Search arXiv⌕ Search

arXiv subjects

Wei Ji

Publications and source records attributed to Wei Ji.

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

Making Image Editing Easier via Adaptive Task Reformulation with Agentic Executions

Instruction guided image editing has advanced substantially with recent generative models, yet it still fails to produce reliable results across many seemingly simple cases. We observe that a large portion of these failures stem not from insufficient model capacity, but from poorly formulated editing tasks, such as those involving small targets, implicit spatial relations, or under-specified instructions. In this work, we frame image editing failures as a task formulation problem and propose an adaptive task reformulation framework that improves editing performance without modifying the underlying model. Our key idea is to transform the original image-instruction pair into a sequence of operations that are dynamically determined and executed by a MLLM agent through analysis, routing, reformulation, and feedback-driven refinement. Experiments on multiple benchmarks, including ImgEdit, PICA, and RePlan, across diverse editing backbones such as Qwen Image Edit and Nano Banana, show consistent improvements, with especially large gains on challenging cases. These results suggest that task reformulation is a critical but underexplored factor, and that substantial gains can be achieved by better matching editing tasks to the effective operating regime of existing models.

cs.CV↗
arXiv2026

AdaSurvMamba: Dynamic Fusion and Semantic Scanning for Multimodal Survival Analysis

Multimodal survival analysis utilizing whole slide images (WSIs) and genomic profiles is fundamental for cancer prognosis. Recently, state-space models like Mamba have emerged as powerful tools for sequence modeling. However, translating this success to complex multimodal tasks is hindered by two critical limitations. First, conventional fusion strategies assume a static multimodal interaction strength, ignoring the fluctuating diagnostic importance of each modality across different patients and local regions. Second, the standard Mamba architecture processes tokens along predefined physical paths. This rigid scanning disrupts the semantic continuity of spatially scattered medical features and exacerbates long-range decay. To address these challenges, we introduce AdaSurvMamba as a novel adaptive framework for multimodal survival analysis. The framework features a Dual-Scale Importance-Aware Reconstruction (DSIR) module to dynamically modulate cross-modal interaction strength. It evaluates diagnostic importance at both the sequence and token levels to reconstruct the input representations. Furthermore, we propose a Semantic Aggregation Scanning (SAS) module to overcome contextual fragmentation. The SAS module dynamically reorganizes discrete tokens into semantically continuous sequences via a shared prototype pool. It explicitly modulates the state transition step size using global modality context and semantic priors to adaptively control the information absorption rate. Experiments across five TCGA cohorts demonstrate consistent gains over existing methods. Code is available at https://github.com/zjlGO/AdaSurvMamba.

cs.LG↗
arXiv2026

Ferromagnetic broadband sensing of axionlike dark matter

Levitated particles have demonstrated ultrahigh sensitivity to magnetic fields and accelerations owing to their extremely low dissipation. Such systems have strong potential for fundamental physics research, particularly for the detection of axions and axionlike particles, well-motivated dark matter candidates spanning a broad mass range. In this context, both high sensitivity and large bandwidth are essential. Here, we demonstrate a levitated magnet magnetometer based on an engineered double-resonance mode, achieving an effective linewidth at its optimal sensitivity that is approximately three orders of magnitude broader than those of previous approaches. Together with a hard-magnet array that enhances the axion-induced signal and soft-ferromagnetic shielding that suppresses environmental magnetic noise, this system constitutes a hybrid ferromagnetic platform for axionlike dark matter searches. We search for axionlike dark matter through its photon coupling $g_{aγ}$ over the $40$-$3000\,\mathrm{Hz}$ frequency range and establish new direct limits in this frequency band. The best sensitivity is achieved near the upper resonance around $276\,\mathrm{Hz}$, where the magnetometer reaches a magnetic-field resolution of $0.7\,\mathrm{fT}$, corresponding to a limit of $g_{aγ}\sim10^{-7}\,\mathrm{GeV}^{-1}$. At this frequency, this result improves upon previous direct limits by more than four orders of magnitude. The demonstrated high-bandwidth levitated sensor may also enable a broad range of applications, including biological sensing and precision measurements.

hep-ex↗
arXiv2026

InvDesMobility: a reliability-gated first-principles feedback framework for closed-loop materials discovery

Inverse materials design starts from target functionality and searches for structures that can realize it. Its value in closed-loop discovery depends not only on prediction performance, but also on whether expensive first-principles results are independently validated, provenance-recorded, and admitted as feedback only when evidence is sufficient. This is especially important for composite properties such as carrier mobility, where a final scalar value hides intermediate quantities, fit quality, convergence history, and workflow assumptions. Here we present InvDesMobility, a reliability-gated first-principles feedback framework that integrates multi-agent automated DFT, evidence stratification, generative structure proposal, acquisition ranking, and auditable release. Using 516 2DMatPedia-derived candidates, the workflow produced 280 QC-passed materials and 573 retained carrier-direction seed channels after channel-level reliability gating. These records were split into two feedback objects: relaxed structures updated the generative model, while retained mobility channels trained the acquisition model and set validation priority. Over multiple iterations, InvDesMobility screened 2.4 x 10^6 structures, submitted 102 candidates for DFT validation, and retained 86 reliability-gated generated channels across 41 formulas. Overall, the main contribution is not a fixed list of high-mobility materials, but a transferable feedback contract that makes closed-loop inverse design both useful and auditable when learning from expensive calculated properties. All source data, retained feedback records, and workflows are available at https://github.com/DreamLufei/invDesMobility, with an accompanying evidence website at https://dreamlufei.github.io/invDesMobility/.

cond-mat.mtrl-sci↗
arXiv2026

Curvature-driven revival of charge density waves in non-Euclidean space

Strongly correlated quantum states, such as charge density waves (CDWs), are exquisitely sensitive to Fermi surface topology and lattice symmetry, and are typically quenched by heavy carrier doping. In two-dimensional (2D) systems, however, macroscopic geometric curvature emerges as a novel structural degree of freedom to modulate microscopic quantum coherence. This raises a compelling physical question: can non-Euclidean geometric deformations compete with extreme electronic perturbations to reshape, or even revive, a quenched macroscopic quantum order? Here, by constructing monolayer TiSe$_2$-NbSe$_2$ heterostructure on a BLG/SiC substrate for the first time, we report the curvature-driven revival of a frustrated charge order in a non-Euclidean space. Low-temperature angle-resolved photoemission spectroscopy (ARPES) reveals a massive interfacial charge transfer, which destroys the global Fermi surface nesting and completely suppresses the long-range CDW order in Euclidean flat regions. Strikingly, high-resolution scanning tunneling microscopy (STM) reveals that a novel, non-linear CDW state miraculously survives, remaining strictly localized within morphologically distorted, non-Euclidean nanoscale curved regions. Atomistic simulations unravel the structural origin of this phenomenon, demonstrating that interfacial twist and lattice mismatch spontaneously generate a corrugated superlattice.

cond-mat.mtrl-sci↗
arXiv2026

RADAR: Redundancy-Aware Diffusion for Multi-Agent Communication Structure Generation

Compared with individual agents, large language model based multi-agent systems have shown great capabilities consistently across diverse tasks, including code generation, mathematical reasoning, and planning, etc. Despite their impressive performance, the effectiveness and robustness of these systems heavily rely on their communication topology, which is often fixed or generated in a single step. This restricts fine-grained structural exploration and flexible composition, resulting in excessive token utilization on simple tasks while limiting capability on complicated tasks. To mitigate this challenge, we introduce RADAR, a redundancy-aware and query-adaptive generative framework that actively reduce communication overhead. Motivated by recent progress in conditional discrete graph diffusion models, we formulate communication topology design as a step-by-step generation process, guided by the effective size of the graph. Comprehensive experiments on six benchmarks demonstrate that RADAR consistently outperforms recent baselines, achieving higher accuracy, lower token consumption, and greater robustness across diverse scenarios. Our code and data are available at https://github.com/cszhangzhen/RADAR.

cs.AI↗
arXiv2026

Turing Patterns for Multimedia: Reaction-Diffusion Multi-Modal Fusion for Language-Guided Video Moment Retrieval

Video-language models are pivotal for tasks such as moment retrieval and highlight detection, yet they often struggle to capture the dynamic, non-linear interactions between temporal video sequences and textual semantics. Existing approaches, relying on static cross-attention or prompt-tuning mechanisms, fail to adaptively model the evolving relationships between modalities, leading to suboptimal alignment and limited generalization. Inspired by systems biology, we propose \textbf{Reaction-Diffusion Multimodal Fusion (RDMF)}, a novel framework that reimagines video-language alignment as a reaction-diffusion (RD) process, drawing on the principles of pattern formation introduced by Alan Turing. In RDMF, video features diffuse across time to capture temporal context, while text-video interactions are modeled as non-linear reactions that amplify relevant features and suppress noise, forming emergent patterns akin to biological systems. Leveraging the Gray-Scott RD model, we design a computationally efficient fusion module that integrates video and text representations, supported by rigorous mathematical analysis of stability and convergence using Turing instability criteria. Our framework is theoretically grounded, employing advanced mathematical tools to ensure stable pattern formation, and is practically viable, incorporating standard components like pretrained encoders and DETR-style heads for moment retrieval and saliency prediction. RDMF represents a pioneering interdisciplinary approach, bridging systems biology and multimedia research to address the limitations of conventional multimodal fusion. Preliminary experiments demonstrate its potential to outperform existing methods in identifying salient video moments, offering a new paradigm for video-language tasks.

cs.CV↗
arXiv2026

Explainable Forensics of Manipulated Segments in Untrimmed Long Videos

The rapid advancement of AI-driven video generation has transformed content creation, while simultaneously increasing the risk of misinformation through localized manipulations in long-form videos. Existing video forensic methods predominantly operate on short, independent clips, and thus fail to capture realistic scenarios where AI-generated content is sparsely embedded within otherwise authentic footage. To bridge this gap, we formulate the task of Temporal AI-Generated Segment Localization and Explanation, which targets authenticity detection, temporal localization, and interpretable analysis of manipulated segments in untrimmed long videos. We further introduce TASLE, a large-scale benchmark comprising 12,472 untrimmed videos with diverse manipulation patterns and rich annotation signals, including temporal boundaries, authenticity labels, and segment-level rationales. In addition, we propose MSLoc, a coarse-to-fine forensic baseline that combines a boundary-sensitive proposal generation module for efficient long-video scanning with an MLLM-based refinement module for precise boundary localization and interpretable reasoning. Experiments validate the effectiveness of the proposed baseline, highlighting the importance of segment-level explainable forensics for long-form AI-generated video analysis. Our dataset and code are publicly available at https://debby-0527.github.io/TASLE.

cs.CV↗
arXiv2026

A tunable feedback-controlled magnetic trap for a magnet in free fall

Ferromagnets in free space are predicted to exhibit pure Larmor precession at near-zero magnetic fields and provide exceptional sensitivity for magnetometry and gyroscopy. Notably, pure Larmor precession has not been observed in a macroscopic ferromagnetic particle, despite its fundamental importance and potential for probing relativistic effects and dark-matter interactions. Realizing such dynamics requires true free fall to eliminate clamping losses and trap-induced systematics. A central challenge is designing a tunable trap that is weak enough to permit near-free evolution yet robust enough to withstand the disturbances of launch and release. Here, we propose and demonstrate a novel master proportional-integral-differential magnetic trap (MPIDMT) combining a PID-controlled coil system with a master control coil system. Implemented in the third-generation drop tower - Einstein-Elevator, during the microgravity phase the system stably levitates a ferromagnetic particle against shock accelerations up to 1.5 g and resolves its motion in both a low-field (0.4 g) configuration and in pure free fall. These results represent a key step toward free-fall ferromagnetic magnetometry, the long-sought direct observation of macroscopic Larmor precession, and future space-based experiments.

physics.atom-ph↗
arXiv2026

GIRL-DETR: Gradient-Isolated Reinforcement Learning for Video Moment Retrieval

Video Moment Retrieval (VMR) task requires accurately localizing temporal boundaries aligned with natural language queries, but many models suffer from a misalignment between continuous surrogate losses and non-differentiable metrics, leading to optimization stagnation during the late stages of training and trapping boundary predictions in suboptimal solutions. Although Reinforcement Learning (RL) post-training successfully optimizes localization results for large models, applying it directly to lightweight networks easily disrupts the fragile feature representations established during the supervised phase. To overcome this optimization bottleneck, we propose Gradient-Isolated Reinforcement Learning for DETR (GIRL-DETR), introducing RL post-training into a lightweight temporal localization framework for the first time. The input video and text features first establish early alignment through Cross-Modal Interaction (CMI) before entering the transformer encoder. Subsequently, a Text-Guided Gating (TGG) mechanism dynamically injects semantic priors into the queries before the transformer decoder generates candidate proposals, providing high signal-to-noise ratio inputs for temporal prediction. After the supervised training reaches convergence, the backbone network is frozen to protect the feature manifold, while the detection head directly optimizes the non-differentiable evaluation metric tIoU to enhance localization accuracy through a Three-stage Progressive Reinforcement Learning (TPRL) strategy. This approach achieves an orthogonal decoupling of state representation and metric optimization. Experiments on Charades-STA, QVHighlights, and TACoS demonstrate that GIRL-DETR effectively resolves surrogate loss degradation and achieves substantial accuracy improvements with minimal parameter updates, providing a robust new pathway for RL applications in lightweight VMR models.

cs.CV↗
arXiv2026

Immuno-VLM: Immunizing Large Vision-Language Models via Generative Semantic Antibodies for Open-World Trustworthiness

Large Vision-Language Models have achieved unprecedented success in zero-shot recognition by aligning visual features with broad semantic concepts. However, this semantic abstraction creates a critical vulnerability in open-world deployment: the ``Hubris of Semantics'', where models force-fit unknown anomalies into known categories with high confidence due to the lack of explicit negative knowledge. To address this \textit{Open-World Trustworthiness Paradox}, we propose \textbf{Immuno-VLM}, a bio-inspired framework that adapts the biological principle of \textbf{Immunological Negative Selection} to high-dimensional latent spaces. Departing from traditional Open-Set Recognition methods that rely on passive density estimation or inefficient pixel-space outlier generation, Immuno-VLM leverages the generative reasoning of Large Language Models to actively hallucinate ``Semantic Antibodies'', textual descriptions of near-distribution outliers (e.g., look-alikes, contextual anomalies) that effectively bound the decision space of known classes.Extensive experiments on ImageNet-1K and four challenging OOD benchmarks reveal that Immuno-VLM establishes a new state-of-the-art.

cs.CV↗
arXiv2026

Towards Unified Vision-Language Models with Incomplete Multi-Modal Inputs

Video-Language Models (VLMs) have demonstrated impressive multi-modal reasoning capabilities across diverse computer vision applications. However, these VLMs are task-specific and assume that both video and language inputs are complete. However, real-world VLM applications might face challenges due to deactivated sensors (e.g., cameras are unavailable due to data privacy), yielding modality-incomplete data and leading to inconsistency between training and testing data. While straightforward incomplete input can boast training generalization-ability and lead to training failure, its potential risks to VLMs regarding safety and trustworthiness have been largely neglected. To this end, we make the first attempt to propose a unified incomplete video-language model to process the incomplete multi-modal inputs. Extensive experimental results show that our method can serve as a plug-and-play module for previous works to improve their performance in various multi-modal tasks.

cs.CV↗
arXiv2026

ConceptSeg-R1: Segment Any Concept via Meta-Reinforcement Learning

Recent progress in promptable segmentation has shifted visual perception from object-level localization toward concept-level understanding. However, the notion of a concept remains under-specified, making it unclear whether current methods truly generalize beyond category recognition. In this work, we formalize generalized concept segmentation through a three-level taxonomy consisting of context-independent (CI), context-dependent (CD), and context-reasoning (CR) concepts, which reveals a clear capability gap across increasing levels of cognitive complexity. To address this challenge, we propose ConceptSeg-R1, a unified framework that reformulates concept segmentation as rule-induced concept grounding. At the core of our method is Meta-GRPO, a meta-reinforcement learning mechanism that learns transferable task rules from visual demonstrations and verifies them through proxy reasoning. The inferred reasoning states are then translated into segmentation-ready concept prompts via a lightweight concept translation module, enabling deductive application to target images. A shortcut routing strategy further preserves the native efficiency of segmentation models on simple cases. To systematically evaluate generalized concept segmentation, we conduct extensive experiments across diverse CI, CD, and CR concept segmentation benchmarks spanning natural, industrial, medical and reasoning-intensive domains. Without bells and whistles, ConceptSeg-R1 achieves strong performance across the full concept hierarchy while maintaining the native capability of promptable segmentation backbones. As an initial step toward segmenting any concept, we hope ConceptSeg-R1 can serve as a practical baseline for advancing segmentation from object-level prediction toward concept-level understanding.

cs.CV↗
arXiv2026

UniPPTBench: A Unified Benchmark for Presentation Generation Across Diverse Input Settings

Existing works typically focus on presentation generation under isolated input settings, whereas real-world use cases span diverse scenarios, including vague user prompts, long documents, multimodal materials, and multiple heterogeneous sources. Moreover, current evaluations are often insufficiently scenario-specific. They mainly rely on generic presentation-quality criteria, such as visual appeal, layout quality, and overall coherence, but fail to assess the core capabilities required by different input settings, including grounded compression, visual-text alignment, and cross-source synthesis. Consequently, the field lacks a unified benchmark and a scenario-aware evaluation framework for faithfully diagnosing presentation-generation systems across diverse real-world settings. We present UniPPTBench, a unified benchmark for presentation generation across four representative input settings: vague-prompt, long-document, multimodal-document, and multi-source generation. We further introduce UniPPTEval, a scenario-aware evaluation protocol that combines shared metrics for cross-setting comparison with scenario-specific metrics tailored to the core requirements of each setting. We also provide transparent reference baselines to support reproducible comparison. Experiments on UniPPTBench reveal substantial performance variation across settings and recurring failure modes in content grounding, multimodal integration, and cross-source synthesis. In particular, strong performance on generic presentation-quality metrics does not necessarily imply strong task fulfillment in grounded scenarios. Together, UniPPTBench and UniPPTEval provide a faithful and diagnostic foundation for evaluating presentation generation across diverse real-world scenarios. Code and data will be publicly available.

cs.CV↗
arXiv2026

An Adaptive Real-Time Forecasting Framework for Cryogenic Fluid Management in Space Systems

Accurate real-time forecasting of cryogenic tank behavior is essential for the safe and efficient operation of propulsion and storage systems in future deep-space missions. While cryogenic fluid management (CFM) systems increasingly require autonomous capabilities, conventional simulation methods remain hindered by high computational cost, model imperfections, and sensitivity to unanticipated boundary condition changes. To address these limitations, this study proposes an Adaptive Real-Time Forecasting Framework for Cryogenic Propellant Management in Space Systems, featuring a lightweight, non-intrusive method named ARCTIC (Adaptive Real-time Cryogenic Tank Inference and Correction). ARCTIC integrates real-time sensor data with precomputed nodal simulations through a data-driven correction layer that dynamically refines forecast accuracy without modifying the underlying model. Two updating mechanisms, auto-calibration and observation and correction, enable continuous adaptation to evolving system states and transient disturbances. The method is first assessed through synthetic scenarios representing self-pressurization, sloshing, and periodic operations, then validated using experimental data from NASA's Multipurpose Hydrogen Test Bed and K-Site facilities. Results demonstrate that ARCTIC significantly improves forecast accuracy under model imperfections, data noise, and boundary fluctuations, offering a robust real-time forecasting capability to support autonomous CFM operations. The framework's compatibility with existing simulation tools and its low computational overhead make it especially suited for onboard implementation in space systems requiring predictive autonomy.

physics.flu-dyn↗
arXiv2026

MD-PNOP: Equation-Recast Neural Operators for Minimal-Data Extrapolation and PDE Solver Acceleration

The computational overhead of traditional numerical solvers for partial differential equations (PDEs) remains a critical bottleneck for large-scale parametric studies and design optimization. We introduce a Minimal-Data Parametric Neural Operator Preconditioning (MD-PNOP) framework, which establishes a new strategy for accelerating parametric PDE solvers while strictly preserving physical constraints. To address the extrapolation limitation of neural operators, parameter-induced operator difference is recast as additional source terms and incorporated into an iterative solution scheme using a pretrained neural operator. This equation-recast formulation enables systematic parameter extrapolation from a single training configuration to a broad range of unseen parameter settings without retraining. The neural operator predictions are then embedded into iterative PDE solvers as improved initial guesses, thereby reducing convergence iterations without sacrificing accuracy. Unlike purely data-driven approaches, MD-PNOP guarantees that the governing equations remain fully enforced, eliminating concerns regarding loss of physics or interpretability. The framework is architecture-agnostic and is demonstrated using both DeepONet and FNO for Boltzmann transport equation solvers in neutron transport applications. Numerical results demonstrate that neural operators trained on a single set of constant parameters successfully accelerate solutions with heterogeneous, sinusoidal, and discontinuous parameter distributions. Moreover, MD-PNOP consistently achieves approximately 50% reduction in computational time while maintaining full-order fidelity for fixed-source, single-group eigenvalue, and multigroup coupled eigenvalue problems.

cs.LG↗
arXiv2026

Towards Unified Surgical Scene Understanding:Bridging Reasoning and Grounding via MLLMs

Surgical scene understanding is a cornerstone of computer-assisted intervention. While recent advances, particularly in surgical image segmentation, have driven progress, real-world clinical applications require a more holistic understanding that jointly captures procedural context, semantic reasoning, and precise visual grounding. However, existing approaches typically address these components in isolation, leading to fragmented representations and limited semantic consistency. To address this limitation, we propose SurgMLLM, a unified surgical scene understanding framework that bridges high-level reasoning and low-level visual grounding within a single model. Given surgical videos, SurgMLLM fine-tunes a multimodal large language model (MLLM) to support structured interpretability reasoning, which is used to jointly model phases, instrument-verb-target (IVT) triplets, and triplet-entity segmentation tokens. These tokens are then temporally aggregated and serve as prompts for a segmentation network, enabling accurate pixel-wise grounding of triplet instruments and targets. The entire framework is trained end-to-end with a unified objective that couples language-based reasoning supervision with visual grounding losses, promoting coherent cross-task learning and clinically consistent scene representations. To facilitate unified evaluation, we introduce CholecT45-Scene, extending CholecT45 dataset with 64,299 frames of pixel-level mask annotations for instruments and targets, aligned with existing triplet labels. Extensive experiments show that SurgMLLM significantly advances surgical scene understanding, improving the primary triplet recognition metric AP_IVT from 40.7% to 46.0% and consistently outperforming prior methods in phase recognition and segmentation. These results highlight the effectiveness of unified reasoning-and-grounding for reliable, context-aware surgical assistance.

cs.CV↗
arXiv2026

Spatially anisotropic Kondo resonance coupled with the superconducting gap in a kagome metal

The chromium-based kagome metal CsCr3Sb5 has garnered significant interest due to its strong electron correlations, intertwined orders and potential for unconventional superconductivity under high pressure. The evolution of magnetic and superconducting interactions as the more frequently studied CsV3Sb5 is doped to CsCr3Sb5 remains poorly understood. Here, we demonstrate the emergence of a spatially anisotropic Kondo resonance intertwined with the superconducting gap, enabled by introducing magnetic Cr impurities into the kagome superconductor CsV3Sb5. The addition of dilute Cr impurities not only weakens long range charge density wave order but also produces local magnetic moments that leads to Kondo resonances. We show that the Kondo resonance forms anisotropic, ripple like spatial patterns around individual Cr atoms, breaking all local mirror symmetries. We further reveal that with the emergence of Kondo screening, the coherence peak and depth of superconducting gap with finite zero-energy conductance are enhanced. This suggests that non superconducting carriers at the Fermi surface in the parent compound participate in the Kondo effect, simultaneously screening Cr magnetic moments and increasing the superfluid density. Our findings offer an opportunity to study the interplay between superconductivity and local magnetism in kagome materials.

cond-mat.supr-con↗