arXiv subjects

Junjie Zhou

Publications and source records attributed to Junjie Zhou.

At least 19 recordsLinked to original sources

arXiv2026

Benchmark Radar: A Living Database and Search Engine for AI Benchmarks and Evaluation

Benchmark researchers and developers of large language models (LLMs) and other AI systems need to find relevant evaluations, locate their benchmark datasets and code, and understand the settings behind reported scores. We present Benchmark Radar, a living database and search engine for retrieval and discovery of AI benchmarks, covering LLM evaluation, agentic and tool-use benchmarks, coding, reasoning, safety, and domain-specific evaluations. The system combines daily discovery of benchmark papers, repositories, datasets, and releases with a searchable benchmark catalog, mentions in model cards and technical reports, and score histories. It retains source identities and citations so readers can inspect candidate benchmarks and their evaluation evidence. Daily discovery draws on 37 sources: 13 direct connectors and 24 first-party research and engineering feeds. The catalog contains 1,283 source records drawn from 4 benchmark catalogs and 12,916 numeric observations on 790 records. We describe collection and retrieval, audit the full catalog, and examine benchmark saturation, adoption trends, and the limits of score comparisons. A worked example walks through a complete prior-art search, showing how to query the catalog and inspect benchmark evidence when designing a new evaluation. We release the web dashboard with a benchmark leaderboard, a Pareto frontier view of score against measured use, saturation and trend views, daily feeds, downloadable evidence, a command-line interface (CLI) for offline queries, and reproducible analysis.

cs.AI↗

arXiv2026

Beyond Chain-of-Thought: Rewrite as a Universal Interface for Generative Multimodal Embeddings

Multimodal Large Language Models (MLLMs) have emerged as a promising foundation for universal multimodal embeddings. Recent studies have shown that reasoning-driven generative multimodal embeddings can outperform discriminative embeddings on several embedding tasks. However, Chain-of-Thought (CoT) reasoning tends to generate redundant thinking steps and introduce semantic ambiguity in the summarized answers in broader retrieval scenarios. To address this limitation, we propose Rewrite-driven Multimodal Embedding (RIME), a unified framework that jointly optimizes generation and embedding through a retrieval-friendly rewrite. Meanwhile, we present the Cross-Mode Alignment (CMA) to bridge the generative and discriminative embedding spaces, enabling flexible mutual retrieval to trade off efficiency and accuracy. Based on this, we also introduce Refine Reinforcement Learning (Refine-RL) that treats discriminative embeddings as stable semantic anchors to guide the rewrite optimization. Extensive experiments on MMEB-V2, MRMR and UVRB demonstrate that RIME substantially outperforms prior generative embedding models while significantly reducing the length of thinking. Code is available at https://github.com/PeppaWu/RIME.

cs.CV↗

arXiv2026

WeMM-Embedding: WeChat Multi-Modal Embedding Technical Report

Universal multimodal embeddings are becoming a core component of modern AI systems, enabling heterogeneous content to be represented in a shared space for applications such as retrieval, recommendation, classification, and agentic systems. In this report, we present WeMM-Embedding, a family of universal multimodal embedding models supporting text, images, videos, visual documents, and arbitrarily interleaved multimodal inputs with flexible output dimensions. The family comprises 2B, 4B, and 9B variants and is trained in two stages: a large-scale multimodal alignment stage, followed by a refinement stage using curated data, fine-grained relevance supervision, and cross-scale knowledge transfer. Across extensive evaluations, WeMM-Embedding achieves leading performance on multiple public benchmarks. Notably, the 2B variant already surpasses the previously leading 8B open-source baseline on MMEB-v2, while the 9B variant further achieves a new state-of-the-art overall score of 80.6. WeMM-Embedding also demonstrates strong practical performance across WeChat applications, with substantial gains on a 26-task in-house benchmark and consistent improvements across 14 online A/B tests. It has been deployed at scale across recommendation and search applications, including WeChat Channels, Official Accounts, Moments, and e-commerce services. We have released the model weights and code to facilitate future research at https://github.com/Tencent/WeMM-Embedding.

cs.CV↗

arXiv2026

AffineTok: Semantic Affine Consistency for Diffusion-Friendly Visual Tokenizer

Visual tokenizers increasingly inject semantic supervision into latent spaces to make downstream diffusion easier. Yet how these semantics should be organized to facilitate denoising remains underexplored. In this paper, we define the semantic recovery objective: the denoising process should recover the semantic content of the clean image from noisy latent, and a good tokenizer should make it easier. Existing approaches train a projector to predict the semantics directly from the noisy latent. We argue that this predicts the average of clean-image semantics, whereas what really needs to be aligned is the semantics of averaged clean latents. More importantly, we demonstrate that the semantic recovery error orthogonally decomposes into the error of the optimal semantic prediction directly from the noisy latent and the error between these two predictions. We therefore identify their consistency as the missing requirement and call it Semantic Affine Consistency (SAC). To examine whether this overlooked requirement is closely related to downstream generation, we introduce M_SAC, a tokenizer-side proxy for SAC. Across the evaluated tokenizers and diffusion model scales, M_SAC closely tracks generation quality, reaching a Pearson correlation of 0.960 with SiT-XL gFID, thereby motivating SAC-guided tokenizer training. We then introduce AffineTok, which promotes SAC through two complementary, training-only components. Global Semantic Coordination Token (GSCT) coordinates the semantic organization of clean latents, keeping semantic averaging meaningful, while Posterior-Mean Semantic Alignment (PMSA) predicts posterior-mean latents from noisy inputs and supervises their semantics. On ImageNet 256, compared with the baseline, AffineTok reduces gFID by 26% at 20 epochs and, with continued training, achieves a new state-of-the-art gFID of 1.21 without classifier-free guidance and 1.10 with guidance.

cs.CV↗

arXiv2026

Predicting large-supercell defect formation energies from machine-learning charge density models trained on small supercells

First-principles defect calculations are often limited by the cost of the large supercells required to suppress image interactions. Machine-learning interatomic potentials (MLIPs) provide another alternative, but training defect MLIPs typically requires thousands of structures and weeks of data generation. Since charge density is the key to density-functional-theory (DFT), we propose a machine-learning charge density (MLCD) route for predicting defect formation energies with higher data efficiency. We optimize the training set by integrating small supercells of varying sizes for better extrapolation, allocating their proportions based on spatial charge-density analysis. With only 96 supercells containing 16--96 atoms as the dataset, MLCD accurately predicts the formation energies of four intrinsic defects in 360-atom supercells, with defect-wise mean absolute error below 0.05 eV. In contrast, MLIPs trained on the same dataset can err by more than 1 eV. These results show that charge-density learning enables more robust cross-size transfer than direct energy-force fitting and that mixed-size data design can substantially reduce the cost of defect prediction.

cond-mat.mtrl-sci↗

arXiv2026

Geo-Embed: Towards Unified Multimodal Embeddings for Urban Understanding

Geospatial and urban applications increasingly require models to compare heterogeneous evidence across street-view imagery, remote-sensing observations, text descriptions, region proposals, and temporal change cues. However, existing multimodal embedding models and benchmarks are still largely designed and evaluated around general-purpose image-text matching, leaving unclear whether unified embedding space can support heterogeneous geospatial tasks involving spatial relationships, fine-grained semantics, and temporal changes. To address this gap, we make three key contributions. First, we introduce GeoMEB, a large-scale multimodal embedding benchmark that standardizes 45 urban evaluation tasks across retrieval, visual question answering, change detection, classification, and visual grounding, together with training collections comprising 1.32M examples and 286K evaluation queries. Second, we present Geo-Embed, a unified embedding model that adapts a shared vision-language backbone to instruction-conditioned query-target matching over heterogeneous geospatial inputs, including single images, multiple images, text, regions, and masks. On GeoMEB, Geo-Embed achieves the strongest overall performance among representative multimodal embedders, with a 15.3% relative improvement over the strongest baseline. These results motivate future geospatial embedders that organize training and evaluation around explicit query-target relations, including semantic, cross-view, region-level, and temporal correspondence.

cs.CV↗

arXiv2026

BayesPO: Bayesian Prompt Optimization via Parallel-Tempered Gradient-Guided Discrete MCMC

Prompt optimization adapts large language models (LLMs) without updating model parameters, but many automatic prompt optimizers remain heuristic search procedures over candidate instructions. This paper studies prompt optimization as Bayesian posterior sampling over discrete prompt tokens. We define a posterior distribution by combining a task likelihood term, which rewards prompts that explain input-output examples, with a language-model prior, which favors fluent instructions. This converts prompt optimization into an energy-based posterior sampling problem, for which gradients can be used to guide discrete Markov chain Monte Carlo (MCMC) proposals over vocabulary tokens. We refer to our framework as BayesPO, short for Bayesian Prompt Optimization. In this paper, BayesPO is instantiated with Markov chain Monte Carlo: it uses a Metropolis-Hastings corrected Gibbs-with-Langevin (GwL) proposal and integrates parallel tempering for global exploration of rugged LLM-induced energy landscapes. The concrete sampler further adapts the GwL sampler to the practical constraints of non-weight-tied LLM embeddings. Experiments with Qwen2.5 models show that the sampler discovers semantically meaningful prompts on diagnostic tasks, that parallel tempering helps escape a local optimum in a poetry completion task, and that post-optimizing APE prompts on 24 instruction-induction subtasks improves average accuracy from 60.04% to 63.23%. The study also reveals two main limitations: energy minimization may overfit small optimization sets, and the current sampler remains computationally expensive. These findings position Bayesian prompt sampling as a principled post-optimization tool and point to a promising direction for probabilistic prompt optimization.

cs.CL↗

arXiv2026

Development and characterization of a millimeter-wave cold load prototype

Superconducting transition-edge sensors (TESs) are crucial detectors for cosmic microwave background (CMB) observations and require stable and tunable millimeter-wave cold loads for optical-efficiency calibration. This work presents the design, fabrication, and preliminary characterization of a 4-20 K millimeter-wave cold load prototype intended for integration into the 1 K stage of a dilution refrigerator and subsequent 40/90 GHz CMB TES calibration experiments. Two absorber prototypes based on commercially available CR-110 and a Stycast 2850FT composite were fabricated and studied. Simulation results show that both absorber structures exhibit small predicted steady-state temperature gradients and low normal-incidence reflectance in the target frequency bands. Room-temperature S11 measurements were used only to screen low-reflectance cold load prototype, and the measured results generally agree with the electromagnetic simulations. The measured S11 of the Stycast 2850FT composite is comparable to that of the commercial absorber TK RAM. Additionally, to explore a more readily obtainable alternative absorber material, TIE280-25AB was preliminarily evaluated by measuring its electromagnetic parameters. Based on the measured parameters, the simulated S11 of the TIE280-25AB pyramidal absorber structure is comparable to those of CR-110 and the Stycast 2850FT composite over 33-110 GHz. These results identify CR-110 and the Stycast 2850FT composite as promising absorbers for subsequent cryogenic evaluation. The absolute low-temperature emissivity, effective radiation temperature, and TES calibration performance remain to be established through future cryogenic radiometric and TES based optical-power measurements.

astro-ph.IM↗

arXiv2026

MSAVBench: Towards Comprehensive and Reliable Evaluation of Multi-Shot Audio-Video Generation

Video generation is rapidly evolving from single-shot synthesis to complex multi-shot audio-video (MSAV) narratives to meet real-world demands. However, evaluating such frontier models remains a fundamental challenge. Existing benchmarks are limited in scope and data diversity, and rely on rigid evaluation pipelines, preventing systematic and reliable assessment of modern MSAV models. To bridge these gaps, we introduce MSAVBench, the first comprehensive benchmark and adaptive hybrid evaluation framework for multi-shot audio-video generation. Our benchmark spans four key dimensions, video, audio, shot, and reference, covering diverse task settings, varying shot counts of up to 15, and challenging non-realistic scenarios. Our evaluation framework improves robustness through an adaptive self-correction mechanism for shot segmentation, instance-wise rubrics for subjective metrics, and tool-grounded evidence extraction for complex judgments. Furthermore, MSAVBench achieves high alignment with human judgments, reaching a Spearman rank correlation of 91.5%. Our systematic evaluation of 19 state-of-the-art closed- and open-source models shows that current systems still struggle with director-level control and fine-grained audio-visual synchronization, while modular or agentic generation pipelines offer a promising path toward narrowing the gap between open- and closed-source models. The benchmark data and evaluation code are publicly available at https://github.com/ali-vilab/MSAVBench.

cs.CV↗

arXiv2026

Nonradiative Multiphonon Model of Deep-Level Transient Spectroscopy: Beyond Henry-Lang Model

Deep-level transient spectroscopy (DLTS) is a key experimental method for defect characterization, yet its analysis remains controversial, and the two widely used models developed by Henry and Lang are conflicting. We show that the Henry-Lang model is valid only under the Condon approximation, as well as high-temperature and strong electron-phonon coupling approximations, which cause incorrect temperature dependence of carrier emission and capture. Here we develop a rigorous nonradiative multiphonon (NMP) model, and demonstrate that the temperature dependence is governed predominantly by effective phonons with large phonon wavefunction overlap and high thermal occupation. The effective phonons are strongly correlated with lattice relaxation.The neglect of this correlation in existing DLTS models introduces substantial errors when they are used to fit DLTS-measured emission rates. Our comparison for 21 different defects in 12 semiconductors, including Si, SiC and Ga$_2$O$_3$, shows that the Henry-Lang model gives a completely different temperature dependence of carrier capture cross section from that obtained using the rigorous NMP model, with errors reaching up to six orders of magnitude at room temperature. Our study highlights the necessity of revisiting previous DLTS analysis studies using the rigorous NMP model.

cond-mat.mtrl-sci↗

arXiv2026

A wrong ground-state structure of HfO$_2$ predicted by machine-learning interatomic potentials based on the PBE functional

Machine-learning interatomic potentials (MLIPs) have become powerful tools for material simulations. Many MLIPs are trained based on density functional theory (DFT) datasets generated with the Perdew-Burke-Ernzerhof (PBE) exchange-correlation functional. Using a PBE-based MLIP for HfO2, we identify a previously unreported low-energy I41/amd structure, which is predicted to be more stable than the well-known ground-state structure, the monoclinic P21/c structure. Since experiments show clearly that HfO2 takes the P21/c structure as the ground state, this is obviously a wrong prediction. Unfortunately, the same prediction is also made by widely used PBE-based foundation models such as NequIP-OAM-L and MatterSim-v1-5M. Comparisons among different DFT functionals show that this error originates from the PBE functional, which overstabilizes low-density structures containing sixfold Hf-O octahedral units, such as the I41/amd and Pbcn phases. The error also affects the calculated energy landscapes and barrier heights along ferroelectric HfO2 polarization switching paths when there are large lattice relaxations. Fortunately, the error can be largely suppressed by other functionals such as PBEsol and local density approximation. Our study serves as a warning about the impact of errors in exchange-correlation functional approximations on the reliability of MLIP simulations of crystal structures and phase transitions.

cond-mat.mtrl-sci↗

arXiv2026

Weak-strong uniqueness of the full coupled Navier-Stokes and Q-tensor system in dimension three

In this paper, we study the weak-strong uniqueness for the Leray-Hopf type weak solutions to the Beris-Edwards model of nematic liquid crystals in $\R^3$ with an arbitrary parameter $ξ\in\R$, which measures the ratio of tumbling and alignment effects caused by the flow. This result is obtained by proposing a new uniqueness criterion in terms of $(ΔQ,\nabla u)$ with regularity $L_t^qL_x^p$ for $\frac{2}{q}+\frac{3}{p}=\frac{3}{2}$ and $2\leq p\leq 6$, which enables us to deal with the additional nonlinear difficulties arising from the parameter $ξ$. Compared with the known results, our finding reveals that the criterion of weak-strong uniqueness for $ξ\ne 0$ is a sub-regime of the one for the corotational case. The associated regularity assumption rises with the nonlinearity of the model. Moreover, we establish the global well-posedness of this model for small initial data in $H^s$-framework.

math.AP↗

arXiv2026

When Do Markets Work? Multiplex Networks and Efficiency

We study an Arrow-Debreu economy with externalities generated by multiplex networks. Market equilibrium prices reflect both the preferences and scarcity of goods, consumers' network centralities arising from goods' externalities, as well as linkages across goods (layers) through the budget constraint. Despite the presence of externalities, competitive markets can still be efficient: the First and Second Welfare Theorems hold if either all networks are regular or all layers share the same network structure. When markets allocate goods inefficiently, a Lindahl equilibrium-implemented through personalized prices-can restore efficiency, but may leave some consumers worse off.

econ.TH↗

arXiv2026

Wan-Image: Pushing the Boundaries of Generative Visual Intelligence

We present Wan-Image, a unified visual generation system explicitly engineered to paradigm-shift image generation models from casual synthesizers into professional-grade productivity tools. While contemporary diffusion models excel at aesthetic generation, they frequently encounter critical bottlenecks in rigorous design workflows that demand absolute controllability, complex typography rendering, and strict identity preservation. To address these challenges, Wan-Image features a natively unified multi-modal architecture by synergizing the cognitive capabilities of large language models with the high-fidelity pixel synthesis of diffusion transformers, which seamlessly translates highly nuanced user intents into precise visual outputs. It is fundamentally powered by large-scale multi-modal data scaling, a systematic fine-grained annotation engine, and curated reinforcement learning data to surpass basic instruction following and unlock expert-level professional capabilities. These include ultra-long complex text rendering, hyper-diverse portrait generation, palette-guided generation, multi-subject identity preservation, coherent sequential visual generation, precise multi-modal interactive editing, native alpha-channel generation, and high-efficiency 4K synthesis. Across diverse human evaluations, Wan-Image exceeds Seedream 5.0 Lite and GPT Image 1.5 in overall performance, reaching parity with Nano Banana Pro in challenging tasks. Ultimately, Wan-Image revolutionizes visual content creation across e-commerce, entertainment, education, and personal productivity, redefining the boundaries of professional visual synthesis.

cs.CV↗

arXiv2026

OmniGen2: Towards Instruction-Aligned Multimodal Generation

In this work, we introduce OmniGen2, a versatile and open-source generative model designed to provide a unified solution for diverse generation tasks, including text-to-image, image editing, and in-context generation. Unlike OmniGen v1, OmniGen2 features two distinct decoding pathways for text and image modalities, utilizing unshared parameters and a decoupled image tokenizer. This design enables OmniGen2 to build upon existing multimodal understanding models without the need to re-adapt VAE inputs, thereby preserving the original text generation capabilities. To facilitate the training of OmniGen2, we developed comprehensive data construction pipelines, encompassing image editing and in-context generation data. Additionally, we introduce a reflection mechanism tailored for image generation tasks and curate a dedicated reflection dataset based on OmniGen2. Despite its relatively modest parameter size, OmniGen2 achieves competitive results on multiple task benchmarks, including text-to-image and image editing. To further evaluate in-context generation, also referred to as subject-driven tasks, we introduce a new benchmark named OmniContext. OmniGen2 achieves state-of-the-art performance among open-source models in terms of consistency. We will release our models, training code, datasets, and data construction pipeline to support future research in this field. Project Page: https://vectorspacelab.github.io/OmniGen2; GitHub Link: https://github.com/VectorSpaceLab/OmniGen2

cs.CV↗

arXiv2026

ReinVBC: A Model-based Reinforcement Learning Approach to Vehicle Braking Controller

Braking system, the key module to ensure the safety and steer-ability of current vehicles, relies on extensive manual calibration during production. Reducing labor and time consumption while maintaining the Vehicle Braking Controller (VBC) performance greatly benefits the vehicle industry. Model-based methods in offline reinforcement learning, which facilitate policy exploration within a data-driven dynamics model, offer a promising solution for addressing real-world control tasks. This work proposes ReinVBC, which applies an offline model-based reinforcement learning approach to deal with the vehicle braking control problem. We introduce useful engineering designs into the paradigm of model learning and utilization to obtain a reliable vehicle dynamics model and a capable braking policy. Several results demonstrate the capability of our method in real-world vehicle braking and its potential to replace the production-grade anti-lock braking system.

cs.RO↗

arXiv2026

ShowTable: Unlocking Creative Table Visualization with Collaborative Reflection and Refinement

While existing generation and unified models excel at general image generation, they struggle with tasks requiring deep reasoning, planning, and precise data-to-visual mapping abilities beyond general scenarios. To push beyond the existing limitations, we introduce a new and challenging task: creative table visualization, requiring the model to generate an infographic that faithfully and aesthetically visualizes the data from a given table. To address this challenge, we propose ShowTable, a pipeline that synergizes MLLMs with diffusion models via a progressive self-correcting process. The MLLM acts as the central orchestrator for reasoning the visual plan and judging visual errors to provide refined instructions, the diffusion execute the commands from MLLM, achieving high-fidelity results. To support this task and our pipeline, we introduce three automated data construction pipelines for training different modules. Furthermore, we introduce TableVisBench, a new benchmark with 800 challenging instances across 5 evaluation dimensions, to assess performance on this task. Experiments demonstrate that our pipeline, instantiated with different models, significantly outperforms baselines, highlighting its effective multi-modal reasoning, generation, and error correction capabilities.

cs.CV↗

arXiv2026

Advancing Cancer Prognosis with Hierarchical Fusion of Genomic, Proteomic and Pathology Imaging Data from a Systems Biology Perspective

To enhance the precision of cancer prognosis, recent research has increasingly focused on multimodal survival methods by integrating genomic data and histology images. However, current approaches overlook the fact that the proteome serves as an intermediate layer bridging genomic alterations and histopathological features while providing complementary biological information essential for survival prediction. This biological reality exposes another architectural limitation: existing integrative analysis studies fuse these heterogeneous data sources in a flat manner that fails to capture their inherent biological hierarchy. To address these limitations, we propose HFGPI, a hierarchical fusion framework that models the biological progression from genes to proteins to histology images from a systems biology perspective. Specifically, we introduce Molecular Tokenizer, a molecular encoding strategy that integrates identity embeddings with expression profiles to construct biologically informed representations for genes and proteins. We then develop Gene-Regulated Protein Fusion (GRPF), which employs graph-aware cross-attention with structure-preserving alignment to explicitly model gene-protein regulatory relationships and generate gene-regulated protein representations. Additionally, we propose Protein-Guided Hypergraph Learning (PGHL), which establishes associations between proteins and image patches, leveraging hypergraph convolution to capture higher-order protein-morphology relationships. The final features are progressively fused across hierarchical layers to achieve precise survival outcome prediction. Extensive experiments on five benchmark datasets demonstrate the superiority of HFGPI over state-of-the-art methods.

cs.CV↗