arXiv subjects

Qian Zhao

Publications and source records attributed to Qian Zhao.

At least 37 records · Page 2Linked to original sources

arXiv2026

Layout-Guided Controllable Pathology Image Generation with In-Context Diffusion Transformers

Controllable pathology image synthesis requires reliable regulation of spatial layout, tissue morphology, and semantic detail. However, existing text-guided diffusion models offer only coarse global control and lack the ability to enforce fine-grained structural constraints. Progress is further limited by the absence of large datasets that pair patch-level spatial layouts with detailed diagnostic descriptions, since generating such annotations for gigapixel whole-slide images is prohibitively time-consuming for human experts. To overcome these challenges, we first develop a scalable multi-agent LVLM annotation framework that integrates image description, diagnostic step extraction, and automatic quality judgment into a coordinated pipeline, and we evaluate the reliability of the system through a human verification process. This framework enables efficient construction of fine-grained and clinically aligned supervision at scale. Building on the curated data, we propose In-Context Diffusion Transformer (IC-DiT), a layout-aware generative model that incorporates spatial layouts, textual descriptions, and visual embeddings into a unified diffusion transformer. Through hierarchical multimodal attention, IC-DiT maintains global semantic coherence while accurately preserving structural and morphological details. Extensive experiments on five histopathology datasets show that IC-DiT achieves higher fidelity, stronger spatial controllability, and better diagnostic consistency than existing methods. In addition, the generated images serve as effective data augmentation resources for downstream tasks such as cancer classification and survival analysis.

cs.CV↗

arXiv2026

Topological Metamaterial for Magnetic Resonance Imaging

Magnetic Resonance Imaging (MRI) is crucial in global healthcare, but the traditional receive coils, as a core component of MRI, SNR enhancement is limited due to the optimization of channel number and magnetic field strength faces high cost and complexity challenges. Here, we demonstrate the use of a topological material to enhance MRI signal reception. Designed with a stack of weak couplings, this material forms quasi-two-dimensional dual topological boundary states. High properties are achieved through low-loss signal transmission via these topological states, as well as only enhanced local magnetic fields and increased number of channels. Initial tests demonstrate superior performance and accessibility compared to commercial coils, suggesting significant potential. This concept introduces a transformative paradigm for all MRI coil designs.

physics.app-ph↗

arXiv2026

STEP3-VL-10B Technical Report

We present STEP3-VL-10B, a lightweight open-source foundation model designed to redefine the trade-off between compact efficiency and frontier-level multimodal intelligence. STEP3-VL-10B is realized through two strategic shifts: first, a unified, fully unfrozen pre-training strategy on 1.2T multimodal tokens that integrates a language-aligned Perception Encoder with a Qwen3-8B decoder to establish intrinsic vision-language synergy; and second, a scaled post-training pipeline featuring over 1k iterations of reinforcement learning. Crucially, we implement Parallel Coordinated Reasoning (PaCoRe) to scale test-time compute, allocating resources to scalable perceptual reasoning that explores and synthesizes diverse visual hypotheses. Consequently, despite its compact 10B footprint, STEP3-VL-10B rivals or surpasses models 10$\times$-20$\times$ larger (e.g., GLM-4.6V-106B, Qwen3-VL-235B) and top-tier proprietary flagships like Gemini 2.5 Pro and Seed-1.5-VL. Delivering best-in-class performance, it records 92.2% on MMBench and 80.11% on MMMU, while excelling in complex reasoning with 94.43% on AIME2025 and 75.95% on MathVision. We release the full model suite to provide the community with a powerful, efficient, and reproducible baseline.

cs.CV↗

arXiv2025

MVGSR: Multi-View Consistent 3D Gaussian Super-Resolution via Epipolar Guidance

Scenes reconstructed by 3D Gaussian Splatting (3DGS) trained on low-resolution (LR) images are unsuitable for high-resolution (HR) rendering. Consequently, a 3DGS super-resolution (SR) method is needed to bridge LR inputs and HR rendering. Early 3DGS SR methods rely on single-image SR networks, which lack cross-view consistency and fail to fuse complementary information across views. More recent video-based SR approaches attempt to address this limitation but require strictly sequential frames, limiting their applicability to unstructured multi-view datasets. In this work, we introduce Multi-View Consistent 3D Gaussian Splatting Super-Resolution (MVGSR), a framework that focuses on integrating multi-view information for 3DGS rendering with high-frequency details and enhanced consistency. We first propose an Auxiliary View Selection Method based on camera poses, making our method adaptable for arbitrarily organized multi-view datasets without the need of temporal continuity or data reordering. Furthermore, we introduce, for the first time, an epipolar-constrained multi-view attention mechanism into 3DGS SR, which serves as the core of our proposed multi-view SR network. This design enables the model to selectively aggregate consistent information from auxiliary views, enhancing the geometric consistency and detail fidelity of 3DGS representations. Extensive experiments demonstrate that our method achieves state-of-the-art performance on both object-centric and scene-level 3DGS SR benchmarks.

cs.CV↗

arXiv2025

Discovering Influential Factors in Variational Autoencoders

In the field of machine learning, it is still a critical issue to identify and supervise the learned representation without manually intervening or intuition assistance to extract useful knowledge or serve for the downstream tasks. In this work, we focus on supervising the influential factors extracted by the variational autoencoder(VAE). The VAE is proposed to learn independent low dimension representation while facing the problem that sometimes pre-set factors are ignored. We argue that the mutual information of the input and each learned factor of the representation plays a necessary indicator of discovering the influential factors. We find the VAE objective inclines to induce mutual information sparsity in factor dimension over the data intrinsic dimension and therefore result in some non-influential factors whose function on data reconstruction could be ignored. We show mutual information also influences the lower bound of the VAE's reconstruction error and downstream classification task. To make such indicator applicable, we design an algorithm for calculating the mutual information for the VAE and prove its consistency. Experimental results on MNIST, CelebA and DEAP datasets show that mutual information can help determine influential factors, of which some are interpretable and can be used to further generation and classification tasks, and help discover the variant that connects with emotion on DEAP dataset.

cs.LG↗

arXiv2025

Every Activation Boosted: Scaling General Reasoner to 1 Trillion Open Language Foundation

We introduce Ling 2.0, a series reasoning-oriented language foundation built upon the principle that every activation boosts reasoning capability. Designed to scale from tens of billions to one trillion parameters under a unified Mixture-of-Experts (MoE) paradigm, Ling 2.0 emphasizes high sparsity, cross-scale consistency, and efficiency guided by empirical scaling laws. The series includes three non-thinking (instruct) models - Ling-mini-2.0, Ling-flash-2.0, and Ling-1T - ranging from 16B to 1T total parameters and achieving up to 7-fold active-compute efficiency compared with dense counterparts. Ling 2.0 integrates coordinated innovations across model architecture, pre-training, post-training, and infrastructure: a high-sparsity MoE with MTP for efficient reasoning, reasoning-oriented data and mid-training CoT activation, reinforcement-based fine-tuning (DFT, Evo-CoT), and full-scale FP8 training with fine-grained heterogeneous pipelines. At the trillion scale, Ling-1T establishes a new Pareto frontier of reasoning accuracy versus computational efficiency, demonstrating that sparse activation, when properly aligned with reasoning objectives, enables scalable and efficient intelligence. Collectively, Ling 2.0 provides a coherent, open, and efficient foundation for advancing future reasoning and thinking models, including the Ring series built upon the same base.

cs.CL↗

arXiv2025

Every Attention Matters: An Efficient Hybrid Architecture for Long-Context Reasoning

In this technical report, we present the Ring-linear model series, specifically including Ring-mini-linear-2.0 and Ring-flash-linear-2.0. Ring-mini-linear-2.0 comprises 16B parameters and 957M activations, while Ring-flash-linear-2.0 contains 104B parameters and 6.1B activations. Both models adopt a hybrid architecture that effectively integrates linear attention and softmax attention, significantly reducing I/O and computational overhead in long-context inference scenarios. Compared to a 32 billion parameter dense model, this series reduces inference cost to 1/10, and compared to the original Ring series, the cost is also reduced by over 50%. Furthermore, through systematic exploration of the ratio between different attention mechanisms in the hybrid architecture, we have identified the currently optimal model structure. Additionally, by leveraging our self-developed high-performance FP8 operator library-linghe, overall training efficiency has been improved by 50%. Benefiting from the high alignment between the training and inference engine operators, the models can undergo long-term, stable, and highly efficient optimization during the reinforcement learning phase, consistently maintaining SOTA performance across multiple challenging complex reasoning benchmarks.

cs.LG↗

arXiv2025

Diffractive Magic Cube Network with Super-high Capacity Enabled by Mechanical Reconfiguration

Free-space wavefront manipulation devices have emerged as powerful platforms for advanced optical information systems. In response to the challenges posed by the exponential growth of optical information, optical multiplexing and dynamic reconfigurable devices are being actively explored to the enhance system capacity. Among them, coarse-grained mechanically reconfigurable mechanism offers a cost-effective and low-complexity approach for capacity enhancement. However, the channel numbers achieved in current studies are insufficient for practical applications because of inadequate mechanical transformations and suboptimal optimization models. In this article, a diffractive magic cube network (DMCN) is proposed to advance the multiplexing capacity of mechanically reconfigurable system. We utilized the diffractive deep neural network (D2NN) model to jointly optimize the subset of channels generated by the combination of three mechanical operations, permutation, translation, and rotation. The 144-channel holograms, 108-channel single/double focus, 60-channel single/multi-mode OAM beam generation were experimentally demonstrated using diffractive optical elements (DOEs). An equivalent connectivity law was formulated to improve model scalability. Our strategy not only provides a novel paradigm to improve system capacity to super-high level with low crosstalk, but also paves the way for new advancements in optical storage, computing, communication, and photolithography.

physics.optics↗

arXiv2025

Deep learning for flash drought forecasting and interpretation

Flash droughts are increasingly occurring worldwide due to climate change, causing widespread socioeconomic and agricultural losses. However, timely and accurate flash drought forecasting remains challenging for operational forecast systems due to uncertain representations of interactive hydroclimate processes. Here, we introduce the Interpretable Transformer for Drought (IT-Drought), a deep learning model based on publicly available hydroclimate data. IT-Drought skillfully forecasts soil moisture up to 42 days and flash droughts up to 11 days ahead on average across the contiguous U.S., substantially outperforming existing forecast systems limited to 10-day skillful soil moisture forecasting and unable to forecast flash droughts. IT-Drought offers novel data-driven interpretability on flash drought onset, benchmarking the memory effects of soil moisture, and revealing heterogeneous impacts and effective time ranges of climatic conditions, predominantly radiation and temperature. The findings highlight the potential of leveraging IT-Drought for early warning and mechanistic investigation of flash droughts, ultimately supporting drought risk mitigation.

physics.geo-ph↗

arXiv2025

Mapping beyond diseases: Controlled variable selection for secondary phenotypes using tilted knockoffs

Researchers in biomedical studies often work with samples that are not selected uniformly at random from the population of interest, a major example being a case-control study. While these designs are motivated by specific scientific questions, it is often of interest to use the data collected to pursue secondary lines of investigations. In these cases, ignoring the fact that observations are not sampled uniformly at random can lead to spurious results. For example, in a case-control study, one might identify a spurious association between an exposure and a secondary phenotype when both affect the case-control status. This phenomenon is known as collider bias in the causal inference literature. While tests of independence under biased sampling are available, these methods typically do not apply when the number of variables is large. Here, we are interested in using the biased sample to select important exposures among a multitude of possible variables with replicability guarantees. While the model-X knockoff framework has been developed to test conditional independence hypotheses with False Discovery Rate (FDR) control, we show that its naive application fails to control FDR in the presence of biased sampling. We show how tilting the population distribution with the selection probability and constructing knockoff variables according to this tilted distribution instead leads to selection with FDR control. We study the FDR and power of the tilted knockoff method using simulated examples, and apply it to identify genetic underpinning of endophenotypes in a case-control study.

stat.ME↗

arXiv2025

Near-Peer Mentoring in Data Science: A Plot for Mutual Growth

Universities have been expanding undergraduate data science programs. Involving graduate students in these new opportunities can foster their growth as data science educators. We describe two programs that employ a near-peer mentoring structure, in which graduate students mentor undergraduates, to (1) strengthen their teaching and mentoring skills and (2) provide research and learning experiences for undergraduates from diverse backgrounds. In the Data Science for Social Good program, undergraduate participants work in teams to tackle a data science project with social impact. Graduate mentors guide project work and provide just-in-time teaching and feedback. The Stanford Mentoring in Data Science course offers training in effective and inclusive mentorship strategies. In an experiential learning framework, enrolled graduate students are paired with undergraduate students from non-R1 schools, whom they mentor through weekly one-on-one remote meetings. In end-of-program surveys, mentors reported growth through both programs. Drawing from these experiences, we developed a self-paced mentor training guide, which engages teaching, mentoring and project management abilities. These initiatives and the shared materials can serve as prototypes of future programs that cultivate mutual growth of both undergraduate and graduate students in a high-touch, inclusive, and encouraging environment.

stat.OT↗

arXiv2025

Generating Cylindrical Vector γ Rays via Beam-Target Interactions: Towards Structured Light at High Energies

Structured γ rays, particularly cylindrical vector γ rays, offer promising tools for sub-nuclear imaging and polarization-sensitive probes in fundamental research and applications, but conventional optical methods face great challenges at such photon energy. Here, we put forward a novel method generating such γ rays through relativistic beam-target interactions. For instance, radially polarized γ rays can be generated by using a dense electron beam striking a multifoil target. We find that the radial polarization is transferred from the generated coherent transition radiation (CTR) fields to $γ$ photons through nonlinear Compton scattering, with the high polarization preserved by phase matching. Three-dimensional spin-resolved simulations demonstrate radial polarization degrees approaching 60\%. Furthermore, these γ rays can decay into azimuthally spin-polarized positrons via the nonlinear Breit-Wheeler process, with their spins aligning along the CTR magnetic field. Our work extends the concept of structured light into the γ-ray regime, offering new prospects for broad fields such as nuclear structure probing, fundamental symmetries tests, polarization-sensitive studies in extreme conditions, and laboratory astrophysical observations.

physics.optics↗

arXiv2025

WSM: Decay-Free Learning Rate Schedule via Checkpoint Merging for LLM Pre-training

Recent advances in learning rate (LR) scheduling have demonstrated the effectiveness of decay-free approaches that eliminate the traditional decay phase while maintaining competitive performance. Model merging techniques have emerged as particularly promising solutions in this domain. We present Warmup-Stable and Merge (WSM), a general framework that establishes a formal connection between learning rate decay and model merging. WSM provides a unified theoretical foundation for emulating various decay strategies-including cosine decay, linear decay and inverse square root decay-as principled model averaging schemes, while remaining fully compatible with diverse optimization methods. Through extensive experiments, we identify merge duration-the training window for checkpoint aggregation-as the most critical factor influencing model performance, surpassing the importance of both checkpoint interval and merge quantity. Our framework consistently outperforms the widely-adopted Warmup-Stable-Decay (WSD) approach across multiple benchmarks, achieving significant improvements of +3.5% on MATH, +2.9% on HumanEval, and +5.5% on MMLU-Pro. The performance advantages extend to supervised fine-tuning scenarios, highlighting WSM's potential for long-term model refinement.

cs.CL↗

arXiv2025

Statistical Theory of Multi-stage Newton Iteration Algorithm for Online Continual Learning

We focus on the critical challenge of handling non-stationary data streams in online continual learning environments, where constrained storage capacity prevents complete retention of historical data, leading to catastrophic forgetting during sequential task training. To more effectively analyze and address the problem of catastrophic forgetting in continual learning, we propose a novel continual learning framework from a statistical perspective. Our approach incorporates random effects across all model parameters and allows the dimension of parameters to diverge to infinity, offering a general formulation for continual learning problems. To efficiently process streaming data, we develop a Multi-step Newton Iteration algorithm that significantly reduces computational costs in certain scenarios by alleviating the burden of matrix inversion. Theoretically, we derive the asymptotic normality of the estimator, enabling subsequent statistical inference. Comprehensive validation through synthetic data experiments and two real datasets analyses demonstrates the effectiveness of our proposed method.

stat.ME↗

arXiv2025

Universal Magnetocaloric Effect near Quantum Critical Point of Magnon Bose-Einstein Condensation

Bose-Einstein condensation (BEC), a macroscopic quantum phenomenon arising from phase coherence and bosonic statistics, has been realized in quantum magnets. Here, we report the observation of a universal magnetocaloric effect (MCE) near a BEC quantum critical point (QCP) in copper sulfate crystal ($CuSO_4 \cdot 5H_2O$). By conducting magnetocaloric and nuclear magnetic resonance measurements, we uncover a field-driven BEC QCP, evidenced by the universal scaling law $T_c \propto (B_c - B)^{2/3}$ and the perfect data collapse of the magnetic Grüneisen ratio. Thermal excitation triggers a dimensional crossover to a 1D quantum-critical regime, where the MCE scaling strictly matches the universality class of 1D Fermi gases. Notably, the quantum-critical MCE enables cooling down to 12.8 mK without helium-3, with very fast thermal relaxation rate that is critical for high cooling power. This work demonstrates the universal MCE in magnon BEC systems, using a common copper sulfate compound as a paradigmatic example, and paves the way for next-generation sub-Kelvin cooling.

cond-mat.str-el↗

arXiv2025

High-Dimensional Matrix-Variate Diffusion Index Models for Time Series Forecasting

This paper proposes a novel diffusion-index model for forecasting when predictors are high-dimensional matrix-valued time series. We apply an $α$-PCA method to extract low-dimensional matrix factors and build a bilinear regression linking future outcomes to these factors, estimated via iterative least squares. To handle weak factor structures, we introduce a supervised screening step to select informative rows and columns. Theoretical properties, including consistency and asymptotic normality, are established. Simulations and real data show that our method significantly improves forecast accuracy, with the screening procedure providing additional gains over standard benchmarks in out-of-sample mean squared forecast error.

econ.EM↗

arXiv2025

MemoCue: Empowering LLM-Based Agents for Human Memory Recall via Strategy-Guided Querying

Agent-assisted memory recall is one critical research problem in the field of human-computer interaction. In conventional methods, the agent can retrieve information from its equipped memory module to help the person recall incomplete or vague memories. The limited size of memory module hinders the acquisition of complete memories and impacts the memory recall performance in practice. Memory theories suggest that the person's relevant memory can be proactively activated through some effective cues. Inspired by this, we propose a novel strategy-guided agent-assisted memory recall method, allowing the agent to transform an original query into a cue-rich one via the judiciously designed strategy to help the person recall memories. To this end, there are two key challenges. (1) How to choose the appropriate recall strategy for diverse forgetting scenarios with distinct memory-recall characteristics? (2) How to obtain the high-quality responses leveraging recall strategies, given only abstract and sparsely annotated strategy patterns? To address the challenges, we propose a Recall Router framework. Specifically, we design a 5W Recall Map to classify memory queries into five typical scenarios and define fifteen recall strategy patterns across the corresponding scenarios. We then propose a hierarchical recall tree combined with the Monte Carlo Tree Search algorithm to optimize the selection of strategy and the generation of strategy responses. We construct an instruction tuning dataset and fine-tune multiple open-source large language models (LLMs) to develop MemoCue, an agent that excels in providing memory-inspired responses. Experiments on three representative datasets show that MemoCue surpasses LLM-based methods by 17.74% in recall inspiration. Further human evaluation highlights its advantages in memory-recall applications.

cs.AI↗

arXiv2025

Enhancing Underwater Imaging with 4-D Light Fields: Dataset and Method

In this paper, we delve into the realm of 4-D light fields (LFs) to enhance underwater imaging plagued by light absorption, scattering, and other challenges. Contrasting with conventional 2-D RGB imaging, 4-D LF imaging excels in capturing scenes from multiple perspectives, thereby indirectly embedding geometric information. This intrinsic property is anticipated to effectively address the challenges associated with underwater imaging. By leveraging both explicit and implicit depth cues present in 4-D LF images, we propose a progressive, mutually reinforcing framework for underwater 4-D LF image enhancement and depth estimation. Specifically, our framework explicitly utilizes estimated depth information alongside implicit depth-related dynamic convolutional kernels to modulate output features. The entire framework decomposes this complex task, iteratively optimizing the enhanced image and depth information to progressively achieve optimal enhancement results. More importantly, we construct the first 4-D LF-based underwater image dataset for quantitative evaluation and supervised training of learning-based methods, comprising 75 underwater scenes and 3675 high-resolution 2K pairs. To craft vibrant and varied underwater scenes, we build underwater environments with various objects and adopt several types of degradation. Through extensive experimentation, we showcase the potential and superiority of 4-D LF-based underwater imaging vis-a-vis traditional 2-D RGB-based approaches. Moreover, our method effectively corrects color bias and achieves state-of-the-art performance. The dataset and code will be publicly available at https://github.com/linlos1234/LFUIE.

cs.CV↗