B. Zheng
YOU?
Author Swipe
View article: Characterization of SiPMs at 40 K for neutrino coherent detection based on pure CsI
Characterization of SiPMs at 40 K for neutrino coherent detection based on pure CsI Open
View article: MeSH: Memory-as-State-Highways for Recursive Transformers
MeSH: Memory-as-State-Highways for Recursive Transformers Open
Recursive transformers reuse parameters and iterate over hidden states multiple times, decoupling compute depth from parameter depth. However, under matched compute, recursive models with fewer parameters often lag behind non-recursive cou…
View article: A nomogram using transition zone PSA density for detecting clinically significant prostate cancer in PI-RADS 3 lesions
A nomogram using transition zone PSA density for detecting clinically significant prostate cancer in PI-RADS 3 lesions Open
View article: Quantum-Dot-Based Molecularly Imprinted Hydrogel for Rapid Detection of Homocysteine
Quantum-Dot-Based Molecularly Imprinted Hydrogel for Rapid Detection of Homocysteine Open
Elevated levels of homocysteine (Hcy) are associated with various pathological conditions including atherosclerosis, hypertension, and cardiovascular diseases. In this work, quantum-dot-based molecularly imprinted hydrogels (QD@MIHs) were …
View article: Novel insights into the transcriptomic changes of the hypothalamus in broilers exposed to high-density stocking
Novel insights into the transcriptomic changes of the hypothalamus in broilers exposed to high-density stocking Open
Background The prevalence of high stocking density ( HD ) in the broiler industry has been significantly increased, exerting profound implications for the physiology, behavior, and welfare of chickens, particularly concerning the regulatio…
View article: The preparation method and application of aluminum alloy flux-cored wire for wire arc additive manufacturing
The preparation method and application of aluminum alloy flux-cored wire for wire arc additive manufacturing Open
Ceramic phase modification is a effective method to improve the performance of wire arc additive manufacturing (WAAM) aluminum alloy component. In this paper, a preparation method of ceramic aluminum alloy flux-cored wire was developed. An…
View article: Masked Self-distilled Transducer-based Keyword Spotting with Semi-autoregressive Decoding
Masked Self-distilled Transducer-based Keyword Spotting with Semi-autoregressive Decoding Open
RNN-T-based keyword spotting (KWS) with autoregressive decoding~(AR) has gained attention due to its streaming architecture and superior performance. However, the simplicity of the prediction network in RNN-T poses an overfitting issue, es…
View article: Pangu Pro MoE: Mixture of Grouped Experts for Efficient Sparsity
Pangu Pro MoE: Mixture of Grouped Experts for Efficient Sparsity Open
The surgence of Mixture of Experts (MoE) in Large Language Models promises a small price of execution cost for a much larger model parameter count and learning capacity, because only a small fraction of parameters are activated for each in…
View article: Optimal and Sustainable Scheduling of Integrated Energy System Coupled with CCS-P2G and Waste-to-Energy Under the “Green-Carbon” Offset Mechanism
Optimal and Sustainable Scheduling of Integrated Energy System Coupled with CCS-P2G and Waste-to-Energy Under the “Green-Carbon” Offset Mechanism Open
Waste-to-energy (WTE) is considered the most promising method for municipal solid waste treatment. An integrated energy system (IES) with carbon capture systems (CCS) and power-to-gas (P2G) can reduce carbon emissions. The incorporation of…
View article: USB: A Comprehensive and Unified Safety Evaluation Benchmark for Multimodal Large Language Models
USB: A Comprehensive and Unified Safety Evaluation Benchmark for Multimodal Large Language Models Open
Despite their remarkable achievements and widespread adoption, Multimodal Large Language Models (MLLMs) have revealed significant security vulnerabilities, highlighting the urgent need for robust safety evaluation benchmarks. Existing MLLM…
View article: Beyond Safe Answers: A Benchmark for Evaluating True Risk Awareness in Large Reasoning Models
Beyond Safe Answers: A Benchmark for Evaluating True Risk Awareness in Large Reasoning Models Open
Despite the remarkable proficiency of \textit{Large Reasoning Models} (LRMs) in handling complex reasoning tasks, their reliability in safety-critical scenarios remains uncertain. Existing evaluations primarily assess response-level safety…
View article: Think-J: Learning to Think for Generative LLM-as-a-Judge
Think-J: Learning to Think for Generative LLM-as-a-Judge Open
LLM-as-a-Judge refers to the automatic modeling of preferences for responses generated by Large Language Models (LLMs), which is of significant importance for both LLM evaluation and reward modeling. Although generative LLMs have made subs…
View article: TranSUN: A Preemptive Paradigm to Eradicate Retransformation Bias Intrinsically from Regression Models in Recommender Systems
TranSUN: A Preemptive Paradigm to Eradicate Retransformation Bias Intrinsically from Regression Models in Recommender Systems Open
Regression models are crucial in recommender systems. However, retransformation bias problem has been conspicuously neglected within the community. While many works in other fields have devised effective bias correction methods, all of the…
View article: Fine Classification of Vegetation Under Complex Surface Cover Conditions with Hyperspectral and High-Spatial Resolution: A Case Study of the Xisha Area, Chongming District, Shanghai
Fine Classification of Vegetation Under Complex Surface Cover Conditions with Hyperspectral and High-Spatial Resolution: A Case Study of the Xisha Area, Chongming District, Shanghai Open
Since both diversity and similarity exist among different vegetation types and since differences and similarities are reflected mainly in geometric morphology and in physical and chemical characteristics, the feedback signals of remote sen…
View article: Equilibrate RLHF: Towards Balancing Helpfulness-Safety Trade-off in Large Language Models
Equilibrate RLHF: Towards Balancing Helpfulness-Safety Trade-off in Large Language Models Open
Fine-tuning large language models (LLMs) based on human preferences, commonly achieved through reinforcement learning from human feedback (RLHF), has been effective in improving their performance. However, maintaining LLM safety throughout…
View article: "See the World, Discover Knowledge": A Chinese Factuality Evaluation for Large Vision Language Models
"See the World, Discover Knowledge": A Chinese Factuality Evaluation for Large Vision Language Models Open
The evaluation of factual accuracy in large vision language models (LVLMs) has lagged behind their rapid development, making it challenging to fully reflect these models' knowledge capacity and reliability. In this paper, we introduce the …
View article: Unlocking Scaling Law in Industrial Recommendation Systems with a Three-step Paradigm based Large User Model
Unlocking Scaling Law in Industrial Recommendation Systems with a Three-step Paradigm based Large User Model Open
Recent advancements in autoregressive Large Language Models (LLMs) have achieved significant milestones, largely attributed to their scalability, often referred to as the "scaling law". Inspired by these achievements, there has been a grow…
View article: AIGuard: A Benchmark and Lightweight Detection for E-commerce AIGC Risks
AIGuard: A Benchmark and Lightweight Detection for E-commerce AIGC Risks Open
View article: Aircraft-Cargo Separation Rate-Driven Ground Risk–Cost Optimization for Urban Logistics UAVs
Aircraft-Cargo Separation Rate-Driven Ground Risk–Cost Optimization for Urban Logistics UAVs Open
The surge in demand for urban logistics has positioned unmanned aerial vehicles (UAVs) as a vital solution for improving last mile delivery efficiency in complex urban environments. To overcome the ground safety bottleneck in the large-sca…
View article: Chinese SafetyQA: A Safety Short-form Factuality Benchmark for Large Language Models
Chinese SafetyQA: A Safety Short-form Factuality Benchmark for Large Language Models Open
View article: Can Large Language Models Detect Errors in Long Chain-of-Thought Reasoning?
Can Large Language Models Detect Errors in Long Chain-of-Thought Reasoning? Open
View article: Predicting the Risk of Incident Cardiovascular Disease Using Retinal Vessel Caliber Measured by Artificial Intelligence in a Chinese Community-Based Population
Predicting the Risk of Incident Cardiovascular Disease Using Retinal Vessel Caliber Measured by Artificial Intelligence in a Chinese Community-Based Population Open
View article: VC4VG: Optimizing Video Captions for Text-to-Video Generation
VC4VG: Optimizing Video Captions for Text-to-Video Generation Open
View article: Chinese SimpleQA: A Chinese Factuality Evaluation for Large Language Models
Chinese SimpleQA: A Chinese Factuality Evaluation for Large Language Models Open
View article: Qwen2.5 Technical Report
Qwen2.5 Technical Report Open
In this report, we introduce Qwen2.5, a comprehensive series of large language models (LLMs) designed to meet diverse needs. Compared to previous iterations, Qwen 2.5 has been significantly improved during both the pre-training and post-tr…
View article: LLaVA-UHD v2: an MLLM Integrating High-Resolution Semantic Pyramid via Hierarchical Window Transformer
LLaVA-UHD v2: an MLLM Integrating High-Resolution Semantic Pyramid via Hierarchical Window Transformer Open
Vision transformers (ViTs) are widely employed in multimodal large language models (MLLMs) for visual encoding. However, they exhibit inferior performance on tasks regarding fine-grained visual perception. We attribute this to the limitati…
View article: Chinese SafetyQA: A Safety Short-form Factuality Benchmark for Large Language Models
Chinese SafetyQA: A Safety Short-form Factuality Benchmark for Large Language Models Open
With the rapid advancement of Large Language Models (LLMs), significant safety concerns have emerged. Fundamentally, the safety of large language models is closely linked to the accuracy, comprehensiveness, and clarity of their understandi…
View article: Efficacy and safety of a patent hemostatic device (PHD) with quantitative pressure for radial artery hemostasis: a first-in-human feasibility study
Efficacy and safety of a patent hemostatic device (PHD) with quantitative pressure for radial artery hemostasis: a first-in-human feasibility study Open
Objectives: Hemostatic devices are commonly used after coronary angiography (CAG) or percutaneous coronary intervention (PCI) via the transradial approach. However, radial artery occlusion (RAO) and bleeding still occur due to uncertaintie…
View article: Chinese SimpleQA: A Chinese Factuality Evaluation for Large Language Models
Chinese SimpleQA: A Chinese Factuality Evaluation for Large Language Models Open
New LLM evaluation benchmarks are important to align with the rapid development of Large Language Models (LLMs). In this work, we present Chinese SimpleQA, the first comprehensive Chinese benchmark to evaluate the factuality ability of lan…
View article: Defect intelligent recognition of membrane product based on deep learning
Defect intelligent recognition of membrane product based on deep learning Open
Defect detection plays a crucial role in the manufacturing industry, ensuring the quality of industrial products. Despite advancements in this field, current defect detection methods face two primary challenges: (1) extracting visually sim…