same pairing-response signal into a dome-shaped $T_c(γ)$ fingerprint, suggesting an experimentally distinguishable response in coherent-drive versus reservoir-dominated moiré devices.

cond-mat.str-el↗

arXiv2026

VisualClaw: A Real-Time, Personalized Agent for the Physical World

Vision language models are serving as general-purpose interfaces for complex multimodal tasks. However, deployment still faces three gaps: VLMs typically incur high latency and cost when processing dense video frames and long prompts, the agent scaffold remains static after deployment, and standard video-QA benchmarks do not test whether agents can use visual evidence inside tool-using workspaces. We present VisualClaw, a self-evolving multimodal agent built around two principles. First, hybrid encoding reduces deployment cost by filtering less informative streaming frames with a cascaded gate and compressing the text skill bank through hot/cold top-k injection. Second, skill evolution lets the agent learn from failures: retrieved memories condition an evolver as direct concatenated context or as guided evidence, producing skill-bank updates that help future questions. Across 4 video-QA benchmarks with 2 VLMs, VisualClaw cuts per-question API cost by an average -98% versus full-frame upload and by -25.9% over the offline uniform 8 frame baseline, while boosting accuracy in most settings, e.g., an average +3.85% and a peak +15.80% on EgoSchema with Gemini 3 Flash. To address the gap, we curate VisualClawArena, a 200-scenario multimodal agentic benchmark built through a strict five-stage pipeline; models must use video evidence, documents, dynamic updates, and executable checks inside a workspace. On VisualClawArena, the same framework with computer-use agent backends improves macro accuracy by +2.9% for Codex (GPT-5.5) and +3.2% for Claude Code (Sonnet 4.6) over no-evolution baselines, with a -9.5% cost reduction compared to the uniform-sampled baseline. These properties make VisualClaw a natural fit for edge applications, where the cascade reduces a 1-hour streaming session from ~3,600 API uploads down to only 5-20 calls and the self-evolution makes it a perfect personalized assistant.

cs.CV↗

arXiv2026

MedVerse: Efficient and Reliable Medical Reasoning via DAG-Structured Parallel Execution

Large language models (LLMs) have demonstrated strong performance and rapid progress in a wide range of medical reasoning tasks. However, their sequential autoregressive decoding forces inherently parallel clinical reasoning, such as differential diagnosis, into a single linear reasoning path, limiting both efficiency and reliability for complex medical problems. To address this, we propose MedVerse, a reasoning framework for complex medical inference that reformulates medical reasoning as a parallelizable directed acyclic graph (DAG) process based on Petri net theory. The framework adopts a full-stack design across data, model architecture, and system execution. For data creation, we introduce the MedVerse Curator, an automated pipeline that synthesizes knowledge-grounded medical reasoning paths and transforms them into Petri net-structured representations. At the architectural level, we propose a topology-aware attention mechanism with adaptive position indices that supports parallel reasoning while preserving logical consistency. Systematically, we develop a customized inference engine that supports parallel execution without additional overhead. Empirical evaluations show that MedVerse improves strong general-purpose LLMs by up to 8.9%. Compared to specialized medical LLMs, MedVerse achieves comparable performance while delivering a 1.3x reduction in inference latency and a 1.7x increase in generation throughput, enabled by its parallel decoding capability. Code is available at https://github.com/aiming-lab/MedVerse.

cs.LG↗

arXiv2026

MetaClaw: Just Talk -- An Agent That Meta-Learns and Evolves in the Wild

Large language model (LLM) agents are increasingly used for complex tasks, yet deployed agents often remain static, failing to adapt as user needs evolve. This creates a tension between the need for continuous service and the necessity of updating capabilities to match shifting task distributions. On platforms like OpenClaw, which handle diverse workloads across 20+ channels, existing methods either store raw trajectories without distilling knowledge, maintain static skill libraries, or require disruptive downtime for retraining. We present MetaClaw, a continual meta-learning framework that jointly evolves a base LLM policy and a library of reusable behavioral skills. MetaClaw employs two complementary mechanisms. Skill-driven fast adaptation analyzes failure trajectories via an LLM evolver to synthesize new skills, enabling immediate improvement with zero downtime. Opportunistic policy optimization performs gradient-based updates via cloud LoRA fine-tuning and Reinforcement Learning with a Process Reward Model (RL-PRM). This is triggered during user-inactive windows by the Opportunistic Meta-Learning Scheduler (OMLS), which monitors system inactivity and calendar data. These mechanisms are mutually reinforcing: a refined policy generates better trajectories for skill synthesis, while richer skills provide higher-quality data for policy optimization. To prevent data contamination, a versioning mechanism separates support and query data. Built on a proxy-based architecture, MetaClaw scales to production-size LLMs without local GPUs. Experiments on MetaClaw-Bench and AutoResearchClaw show that skill-driven adaptation improves accuracy by up to 32% relative. The full pipeline advances Kimi-K2.5 accuracy from 21.4% to 40.6% and increases composite robustness by 18.3%. Code is available at https://github.com/aiming-lab/MetaClaw.

cs.LG↗

arXiv2026

SkillRL: Evolving Agents via Recursive Skill-Augmented Reinforcement Learning

Large Language Model (LLM) agents have shown stunning results in complex tasks, yet they often operate in isolation, failing to learn from past experiences. Existing memory-based methods primarily store raw trajectories, which are often redundant and noise-heavy. This prevents agents from extracting high-level, reusable behavioral patterns that are essential for generalization. In this paper, we propose SkillRL, a framework that bridges the gap between raw experience and policy improvement through automatic skill discovery and recursive evolution. Our approach introduces an experience-based distillation mechanism to build a hierarchical skill library SkillBank, an adaptive retrieval strategy for general and task-specific heuristics, and a recursive evolution mechanism that allows the skill library to co-evolve with the agent's policy during reinforcement learning. These innovations significantly reduce the token footprint while enhancing reasoning utility. Experimental results on ALFWorld, WebShop and seven search-augmented tasks demonstrate that SkillRL achieves state-of-the-art performance, outperforming strong baselines over 15.3% and maintaining robustness as task complexity increases. Code is available at this https://github.com/aiming-lab/SkillRL.

cs.LG↗

arXiv2021

Leveraging Large-scale Computational Database and Deep Learning for Accurate Prediction of Material Properties

Accurately predicting the physical and chemical properties of materials remains one of the most challenging tasks in material design, and one effective strategy is to construct a reliable data set and use it for training a machine learning model. In this study, we constructed a large-scale material genome database (Matgen) containing 76,463 materials collected from experimentally-observed database, and computed their bandgap properties through the Density functional theory (DFT) method with Perdew-Burke-Ernzehof (PBE) functional. We verified the computation method by comparing part of our results with those from the open Material Project (MP) and Open Quantum Materials Database (OQMD), all with PBE computations, and found that Matgen achieved the same computation accuracy based on both measured and computed bandgap properties. Based on the computed properties of our comprehensive dataset, we have developed a new graph-based deep learning model, namely CrystalNet, through our recently developed Communicative Message Passing Neural Network (CMPNN) framework. The model was shown to outperform other state-of-the-art prediction models. A further fine-tuning on 1716 experimental bandgap values (CrystalNet-TL) achieved a superior performance with mean absolute error (MAE) of 0.77 eV on independent test, which has outperformed the pure PBE (1.14~1.45 eV). Moreover, the model was proven applicable to hypothetical materials with MAE of 0.77 eV as referred by computations from HSE, a highly accurate quantum mechanics (QM) method, consist better than PBE (MAE=1.13eV). We also made material structures, computed properties by PBE, and the CrystalNet models publically available at https://matgen.nscc-gz.cn.

cond-mat.mtrl-sci↗

arXiv2021

Learning Attributed Graph Representations with Communicative Message Passing Transformer

Constructing appropriate representations of molecules lies at the core of numerous tasks such as material science, chemistry and drug designs. Recent researches abstract molecules as attributed graphs and employ graph neural networks (GNN) for molecular representation learning, which have made remarkable achievements in molecular graph modeling. Albeit powerful, current models either are based on local aggregation operations and thus miss higher-order graph properties or focus on only node information without fully using the edge information. For this sake, we propose a Communicative Message Passing Transformer (CoMPT) neural network to improve the molecular graph representation by reinforcing message interactions between nodes and edges based on the Transformer architecture. Unlike the previous transformer-style GNNs that treat molecules as fully connected graphs, we introduce a message diffusion mechanism to leverage the graph connectivity inductive bias and reduce the message enrichment explosion. Extensive experiments demonstrated that the proposed model obtained superior performances (around 4$\%$ on average) against state-of-the-art baselines on seven chemical property datasets (graph-level tasks) and two chemical shift datasets (node-level tasks). Further visualization studies also indicated a better representation capacity achieved by our model.

cs.LG↗

arXiv2021

Lossless Point Cloud Attribute Compression with Normal-based Intra Prediction

The sparse LiDAR point clouds become more and more popular in various applications, e.g., the autonomous driving. However, for this type of data, there exists much under-explored space in the corresponding compression framework proposed by MPEG, i.e., geometry-based point cloud compression (G-PCC). In G-PCC, only the distance-based similarity is considered in the intra prediction for the attribute compression. In this paper, we propose a normal-based intra prediction scheme, which provides a more efficient lossless attribute compression by introducing the normals of point clouds. The angle between normals is used to further explore accurate local similarity, which optimizes the selection of predictors. We implement our method into the G-PCC reference software. Experimental results over LiDAR acquired datasets demonstrate that our proposed method is able to deliver better compression performance than the G-PCC anchor, with $2.1\%$ gains on average for lossless attribute coding.

eess.IV↗

arXiv2021

RAI-Net: Range-Adaptive LiDAR Point Cloud Frame Interpolation Network

LiDAR point cloud frame interpolation, which synthesizes the intermediate frame between the captured frames, has emerged as an important issue for many applications. Especially for reducing the amounts of point cloud transmission, it is by predicting the intermediate frame based on the reference frames to upsample data to high frame rate ones. However, due to high-dimensional and sparse characteristics of point clouds, it is more difficult to predict the intermediate frame for LiDAR point clouds than videos. In this paper, we propose a novel LiDAR point cloud frame interpolation method, which exploits range images (RIs) as an intermediate representation with CNNs to conduct the frame interpolation process. Considering the inherited characteristics of RIs differ from that of color images, we introduce spatially adaptive convolutions to extract range features adaptively, while a high-efficient flow estimation method is presented to generate optical flows. The proposed model then warps the input frames and range features, based on the optical flows to synthesize the interpolated frame. Extensive experiments on the KITTI dataset have clearly demonstrated that our method consistently achieves superior frame interpolation results with better perceptual quality to that of using state-of-the-art video frame interpolation methods. The proposed method could be integrated into any LiDAR point cloud compression systems for inter prediction.

eess.IV↗

arXiv2021

An Unsupervised Optical Flow Estimation For LiDAR Image Sequences

In recent years, the LiDAR images, as a 2D compact representation of 3D LiDAR point clouds, are widely applied in various tasks, e.g., 3D semantic segmentation, LiDAR point cloud compression (PCC). Among these works, the optical flow estimation for LiDAR image sequences has become a key issue, especially for the motion estimation of the inter prediction in PCC. However, the existing optical flow estimation models are likely to be unreliable for LiDAR images. In this work, we first propose a light-weight flow estimation model for LiDAR image sequences. The key novelty of our method lies in two aspects. One is that for the different characteristics (with the spatial-variation feature distribution) of the LiDAR images w.r.t. the normal color images, we introduce the attention mechanism into our model to improve the quality of the estimated flow. The other one is that to tackle the lack of large-scale LiDAR-image annotations, we present an unsupervised method, which directly minimizes the inconsistency between the reference image and the reconstructed image based on the estimated optical flow. Extensive experimental results have shown that our proposed model outperforms other mainstream models on the KITTI dataset, with much fewer parameters.

eess.IV↗

arXiv2018

An AI Based Super Nodes Selection Algorithm in BlockChain Networks

In blockchain systems, especially cryptographic currencies such as Bitcoin, the double-spending and Byzantine-general-like problem are solved by reaching consensus protocols among all nodes. The state-of-the-art protocols include Proof-of-Work, Proof-of-Stake and Delegated-Proof-of-Stake. Proof-of-Work urges nodes to prove their computing power measured in hash rate in a crypto-puzzle solving competition. The other two take into account the amount of stake of each nodes and even design a vote in Delegated-Proof-of-Stake. However, these frameworks have several drawbacks, such as consuming a large number of electricity, leading the whole blockchain to a centralized system and so on. In this paper, we propose the conceptual framework, fundamental theory and research methodology, based on artificial intelligence technology that exploits nearly complementary information of each nodes. And we designed a particular convolutional neural network and a dynamic threshold, which obtained the super nodes and the random nodes, to reach the consensus. Experimental results demonstrate that our framework combines the advantages of Proof-of-Work, Proof-of-Stake and Delegated-Proof-of-Stake by avoiding complicated hash operation and monopoly. Furthermore, it compares favorably to the three state-of-the-art consensus frameworks, in terms of security and the speed of transaction confirmation.

cs.CR↗

arXiv2012

A New Weighted Spearman's Footrule as A Measure of Distance between Rankings

Many applications motivate the distance measure between rankings, such as comparing top-k lists and rank aggregation for voting, and intrigue great interest to researchers. For example, for a search engine, the use of different ranking algorithms may return different ranking lists. The effect of a ranking algorithm can be estimated by computing the distance (similarity) between the result ranking it returns and the appropriate ranking people expect. People may be interested in only the first few items of result ranking, therefore the metric for measuring the distance should emphasize on the items in higher positions. Besides, in an extreme case, if a result ranking is the total reverse of the expected ranking, then it is considered to be the worst ranking with the maximum distance. Therefore, a metric is called for, which can satisfy both of the two intuitions. To address this problem, we present a weighted metric based on the classical Spearman's footrule metric to measure the distance between two permutations of n objects. This metric can be applied in rank aggregation problem with a polynomial time algorithm, and produces a 2-approximation for adopting the weighted Kendall's tau distance proposed by Farnoud et al.

cs.DM↗