arXiv subjects

Bin Chen

Publications and source records attributed to Bin Chen.

At least 199 records · Page 11Linked to original sources

arXiv2025

A survey of using EHR as real-world evidence for discovering and validating new drug indications

Electronic Health Records (EHRs) have been increasingly used as real-world evidence (RWE) to support the discovery and validation of new drug indications. This paper surveys current approaches to EHR-based drug repurposing, covering data sources, processing methodologies, and representation techniques. It discusses study designs and statistical frameworks for evaluating drug efficacy. Key challenges in validation are discussed, with emphasis on the role of large language models (LLMs) and target trial emulation. By synthesizing recent developments and methodological advances, this work provides a foundational resource for researchers aiming to translate real-world data into actionable drug-repurposing evidence.

stat.AP↗

arXiv2025

Chain-of-Generation: Progressive Latent Diffusion for Text-Guided Molecular Design

Text-conditioned molecular generation aims to translate natural-language descriptions into chemical structures, enabling scientists to specify functional groups, scaffolds, and physicochemical constraints without handcrafted rules. Diffusion-based models, particularly latent diffusion models (LDMs), have recently shown promise by performing stochastic search in a continuous latent space that compactly captures molecular semantics. Yet existing methods rely on one-shot conditioning, where the entire prompt is encoded once and applied throughout diffusion, making it hard to satisfy all the requirements in the prompt. We discuss three outstanding challenges of one-shot conditioning generation, including the poor interpretability of the generated components, the failure to generate all substructures, and the overambition in considering all requirements simultaneously. We then propose three principles to address those challenges, motivated by which we propose Chain-of-Generation (CoG), a training-free multi-stage latent diffusion framework. CoG decomposes each prompt into curriculum-ordered semantic segments and progressively incorporates them as intermediate goals, guiding the denoising trajectory toward molecules that satisfy increasingly rich linguistic constraints. To reinforce semantic guidance, we further introduce a post-alignment learning phase that strengthens the correspondence between textual and molecular latent spaces. Extensive experiments on benchmark and real-world tasks demonstrate that CoG yields higher semantic alignment, diversity, and controllability than one-shot baselines, producing molecules that more faithfully reflect complex, compositional prompts while offering transparent insight into the generation process.

cs.LG↗

arXiv2025

FaSDiff: Balancing Perception and Semantics in Face Compression via Stable Diffusion Priors

With the increasing deployment of facial image data across a wide range of applications, efficient compression tailored to facial semantics has become critical for both storage and transmission. While recent learning-based face image compression methods have achieved promising results, they often suffer from degraded reconstruction quality at low bit rates. Directly applying diffusion-based generative priors to this task leads to suboptimal performance in downstream machine vision tasks, primarily due to poor preservation of high-frequency details. In this work, we propose FaSDiff (\textbf{Fa}cial Image Compression with a \textbf{S}table \textbf{Diff}usion Prior), a novel diffusion-driven compression framework designed to enhance both visual fidelity and semantic consistency. FaSDiff incorporates a high-frequency-sensitive compressor to capture fine-grained details and generate robust visual prompts for guiding the diffusion model. To address low-frequency degradation, we further introduce a hybrid low-frequency enhancement module that disentangles and preserves semantic structures, enabling stable modulation of the diffusion prior during reconstruction. By jointly optimizing perceptual quality and semantic preservation, FaSDiff effectively balances human visual fidelity and machine vision accuracy. Extensive experiments demonstrate that FaSDiff outperforms state-of-the-art methods in both perceptual metrics and downstream task performance.

cs.CV↗

arXiv2025

SWE-Compass: Towards Unified Evaluation of Agentic Coding Abilities for Large Language Models

Evaluating large language models (LLMs) for software engineering has been limited by narrow task coverage, language bias, and insufficient alignment with real-world developer workflows. Existing benchmarks often focus on algorithmic problems or Python-centric bug fixing, leaving critical dimensions of software engineering underexplored. To address these gaps, we introduce SWE-Compass1, a comprehensive benchmark that unifies heterogeneous code-related evaluations into a structured and production-aligned framework. SWE-Compass spans 8 task types, 8 programming scenarios, and 10 programming languages, with 2000 high-quality instances curated from authentic GitHub pull requests and refined through systematic filtering and validation. We benchmark ten state-of-the-art LLMs under two agentic frameworks, SWE-Agent and Claude Code, revealing a clear hierarchy of difficulty across task types, languages, and scenarios. Moreover, by aligning evaluation with real-world developer practices, SWE-Compass provides a rigorous and reproducible foundation for diagnosing and advancing agentic coding capabilities in large language models.

cs.SE↗

arXiv2025

The Exospace Weather Frontier

Space weather is among the most powerful and least understood forces shaping planetary atmospheres. In our Solar System, we observe its effects directly: atmospheric escape, chemical disruption, and spectacular auroral displays. Yet for exoplanets, we lack the tools and data to comprehensively assess the impacts of space weather, especially invisible elements like stellar winds, coronal mass ejections, energetic particles, and variable interplanetary magnetic fields. This problem lies at the intersection of four key fields: heliophysics, planetary science, astrobiology, and astrophysics. In 2023--2025, experts from these four fields convened at the W. M. Keck Institute for Space Studies to explore pathways for advancing the study of exospace weather. Organizing the subject into five core themes -- planets and their stellar particle environments, stellar magnetism and space weather modeling, quasi-steady stellar winds, transient events, and programmatic pathways -- our team synthesized concepts from across relevant fields and identified a wide array of opportunities for progress. This report is the product of that effort. It assembles cross-disciplinary knowledge; highlights outstanding theoretical challenges; explores promising innovations in observation, modeling, methodology, and instrumentation; and makes recommendations for accelerating community-wide progress. Together, these lay out a path to transforming the challenging, yet tractable problem of exospace weather into a foundational element of our understanding exoplanetary systems, and our own Solar System, in their entirety.

astro-ph.IM↗

arXiv2025

KAT-Coder Technical Report

Recent advances in large language models (LLMs) have enabled progress in agentic coding, where models autonomously reason, plan, and act within interactive software development workflows. However, bridging the gap between static text-based training and dynamic real-world agentic execution remains a core challenge. In this technical report, we present KAT-Coder, a large-scale agentic code model trained through a multi-stage curriculum encompassing Mid-Term Training, Supervised Fine-Tuning (SFT), Reinforcement Fine-Tuning (RFT), and Reinforcement-to-Deployment Adaptation. The Mid-Term stage enhances reasoning, planning, and reflection capabilities through a corpus of real software engineering data and synthetic agentic interactions. The SFT stage constructs a million-sample dataset balancing twenty programming languages, ten development contexts, and ten task archetypes. The RFT stage introduces a novel multi-ground-truth reward formulation for stable and sample-efficient policy optimization. Finally, the Reinforcement-to-Deployment phase adapts the model to production-grade IDE environments using Error-Masked SFT and Tree-Structured Trajectory Training. In summary, these stages enable KAT-Coder to achieve robust tool-use reliability, instruction alignment, and long-context reasoning, forming a deployable foundation for real-world intelligent coding agents. Our KAT series 32B model, KAT-Dev, has been open-sourced on https://huggingface.co/Kwaipilot/KAT-Dev.

cs.CL↗

arXiv2025

Probing the Penrose Process: Images of Split Hotspots and Their Observational Signatures

While theoretically established for decades, the Penrose process - energy extraction from rotating black holes - still lacks clear observational evidence. A promising theoretical framework posits magnetic reconnection in the ergosphere as a trigger, causing a plasmoid to separate into an escaping positive-energy fragment and an infalling negative-energy one. In this work, we investigate the observational imprints of this scenario. We treat the energized plasmoid as a hotspot and calculate its light curves for a realistic plasma magnetization. In particular, we further compare with the scenario in which the plasmoid, after fragmentation, falls into the black hole with positive energy, while all other conditions remain unchanged. Our results reveal that the process of fragmentation generates distinct flares, whose characteristics depend heavily on whether the infalling fragment carries negative or positive energy. We propose that these differences serve as identifiable signatures of the Penrose process.

astro-ph.HE↗

arXiv2025

A Study on Privacy-Preserving Scholarship Evaluation Based on Decentralized Identity and Zero-Knowledge Proofs

Traditional centralized scholarship evaluation processes typically require students to submit detailed academic records and qualification information, which exposes them to risks of data leakage and misuse, making it difficult to simultaneously ensure privacy protection and transparent auditability. To address these challenges, this paper proposes a scholarship evaluation system based on Decentralized Identity (DID) and Zero-Knowledge Proofs (ZKP). The system aggregates multidimensional ZKPs off-chain, and smart contracts verify compliance with evaluation criteria without revealing raw scores or computational details. Experimental results demonstrate that the proposed solution not only automates the evaluation efficiently but also maximally preserves student privacy and data integrity, offering a practical and trustworthy technical paradigm for higher education scholarship programs.

cs.CR↗

arXiv2025

Grounding Language with Vision: A Conditional Mutual Information Calibrated Decoding Strategy for Reducing Hallucinations in LVLMs

Large Vision-Language Models (LVLMs) are susceptible to hallucinations, where generated responses seem semantically plausible yet exhibit little or no relevance to the input image. Previous studies reveal that this issue primarily stems from LVLMs' over-reliance on language priors while disregarding the visual information during decoding. To alleviate this issue, we introduce a novel Conditional Pointwise Mutual Information (C-PMI) calibrated decoding strategy, which adaptively strengthens the mutual dependency between generated texts and input images to mitigate hallucinations. Unlike existing methods solely focusing on text token sampling, we propose to jointly model the contributions of visual and textual tokens to C-PMI, formulating hallucination mitigation as a bi-level optimization problem aimed at maximizing mutual information. To solve it, we design a token purification mechanism that dynamically regulates the decoding process by sampling text tokens remaining maximally relevant to the given image, while simultaneously refining image tokens most pertinent to the generated response. Extensive experiments across various benchmarks reveal that the proposed method significantly reduces hallucinations in LVLMs while preserving decoding efficiency.

cs.CL↗

arXiv2025

HiPO: Hybrid Policy Optimization for Dynamic Reasoning in LLMs

Large Language Models (LLMs) increasingly rely on Chain-of-Thought (CoT) reasoning to improve accuracy on complex tasks. However, always generating lengthy reasoning traces is inefficient, leading to excessive token usage and higher inference costs. This paper introduces the Hybrid Policy Optimization (i.e., HiPO), a framework for adaptive reasoning control that enables LLMs to selectively decide when to engage in detailed reasoning (Think-on) and when to respond directly (Think-off). Specifically, HiPO combines a hybrid data pipelineproviding paired Think-on and Think-off responseswith a hybrid reinforcement learning reward system that balances accuracy and efficiency while avoiding over-reliance on detailed reasoning. Experiments across mathematics and coding benchmarks demonstrate that HiPO can substantially reduce token length while maintaining or improving accuracy. Finally, we hope HiPO a can be a principled approach for efficient adaptive reasoning, advancing the deployment of reasoning-oriented LLMs in real-world, resource-sensitive settings.

cs.CL↗

arXiv2025

Quantum field theory in flat spacetime with multiple time directions

This letter serves as the generalization of the work 2505.16436, where we investigated the quantum field theory in Klein space which has two time directions. We extend studies to the general spacetime $\mathbb{R}^{n,d-n}\,(n,d-n\geq2)$ in a similar manner: the ``length of time'' $q$ is regarded as the evolution direction of the system; additional modes beyond the plane wave modes are introduced in the canonical quantization. We show that this novel formulation is consistent with the analytical continuation of the results in Minkowski spacetime.

hep-th↗

arXiv2025

Starspots as the origin of ultrafast drifting radio bursts from an active M dwarf

Detecting coherent radio bursts from nearby M dwarfs provides opportunities for exploring their magnetic activity and interaction with orbiting exoplanets. However, it remains uncertain if the emission is related to flare-like activity similar to the Sun or magnetospheric process akin to magnetized planets. Using observations (1.0 - 1.5 GHz) taken by the Five-hundred-meter Aperture Spherical radio Telescope, we found a type of millisecond-scale radio bursts with exceptionally high frequency drift rates ($\sim 8\;\rm{GHz\;s^{-1}}$) from an active M dwarf, AD Leo. The ultrafast drift rates point to a source region with a notably low magnetic scale height ($<0.15\; r_\star$, $r_\star$ as the stellar radius), a feature not expected in a commonly assumed dipole-like global field but highly possible in localized strong-field structures, i.e. starspots. Our findings suggest that a concentrated magnetic field above starspots could be responsible for some of the most intense radio bursts from M dwarfs, supporting a solar-like electron acceleration mechanism.

astro-ph.SR↗

arXiv2025

Enigmatic centi-SFU and mSFU nonthermal radio transients detected in the middle corona

Decades of solar coronal observations have provided substantial evidence for accelerated particles in the corona. In most cases, the location of particle acceleration can be roughly identified by combining high spatial and temporal resolution data from multiple instruments across a broad frequency range. In almost all cases, these nonthermal particles are associated with quiescent active regions, flares, and coronal mass ejections (CMEs). Only recently, some evidence of the existence of nonthermal electrons at locations outside these well-accepted regions has been found. Here, we report for the first time multiple cases of transient nonthermal emissions, in the heliocentric range of $\sim 3-7R_\odot$, which do not have any obvious counterparts in other wavebands, like white-light and extreme ultra-violet. These detections were made possible by the regular availability of high dynamic range low-frequency radio images from the Owens Valley Radio Observatory's Long Wavelength Array. While earlier detections of nonthermal emissions at these high heliocentric distances often had comparable extensions in the plane-of-sky, they were primarily been associated with radio CMEs, unlike the cases reported here. Thus, these results add on to the evidence that the middle corona is extremely dynamic and contains a population of nonthermal electrons, which is only becoming visible with high dynamic range low-frequency radio images.

astro-ph.SR↗

arXiv2025

Possible First Detection of Gyroresonance Emission from a Coronal Mass Ejection in the Middle Corona

Routine measurements of the magnetic field of coronal mass ejections (CMEs) have been a key challenge in solar physics. Making such measurements is important both from a space weather perspective and for understanding the detailed evolution of the CME. In spite of significant efforts and multiple proposed methods, achieving this goal has not been possible to date. Here we report the first possible detection of gyroresonance emission from a CME. Assuming that the emission is happening at the third harmonic, we estimate that the magnetic field strength ranges from 7.9--5.6 G between 4.9-7.5 $R_\odot$. We also demonstrate that this high magnetic field is not the average magnetic field inside the CME, but most probably is related to small magnetic islands, which are also being observed more frequently with the availability of high-resolution and high-quality white-light images.

astro-ph.SR↗

arXiv2025

Generative artificial intelligence improves projections of climate extremes

Climate change is amplifying extreme events, posing escalating risks to biodiversity, human health, and food security. GCMs are essential for projecting future climate, yet their coarse resolution and high computational costs constrain their ability to represent extremes. Here, we introduce FuXi-CMIPAlign, a generative deep learning framework for downscaling CMIP outputs. The model integrates Flow Matching for generative modeling with domain adaptation via MMD loss to align feature distributions between training data and inference data, thereby mitigating input discrepancies and improving accuracy, stability, and generalization across emission scenarios. FuXi-CMIPAlign performs spatial, temporal, and multivariate downscaling, enabling more realistic simulation of compound extremes such as TCs.

physics.ao-ph↗

arXiv2025

GI-NAS: Boosting Gradient Inversion Attacks Through Adaptive Neural Architecture Search

Gradient Inversion Attacks invert the transmitted gradients in Federated Learning (FL) systems to reconstruct the sensitive data of local clients and have raised considerable privacy concerns. A majority of gradient inversion methods rely heavily on explicit prior knowledge (e.g., a well pre-trained generative model), which is often unavailable in realistic scenarios. This is because real-world client data distributions are often highly heterogeneous, domain-specific, and unavailable to attackers, making it impractical for attackers to obtain perfectly matched pre-trained models, which inevitably suffer from fundamental distribution shifts relative to target private data. To alleviate this issue, researchers have proposed to leverage the implicit prior knowledge of an over-parameterized network. However, they only utilize a fixed neural architecture for all the attack settings. This would hinder the adaptive use of implicit architectural priors and consequently limit the generalizability. In this paper, we further exploit such implicit prior knowledge by proposing Gradient Inversion via Neural Architecture Search (GI-NAS), which adaptively searches the network and captures the implicit priors behind neural architectures. Extensive experiments verify that our proposed GI-NAS can achieve superior attack performance compared to state-of-the-art gradient inversion methods, even under more practical settings with high-resolution images, large-sized batches, and advanced defense strategies. To the best of our knowledge, we are the first to successfully introduce NAS to the gradient inversion community. We believe that this work exposes critical vulnerabilities in real-world federated learning by demonstrating high-fidelity reconstruction of sensitive data without requiring domain-specific priors, forcing urgent reassessment of FL privacy safeguards.

cs.AI↗

arXiv2025

OpenRLHF: An Easy-to-use, Scalable and High-performance RLHF Framework

Large Language Models (LLMs) fine-tuned via Reinforcement Learning from Human Feedback (RLHF) and Reinforcement Learning with Verifiable Rewards (RLVR) significantly improve the alignment of human-AI values, further raising the upper bound of AI capabilities, particularly in reasoning-intensive, long-context Chain-of-Thought (CoT) tasks. However, existing frameworks commonly face challenges such as inference bottlenecks and complexity barriers, which restrict their accessibility to newcomers. To bridge this gap, we introduce \textbf{OpenRLHF}, a user-friendly, scalable, and easy-to-learn open-source RLHF framework built upon Ray, vLLM, DeepSpeed, and HuggingFace Transformers, featuring a simplified design, clear code structure, and comprehensive documentation to facilitate entry for researchers and practitioners. Experimental results show that OpenRLHF achieves superior training efficiency, with speedups ranging from 1.22x to 1.68x across different model sizes, compared to state-of-the-art frameworks. Additionally, it requires significantly fewer lines of code for implementation. OpenRLHF is publicly available at https://github.com/OpenRLHF/OpenRLHF, and has already been adopted by leading institutions to accelerate RLHF research and learning.

cs.AI↗

arXiv2025

CalibCLIP: Contextual Calibration of Dominant Semantics for Text-Driven Image Retrieval

Existing Visual Language Models (VLMs) suffer structural limitations where a few low contribution tokens may excessively capture global semantics, dominating the information aggregation process and suppressing the discriminative features in text-driven image retrieval tasks. To address this, we introduce \textbf{CalibCLIP}, a training-free method designed to calibrate the suppressive effect of dominant tokens. Specifically, in the visual space, we propose the Contrastive Visual Enhancer (CVE), which decouples visual features into target and low information regions. Subsequently, it identifies dominant tokens and dynamically suppresses their representations.In the textual space, we introduce the Discriminative Concept Calibrator (DCC), which aims to differentiate between general and discriminative concepts within the text query. By mitigating the challenges posed by generic concepts and improving the representations of discriminative concepts, DCC strengthens the differentiation among similar samples. Finally, extensive experiments demonstrate consistent improvements across seven benchmarks spanning three image retrieval tasks, underscoring the effectiveness of CalibCLIP. Code is available at: https://github.com/kangbin98/CalibCLIP

cs.CV↗