Sheng Li

Generalization Bounds for Equivariant Networks on Markov Data

Hui Li, Zhiguo Wang, Sheng Li, Bohui Chen

Abstract. Equivariant neural networks play a pivotal role in analyzing datasets with symmetry properties, particularly in complex data structures. However, integrating equivariance with Markov properties presents notable challenges due to the inherent dependencies within such data. Previous research has primarily concentrated on …

cn (code pays fourni par la source)

Noisy probing dose facilitated dose prediction for pencil beam scanning proton therapy: Physics enhances generalizability

Lian Zhang, Jason Holmes, Zhengliang Liu, Hongying Feng et autres

BACKGROUND: Accurate and efficient dose calculation is essential for online adaptive planning in proton therapy. Deep learning (DL) has shown promising dose prediction results for pencil beam scanning proton therapy (PBSPT) in recent years, but existing DL-based dose prediction methods still suffer …

us, de (code pays fourni par la source)

A machine learning model for predicting consumer preferences of melon flavor and sensory attributes

Qianqian Jia, Haojie Sha, Yao Tian, Wen‐Lu Bi et autres

Melon flavor, primarily driven by volatile organic compounds (VOCs), significantly influences consumer preferences. This study analyzed the volatile compounds of 74 melon varieties using GC-IMS. A total of 56 compounds were identified, with esters constituting the highest proportion, followed by ketones, aldehydes, …

Characterizing prognostic and immunological traits of kidney renal clear cell carcinoma via mitochondria-associated membranes and identifying DNM1L as a potential therapeutic target using machine learning

Sheng Li, Jinkang Lin, Fucun Zheng, Xiaoqiang Liu et autres

BACKGROUND: Mitochondrial-associated membranes (MAMs) participate in cellular metabolism, calcium signaling, and cancer reprogramming, but their role in kidney renal clear cell carcinoma (KIRC) remains unclear. METHODS: Clinical/transcriptomic data of KIRC and previously reported MAMs-related genes were obtained from TCGA, E-MTAB-1980, and GSE29609. …

cn (code pays fourni par la source)

COMPASS: Enhancing Agent Long-Horizon Reasoning with Evolving Context

Guoyang Wan, Mingyang Ling, Xiaoqi Ren, Rujun Han et autres

Long-horizon tasks that require sustained reasoning and multiple tool interactions remain challenging for LLM agents: small errors compound across steps, and even state-of-the-art models often hallucinate or lose coherence.We identify context management as the central bottleneck-extended histories cause agents to overlook critical …

us (code pays fourni par la source)

Digital Twin System for Urban Building Energy Efficiency: A Case Study in Shenzhen

Hang Su, Songbo Zhang, Tianyu Li, Biao Yang et autres

Digital twin technology is increasingly recognized as a cornerstone for sustainable digital cities, enabling real-time integration between physical environments and their virtual counterparts. This study develops a Unity-based digital twin system that integrates Internet of Things (IoT) devices with occupancy-centric control (OCC) …

cn (code pays fourni par la source)

Large language models management of complex medication regimens: a case-based evaluation

Aaron M. Chase, Amoreena Most, Shaochen Xu, Erin Barreto et autres

Background: Large language models (LLMs) have shown the ability to diagnose complex medical cases, but only limited studies have evaluated the performance of LLMs in the development of evidence-based treatment plans. The purpose of this evaluation was to test four LLMs on …

us (code pays fourni par la source)

Enhancing solid-liquid suspension in high-solids vanadium-bearing smelting slag leaching reactor via eccentric double-impeller coupling generated asymmetric flow

Jiang Huang, Yimin Zhang, Qiushi Zheng, Nannan Xue et autres

cn (code pays fourni par la source)

Fracture behavior of the wave-shaped flaw in rock-like material under direct shear using XFEM

Sheng Li, Suifeng Wang, Hongchang Zhou, Tao Xu et autres

cn, jp (code pays fourni par la source)

SimuFreeMark: A Noise-Simulation-Free Robust Watermarking Against Image Editing

Yichao Tang, Mingyang Li, Di Miao, Sheng Li et autres

The advancement of artificial intelligence generated content (AIGC) has created a pressing need for robust image watermarking that can withstand both conventional signal processing and novel semantic editing attacks. Current deep learning-based methods rely on training with hand-crafted noise simulation layers, which …

Backdoor in Seconds: Unlocking Vulnerabilities in Large Pre-trained Models via Model Editing

Dongliang Guo, Mengxuan Hu, Zihan Guan, Junfeng Guo et autres

Large pre-trained models have achieved notable success across a range of downstream tasks. However, recent research shows that a type of adversarial attack (i.e., backdoor attack) can manipulate the behavior of machine learning models through contaminating their training dataset, posing significant threat …

us (code pays fourni par la source)

PharmacyGPT: exploration of artificial intelligence for medication management in the intensive care unit

Zhengliang Liu, Shaochen Xu, Zihao Wu, Brian Murray et autres

BACKGROUND: The purpose of this evaluation was to develop a novel framework for large language model (LLMs) iterative prompt optimization for the specialized domain of medication decision-making in the intensive care unit (ICU). This serves as a first step towards the use …

us (code pays fourni par la source)