arXiv subjects

Yijie Zhu

Publications and source records attributed to Yijie Zhu.

At least 19 recordsLinked to original sources

arXiv2026

Learning to Use Imagination: Progress-Conditioned Future Utilization for World Action Models

World Action Models (WAMs) extend Vision-Language-Action (VLA) models by incorporating future visual dynamics into action generation. However, existing WAMs often utilize imagined futures with limited adaptation to evolving execution progress, potentially introducing distracting or unreliable predictive cues. This limitation arises from two empirically identified forms of non-uniformity in future utility: (i) at the inter-progress level, the utility of imagined futures varies across execution stages as control demands change; and (ii) at the intra-progress level, individual future latents exhibit heterogeneous relevance within the same progress state. To address these limitations, we propose ProWAM, a Progress-Conditioned World Action Model that introduces execution progress as an explicit intermediate representation for adaptive imagination utilization. ProWAM comprises two tightly coupled components: (1) To obtain a reliable representation of execution progress, we propose the Self-Supervised Dual-Temporal Progress Encoder (SS-DTPE). SS-DTPE couples short-term action-observation interaction modeling with long-term recurrent progress aggregation to capture recent execution feedback and accumulated task history. (2) Conditioned on the progress representation from SS-DTPE, we propose the Hierarchical Progress-Conditioned Imagination Modulation (HPIM) to adapt imagination utilization to execution progress. HPIM operates at two complementary levels: an inter-progress global modulation mechanism adapts future utilization across execution stages, while an intra-progress relevance mechanism differentiates individual future latents within each progress state. Extensive experiments demonstrate consistent gains over strong VLA and WAM baselines.

cs.CV↗

arXiv2026

Computationally Efficient Pathology Segmentation using Knowledge Distillation from Foundation Models

Automatic tissue segmentation is essential for large-scale analysis of histopathology whole-slide images (WSIs), but accurate pixel-level segmentation remains challenging. Pixel-level annotations are expensive to obtain, models pre-trained on natural images may transfer poorly to histopathology, and pathology foundation models, despite their strong representations, are computationally expensive to deploy at scale. We address these challenges with a foundation-model knowledge distillation framework for efficient tissue segmentation. We first train Virchow2-based segmentation teachers, including a LoRA-adapted variant, achieving state-of-the-art or highly competitive performance across four datasets: PUMA, IGNITE, BEETLE, and a private blood vessel segmentation dataset. We then transfer response-level and feature-level knowledge from these teachers into compact student networks. Distillation consistently improves student performance over supervised training alone, producing state-of-the-art or near state-of-the-art results with substantially fewer parameters and up to ten-fold higher inference throughput than foundation-model-based segmentation networks. These results show that foundation-model knowledge can be effectively transferred to efficient segmentation models for scalable deployment without compromising performance. Models trained on the public datasets will be released through TIAToolbox.

eess.IV↗

arXiv2026

Cartesian tensor equivariant machine-learning force field for spin-dependent atomistic simulations

Magnetic materials exhibit an intricate coupling between atomic structure and spin degrees of freedom, posing a fundamental challenge for atomistic simulations across experimentally relevant length and time scales. Here we introduce HotPP-Spin, a spin-dependent extension of HotPP for magnetic machine learning interatomic potentials, built on Cartesian tensor equivariant message passing. Atomic magnetic moments are treated as explicit axial-vector degrees of freedom, while spatial-inversion and time-reversal parities are propagated through the tensor couplings. This construction provides a unified representation of exchange-dominated and spin-orbit-induced interactions without imposing predefined analytical interaction forms. A scalar spin-dependent potential energy surface yields energy-conserving atomic forces and magnetic effective fields through differentiation. Benchmarks spanning collinear magnetism, noncollinear magnetism, and spin-orbit-coupling-induced magnetic anisotropy show that HotPP-Spin accurately describes magnetic energy landscapes, magnetic forces, and magnetic-order-dependent energy-volume relations within the same general framework. For H-phase monolayer VSe$_2$, stochastic spin-dynamics simulations using the learned magnetic effective fields locate the finite-size magnetic ordering crossover at 415--435~K, in close numerical agreement with the reported experimental value of $418.5\pm7.8$~K. These results establish Cartesian tensor message passing as a general route for connecting first-principles magnetic energetics with large-scale atomistic simulations of coupled structural and spin phenomena.

physics.comp-ph↗

arXiv2026

AffectAgent: Collaborative Multi-Agent Reasoning for Retrieval-Augmented Multimodal Emotion Recognition

LLM-based multimodal emotion recognition relies on static parametric memory and often hallucinates when interpreting nuanced affective states. In this paper, given that single-round retrieval-augmented generation is highly susceptible to modal ambiguity and therefore struggles to capture complex affective dependencies across modalities, we introduce AffectAgent, an affect-oriented multi-agent retrieval-augmented generation framework that leverages collaborative decision-making among agents for fine-grained affective understanding. Specifically, AffectAgent comprises three jointly optimized specialized agents, namely a query planner, an evidence filter, and an emotion generator, which collaboratively perform analytical reasoning to retrieve cross-modal samples, assess evidence, and generate predictions. These agents are optimized end-to-end using Multi-Agent Proximal Policy Optimization (MAPPO) with a shared affective reward to ensure consistent emotion understanding. Furthermore, we introduce Modality-Balancing Mixture of Experts (MB-MoE) and Retrieval-Augmented Adaptive Fusion (RAAF), where MB-MoE dynamically regulates the contributions of different modalities to mitigate representation mismatch caused by cross-modal heterogeneity, while RAAF enhances semantic completion under missing-modality conditions by incorporating retrieved audiovisual embeddings. Extensive experiments on MER-UniBench demonstrate that AffectAgent achieves superior performance across complex scenarios. Our code will be released at: https://github.com/Wz1h1NG/AffectAgent.

cs.CV↗

arXiv2026

MIBE: Multi-subject Interaction Benchmark and Evaluator for Personalized Image Generation

Multi-subject personalized image generation requires the precise rendering of all requested reference identities and their specified interactions based on a guiding prompt. However, state-of-the-art models still struggle with this process, frequently omitting subjects, failing to preserve reference appearances, or misattributing interactions. Furthermore, existing metrics designed primarily for single-subject fidelity cannot reliably capture these errors, suffering severe degradation in ranking separability and failing to align with human preference as the subject count increases. To address this gap, we introduce Multi-subject Interaction Benchmark and Evaluator (MIBE), a unified framework comprising a Multi-subject Interaction Benchmark (MIB) and a Multi-subject Interaction Evaluator (MIE). MIB systematically covers diverse relation types and scene complexities through a decoupled data regime. This consists of a 60K-pair VLM-labeled Silver Set for scalable metric training and a 4K-pair double-blind Human Evaluation Gold Set covering a diverse range of state-of-the-art generators, with the Silver Set reaching 95.1% cross-VLM preference agreement. To demonstrate the utility of this benchmark, we present MIE, a lightweight, reference-conditioned evaluator trained exclusively on the Silver Set with a dual-head ranking and diagnosis objective. MIE exhibits strong cross-generator generalization on the Gold Set, achieving 0.922 overall pairwise accuracy against human preference, including 0.982 on seen generators and 0.884 on unseen generators. By outperforming a broad spectrum of baseline metrics, including CLIP and DINO variants, MIE demonstrates that diagnostic supervision can preserve ranking separability and human alignment where traditional evaluators collapse.

cs.CV↗

arXiv2026

PUMA: Layer-Pruned Language Model for Efficient Unified Multimodal Retrieval with Modality-Adaptive Learning

As multimedia content expands, the demand for unified multimodal retrieval (UMR) in real-world applications increases. Recent work leverages multimodal large language models (MLLMs) to tackle this task. However, their large parameter size results in high training costs and low inference efficiency. To address this, we propose PUMA: a Layer-Pruned Language Model for Efficient Unified Multimodal Retrieval with Modality-Adaptive Learning. Our approach improves UMR from both structural and learning perspectives. (1) Structurally, we propose Layer-Pruned Self-Distillation, which prunes MLLMs by keeping only shallow layers while distilling features from dropped deep layers as teacher signals. This reduces parameters and preserves representation capability. (2) On the learning side, we introduce Modality-Adaptive Contrastive Learning Loss (MAC-Loss), which separates in-batch negatives into harder intra-modality and easier inter-modality groups based on the target modality, assigning different temperature strategies to enhance learning efficiency. Experiments show our method significantly reduces resource usage while maintaining strong performance.

cs.MM↗

arXiv2026

UniEmo: Unifying Emotional Understanding and Generation with Learnable Expert Queries

Emotional understanding and generation are often treated as separate tasks, yet they are inherently complementary and can mutually enhance each other. In this paper, we propose the UniEmo, a unified framework that seamlessly integrates these two tasks. The key challenge lies in the abstract nature of emotions, necessitating the extraction of visual representations beneficial for both tasks. To address this, we propose a hierarchical emotional understanding chain with learnable expert queries that progressively extracts multi-scale emotional features, thereby serving as a foundational step for unification. Simultaneously, we fuse these expert queries and emotional representations to guide the diffusion model in generating emotion-evoking images. To enhance the diversity and fidelity of the generated emotional images, we further introduce the emotional correlation coefficient and emotional condition loss into the fusion process. This step facilitates fusion and alignment for emotional generation guided by the understanding. In turn, we demonstrate that joint training allows the generation component to provide implicit feedback to the understanding part. Furthermore, we propose a novel data filtering algorithm to select high-quality and diverse emotional images generated by the well-trained model, which explicitly feedback into the understanding part. Together, these generation-driven dual feedback processes enhance the model's understanding capacity. Extensive experiments show that UniEmo significantly outperforms state-of-the-art methods in both emotional understanding and generation tasks. The code for the proposed method is available at https://github.com/JiuTian-VL/UniEmo.

cs.CV↗

arXiv2026

Navigating the Emotion Tree: Hierarchical Hyperbolic RAG for Multimodal Emotion Recognition

Multimodal emotion recognition aims to integrate text, audio, and video sources to understand human affective states. Although multimodal large language models excel at multimodal reasoning, they typically treat emotion categories as independent labels, ignoring the rich hierarchical taxonomy of human psychology. Moreover, lacking external contextual knowledge makes them highly susceptible to over-interpreting noisy cues, further complicating fine-grained emotion classification. To address these issues, we propose \textbf{HyperEmo-RAG}, a retrieval-augmented generation framework that leverages a structured emotional knowledge base. Our framework introduces two key innovations. 1) Hierarchical hyperbolic grounding. Recognizing the inherent hierarchical tree structure of emotion taxonomies, we jointly embed hierarchical emotion labels and multimodal samples into a continuous hyperbolic space (Poincaré ball) and design a hierarchical beam-search deliberation process that progressively retrieves samples from coarse to fine-grained levels. 2) Structured evidence injection. Based on the retrieved evidence, we construct an evidence graph and inject the structured knowledge as explicit cognitive context into the LLM through a Tree-Aware Attention mechanism and an EmotionGraphFormer, preserving the integrity of graph-structured information. Experiments on multiple datasets demonstrate that HyperEmo-RAG significantly outperforms existing methods.

cs.LG↗

arXiv2026

When Identities Collapse: A Stress-Test Benchmark for Multi-Subject Personalization

Subject-driven text-to-image diffusion models have achieved remarkable success in preserving single identities, yet their ability to compose multiple interacting subjects remains largely unexplored and highly challenging. Existing evaluation protocols typically rely on global CLIP metrics, which are insensitive to local identity collapse and fail to capture the severity of multi-subject entanglement. In this paper, we identify a pervasive "Illusion of Scalability" in current models: while they excel at synthesizing 2-4 subjects in simple layouts, they suffer from catastrophic identity collapse when scaled to 6-10 subjects or tasked with complex physical interactions. To systematically expose this failure mode, we construct a rigorous stress-test benchmark comprising 75 prompts distributed across varying subject counts and interaction difficulties (Neutral, Occlusion, Interaction). Furthermore, we demonstrate that standard CLIP-based metrics are fundamentally flawed for this task, as they often assign high scores to semantically correct but identity-collapsed images (e.g., generating generic clones). To address this, we introduce the Subject Collapse Rate (SCR), a novel evaluation metric grounded in DINOv2's structural priors, which strictly penalizes local attention leakage and homogenization. Our extensive evaluation of state-of-the-art models (MOSAIC, XVerse, PSR) reveals a precipitous drop in identity fidelity as scene complexity grows, with SCR approaching 100% at 10 subjects. We trace this collapse to the semantic shortcuts inherent in global attention routing, underscoring the urgent need for explicit physical disentanglement in future generative architectures.

cs.CV↗

arXiv2026

GPU-MetaD: Full-Life-Cycle GPU Accelerated Metadynamics with Machine Learning Potentials

Large-scale molecular dynamics simulations with high accuracy have been increasingly popular for their capability to bridge the gap between atomistic modeling and mesoscale phenomena. Both machine learning potentials and enhanced sampling approaches offer substantial improvements in high-accuracy simulation efficiency, which can be further boosted through GPU acceleration. However, an efficient framework combining these advances for extending simulations to large systems and long timescales remains elusive. In this work, we proposed a full-life-cycle GPU accelerated metadynamics simulations package GPU-MetaD. Benchmarking across molecular, interface, and bulk systems demonstrates that GPU-MetaD efficiently handles diverse atomic systems and delivers an order-of-magnitude performance improvement. Building on this demonstrated capability, it enables ab-initio-level rare-event sampling for systems comprising millions of atoms on a typical single GPU. This capability allows us to reveal a previously unknown size-dependent two-step nucleation mechanism in gallium nitride (GaN), highlighting the potential of GPU-MetaD for uncovering complex rare events in realistic large-scale materials systems.

physics.comp-ph↗

arXiv2026

$Δ$VLA: Prior-Guided Vision-Language-Action Models via World Knowledge Variation

Recent vision-language-action (VLA) models have significantly advanced robotic manipulation by unifying perception, reasoning, and control. To achieve such integration, recent studies adopt a predictive paradigm that models future visual states or world knowledge to guide action generation. However, these models emphasize forecasting outcomes rather than reasoning about the underlying process of change, which is essential for determining how to act. To address this, we propose $Δ$VLA, a prior-guided framework that models world-knowledge variations relative to an explicit current-world knowledge prior for action generation, rather than regressing absolute future world states. Specifically, 1) to construct the current world knowledge prior, we propose the Prior-Guided WorldKnowledge Extractor (PWKE). It extracts manipulable regions, spatial relations, and semantic cues from the visual input, guided by auxiliary heads and prior pseudo labels, thus reducing redundancy. 2) Building upon this, to represent how world knowledge evolves under actions, we introduce the Latent World Variation Quantization (LWVQ). It learns a discrete latent space via a VQ-VAE objective to encode world knowledge variations, shifting prediction from full modalities to compact latent. 3)Moreover, to mitigate interference during variation modeling, we design the Conditional Variation Attention (CV-Atten), whichpromotes disentangled learning and preserves the independence of knowledge representations. Extensive experiments on both simulated benchmarks and real-world robotic tasks demonstrate $Δ$VLA achieves state-of-the-art performance while improving efficiency. Code and real-world execution videos are available at https://github.com/JiuTian-VL/DeltaVLA.

cs.CV↗

arXiv2026

Phantom Mirage from Axion Dark Energy

Supernova (SN) and baryon acoustic oscillation (BAO) distance measures have recently provided hints that the dark energy is not only dynamical but apparently evolves from normal to phantom dark energy between redshifts $0<z<1$. A normal axion dark energy component in the mass range just below the Hubble scale can mimic a phantom component by appearing as dark energy at $z=1$ and dark matter at $z=0$, raising the possibility of a phantom mirage. We show that there is a wide range of axion dark energy contributions that can resolve the SN-BAO tension as well as thawing quintessence does, leaving BAO tension with the cosmic microwave background (CMB) for the distance measures from $z\sim 1$ to recombination to be resolved at high redshifts. With axions, raising the optical depth to reionization to $τ\approx 0.1$ works essentially as well as $w_0-w_a$ phantom dark energy for all but the lowE CMB data, with a remaining $Δχ^2\sim -16$ compared with $Λ$CDM, whereas a small spatial curvature of $Ω_K \sim 0.003$ can largely relax the full SN-BAO-CMB tension with a total $Δχ^2 \sim -12$.

astro-ph.CO↗

arXiv2025

H-GAR: A Hierarchical Interaction Framework via Goal-Driven Observation-Action Refinement for Robotic Manipulation

Unified video and action prediction models hold great potential for robotic manipulation, as future observations offer contextual cues for planning, while actions reveal how interactions shape the environment. However, most existing approaches treat observation and action generation in a monolithic and goal-agnostic manner, often leading to semantically misaligned predictions and incoherent behaviors. To this end, we propose H-GAR, a Hierarchical interaction framework via Goal-driven observation-Action Refinement.To anchor prediction to the task objective, H-GAR first produces a goal observation and a coarse action sketch that outline a high-level route toward the goal. To enable explicit interaction between observation and action under the guidance of the goal observation for more coherent decision-making, we devise two synergistic modules. (1) Goal-Conditioned Observation Synthesizer (GOS) synthesizes intermediate observations based on the coarse-grained actions and the predicted goal observation. (2) Interaction-Aware Action Refiner (IAAR) refines coarse actions into fine-grained, goal-consistent actions by leveraging feedback from the intermediate observations and a Historical Action Memory Bank that encodes prior actions to ensure temporal consistency. By integrating goal grounding with explicit action-observation interaction in a coarse-to-fine manner, H-GAR enables more accurate manipulation. Extensive experiments on both simulation and real-world robotic manipulation tasks demonstrate that H-GAR achieves state-of-the-art performance.

cs.RO↗

arXiv2025

REAP: Enhancing RAG with Recursive Evaluation and Adaptive Planning for Multi-Hop Question Answering

Retrieval-augmented generation (RAG) has been extensively employed to mitigate hallucinations in large language models (LLMs). However, existing methods for multi-hop reasoning tasks often lack global planning, increasing the risk of falling into local reasoning impasses. Insufficient exploitation of retrieved content and the neglect of latent clues fail to ensure the accuracy of reasoning outcomes. To overcome these limitations, we propose Recursive Evaluation and Adaptive Planning (REAP), whose core idea is to explicitly maintain structured sub-tasks and facts related to the current task through the Sub-task Planner (SP) and Fact Extractor (FE) modules. SP maintains a global perspective, guiding the overall reasoning direction and evaluating the task state based on the outcomes of FE, enabling dynamic optimization of the task-solving trajectory. FE performs fine-grained analysis over retrieved content to extract reliable answers and clues. These two modules incrementally enrich a logically coherent representation of global knowledge, enhancing the reliability and the traceability of the reasoning process. Furthermore, we propose a unified task paradigm design that enables effective multi-task fine-tuning, significantly enhancing SP's performance on complex, data-scarce tasks. We conduct extensive experiments on multiple public multi-hop datasets, and the results demonstrate that our method significantly outperforms existing RAG methods in both in-domain and out-of-domain settings, validating its effectiveness in complex multi-hop reasoning tasks.

cs.CL↗

arXiv2025

CoEmoGen: Towards Semantically-Coherent and Scalable Emotional Image Content Generation

Emotional Image Content Generation (EICG) aims to generate semantically clear and emotionally faithful images based on given emotion categories, with broad application prospects. While recent text-to-image diffusion models excel at generating concrete concepts, they struggle with the complexity of abstract emotions. There have also emerged methods specifically designed for EICG, but they excessively rely on word-level attribute labels for guidance, which suffer from semantic incoherence, ambiguity, and limited scalability. To address these challenges, we propose CoEmoGen, a novel pipeline notable for its semantic coherence and high scalability. Specifically, leveraging multimodal large language models (MLLMs), we construct high-quality captions focused on emotion-triggering content for context-rich semantic guidance. Furthermore, inspired by psychological insights, we design a Hierarchical Low-Rank Adaptation (HiLoRA) module to cohesively model both polarity-shared low-level features and emotion-specific high-level semantics. Extensive experiments demonstrate CoEmoGen's superiority in emotional faithfulness and semantic coherence from quantitative, qualitative, and user study perspectives. To intuitively showcase scalability, we curate EmoArt, a large-scale dataset of emotionally evocative artistic images, providing endless inspiration for emotion-driven artistic creation. The dataset and code are available at https://github.com/yuankaishen2001/CoEmoGen.

cs.CV↗

arXiv2025

Attention-based Neural Network Emulators for Multi-Probe Data Vectors Part III: Modeling The Next Generation Surveys

Machine learning can accelerate cosmological inferences that involve many sequential evaluations of computationally expensive data vectors. Previous works in this series have examined how machine learning architectures impact emulator accuracy and training time for optical shear and galaxy clustering 2-point function. In this final manuscript, we explore neural network performance when emulating Cosmic Microwave Background temperature and polarization power spectra. We maximize the volume of applicability in the parameter space of our emulators within the standard $Λ$-cold-dark-matter model while ensuring that errors are below cosmic variance. Relative to standard multi-layer perceptron architectures, we find the dot-product-attention mechanism reduces the number of outliers among testing cosmologies, defined as the fraction of testing points with $Δχ^2 > 0.2$ relative to \textsc{CAMB} outputs, for a wide range of training set sizes. Such precision enables attention-based emulators to be directly applied to real data without requiring any additional correction via importance sampling. Combined with pre-processing techniques and optimized activation and loss functions, attention-based models can meet the precision criteria set by current and future CMB and lensing experiments. For each of Planck, Simons Observatory, CMB S4, and CMB HD, we find the fraction of outlier points to be less than $10\%$ with around $2\times10^5$ to $4\times10^5$ training data vectors. We further explore the applications of these methods to supernova distance, weak lensing, and galaxy clustering, as well as alternative architectures and pre-processing techniques.

astro-ph.CO↗

arXiv2025

Leveraging Pathology Foundation Models for Panoptic Segmentation of Melanoma in H&E Images

Melanoma is an aggressive form of skin cancer with rapid progression and high metastatic potential. Accurate characterisation of tissue morphology in melanoma is crucial for prognosis and treatment planning. However, manual segmentation of tissue regions from haematoxylin and eosin (H&E) stained whole-slide images (WSIs) is labour-intensive and prone to inter-observer variability, this motivates the need for reliable automated tissue segmentation methods. In this study, we propose a novel deep learning network for the segmentation of five tissue classes in melanoma H&E images. Our approach leverages Virchow2, a pathology foundation model trained on 3.1 million histopathology images as a feature extractor. These features are fused with the original RGB images and subsequently processed by an encoder-decoder segmentation network (Efficient-UNet) to produce accurate segmentation maps. The proposed model achieved first place in the tissue segmentation task of the PUMA Grand Challenge, demonstrating robust performance and generalizability. Our results show the potential and efficacy of incorporating pathology foundation models into segmentation networks to accelerate computational pathology workflows.

eess.IV↗

arXiv2025

M-ary Precomputation-Based Accelerated Scalar Multiplication Algorithms for Enhanced Elliptic Curve Cryptography

Efficient scalar multiplication is critical for enhancing the performance of elliptic curve cryptography (ECC), especially in applications requiring large-scale or real-time cryptographic operations. This paper proposes an M-ary precomputation-based scalar multiplication algorithm, aiming to optimize both computational efficiency and memory usage. The method reduces the time complexity from $Θ(Q \log p)$ to $Θ\left(\frac{Q \log p}{\log Q}\right)$ and achieves a memory complexity of $Θ\left(\frac{Q \log p}{\log^2 Q}\right)$. Experiments on ElGamal encryption and NS3-based communication simulations validate its effectiveness. On secp256k1, the proposed method achieves up to a 59\% reduction in encryption time and 30\% memory savings. In network simulations, the binary-optimized variant reduces communication time by 22.1\% on secp384r1 and simulation time by 25.4\% on secp521r1. The results demonstrate the scalability, efficiency, and practical applicability of the proposed algorithm. The source code will be publicly released upon acceptance.

cs.CR↗