A5078564881- Reliability and Maintenance Optimization
- Gamma-ray bursts and supernovae
- Statistical Distribution Estimation and Applications
- Advanced Optical Network Technologies
- Bayesian Modeling and Causal Inference
- Advanced Statistical Process Monitoring
- Atomic and Subatomic Physics Research
- Human Pose and Action Recognition
- Machine Learning and Algorithms
- Semantic Web and Ontologies
- Business Process Modeling and Analysis
- Natural Language Processing Techniques
- Intelligent Tutoring Systems and Adaptive Learning
- Advanced Photonic Communication Systems
- Fault Detection and Control Systems
- 3D Shape Modeling and Analysis
- Anatomy and Medical Technology
- Optical Network Technologies
- semigroups and automata theory
- Particle Detector Development and Performance
- Pulsars and Gravitational Waves Research
- Particle physics theoretical and experimental studies
- Machine Fault Diagnosis Techniques
Shanghai Jiao Tong University
2023-2024
China Automotive Engineering Research Institute
2023
Current approaches for training Process Reward Models (PRMs) often involve breaking down responses into multiple reasoning steps using rule-based techniques, such as predefined placeholder tokens or setting the step's length a fixed size. These overlook fact that specific words do not typically mark true decision points in text. To address this, we propose AdaptiveStep, method divides based on model's confidence predicting next word. This division provides more decision-making information at...
Most existing tests in the literature for model checking do not work high dimension settings due to challenges arising from "curse of dimensionality", or dependencies on normality parameter estimators. To address these challenges, we proposed a new goodness fit test based random projections generalized linear models, when covariates may substantially exceed sample size. The only require convergence rate estimators derive limiting distribution. growing is allowed be exponential order relation...
With the advancement of computer and medical imaging technologies, a number high-resolution, voxel-based, full-body human anatomical models have been developed for education, industrial design, physics simulation studies. However, these are limited in many applications because they often only an upstanding posture.To quickly develop multi-pose different applications. A semi-automatic framework voxel deformation is proposed study.This paper describes pose based on three-dimensional (3D)...
Fault feature extraction is extremely important for machine condition monitoring (MCM). Recently proposed learning-based optimized weights have been proven to own fully physical interpretability learn informative fault features MCM. Nevertheless, learning algorithms based still can not be intuitively understood, which restricts the utilization of optimization In this paper, we first propose an equivalent and easier computation approach calculate aforementioned interpretable weights, i.e.,...