A5090328003- Drilling and Well Engineering
- Tunneling and Rock Mechanics
- Mineral Processing and Grinding
- Hydraulic Fracturing and Reservoir Analysis
- Fault Detection and Control Systems
- Advanced machining processes and optimization
- Iterative Learning Control Systems
- Adaptive Control of Nonlinear Systems
- Landslides and related hazards
- Oil and Gas Production Techniques
- Reservoir Engineering and Simulation Methods
- Advanced Memory and Neural Computing
- Neural Networks Stability and Synchronization
- Dynamics and Control of Mechanical Systems
- Geoscience and Mining Technology
- Piezoelectric Actuators and Control
- Cryospheric studies and observations
- Cloud Computing and Resource Management
- Genetics and Plant Breeding
- Genetic Mapping and Diversity in Plants and Animals
- Rock Mechanics and Modeling
- Advanced Computational Techniques and Applications
- Flood Risk Assessment and Management
- Machine Fault Diagnosis Techniques
- Advanced Statistical Process Monitoring
China University of Geosciences
2016-2025
Shanxi Agricultural University
2024-2025
Shandong Institute of Automation
2017-2024
Ministry of Education of the People's Republic of China
2020-2024
Swinburne University of Technology
2017-2019
We introduce our first-generation reasoning models, DeepSeek-R1-Zero and DeepSeek-R1. DeepSeek-R1-Zero, a model trained via large-scale reinforcement learning (RL) without supervised fine-tuning (SFT) as preliminary step, demonstrates remarkable capabilities. Through RL, naturally emerges with numerous powerful intriguing behaviors. However, it encounters challenges such poor readability, language mixing. To address these issues further enhance performance, we DeepSeek-R1, which incorporates...
We present DeepSeek-V3, a strong Mixture-of-Experts (MoE) language model with 671B total parameters 37B activated for each token. To achieve efficient inference and cost-effective training, DeepSeek-V3 adopts Multi-head Latent Attention (MLA) DeepSeekMoE architectures, which were thoroughly validated in DeepSeek-V2. Furthermore, pioneers an auxiliary-loss-free strategy load balancing sets multi-token prediction training objective stronger performance. pre-train on 14.8 trillion diverse...
This study presents a two-dimensional (2-D) repetitive control method to address the issues of periodic tracking and disturbance suppression in uncertain Takagi–Sugeno systems. The uncertainty are treated as an equivalent-input-disturbance (EID). However, conventional EID estimators typically suppress through high gain. Meanwhile, low-pass filter associated with causes certain degree phase lag. A proportional–integral (PI) is integrated estimator develop PI-EID structure improve estimation...
The stay-green (SG) trait enhances photosynthetic activity during the late grain-filling period, benefiting grain yield under drought and heat stresses. CH7034 is a wheat breeding line with SG. To clarify SG loci carried by obtain linked molecular markers, in this study, recombinant inbred (RIL) population derived from cross between non-SG SY95-71 was genotyped using Wheat17K single-nucleotide polymorphism (SNP) array, high-density genetic map covering 21 chromosomes consisting of 2159 SNP...
This paper is concerned with energy-to-peak state estimation on static neural networks (SNNs) interval time-varying delays. The objective to design suitable delay-dependent estimators such that the peak value of error can be minimized for all disturbances bounded energy. Note Lyapunov-Krasovskii functional (LKF) method plus proper integral inequalities provides a powerful tool in stability analysis and delayed NNs. main contribution this lies three points: 1) relationship between two based...
In this article, a systematic linear parameter-varying (LPV) model and gain-scheduled control methodology for drill-string systems are proposed to analyze dynamics suppress stick-slip vibrations, finally achieving efficient drilling. First, the changing length of drill string over entire drilling process is emphasized corresponding LPV presented by combining existing multi-degree-of-freedom (DOF) model, so as capture length-varying effect. Then, we construct generalized structure based on H...
Stick–slip vibrations are widespread and hazardous phenomena in drilling practice. The objective of this paper is to develop a simple general control strategy minimise such together with unknown disturbances. Taking into account the transmission delay torsional energy from surface downhole, neutral-type drill-string model established applied rather than lumped-parameter that cannot provide satisfactory modelling precision. Based on model, major contribution consists combination disturbance...
Geological drilling processes involve many variables, and their relationships dynamic characteristics are highly complicated. In the geological processes, changes in operating performance may be invisible under non-optimal conditions, while data distribution have significant deviations. The quality of collection is difficult to guarantee due underground measurement transmission environment, which increases uncertainty condition. Operators struggled manage monitoring for over decades. To...
The fault diagnosis during drilling is necessary to prevent the accidents develop more serious status. Data-driven methods have great advantages in nonlinear industrial process, however, problem of limited samples restricts its further application. This article proposes a data augmentation method based on synthetic generation and updating for with samples. First, generator trained generative adversarial nets (GAN), GAN improved by design parameter selection module, loss function model. Then,...
The compound directional drilling process, including slip and rotation, is the key to realizing long-hole in coal mines. First, a process control scheme proposed realize intelligent of process. An optimization method was designed improve efficiency. A robust controller for rate based on gain scheduling achieve stable during sliding deflection drilling. hybrid sensitivity rotary inclined stabilization results this study can provide theoretical basis
The leaf is not only the main site of photosynthesis, but also an important organ reflecting plant salt tolerance. Discovery salt-stress-responding genes in great significance for molecular improvement tolerance wheat varieties. In this study, transcriptome sequencing was conducted on leaves salt-tolerant germplasm CH7034 seedlings at 0, 1, 6, 24, and 48 h after NaCl treatment. Based weighted gene correlation network analysis differentially expressed (DEGs) under stress, 12 co-expression...
This paper is concerned with the stubborn state estimation of delayed neural networks that subject to a general class disturbances in measurements, including outliers and impulsive as its special cases. may be unbounded, irregular, assorted; therefore, they can hardly suppressed by existing identification-based approaches. In this paper, estimator constructed intentionally devising saturation scheme on injection output error. The embedded effectively resist influences from these measurement...