A5078600050- Stochastic Gradient Optimization Techniques
- Domain Adaptation and Few-Shot Learning
- Advanced Neural Network Applications
- advanced mathematical theories
- Spectral Theory in Mathematical Physics
- Error Correcting Code Techniques
- Neural Networks and Applications
- Mathematical Dynamics and Fractals
- Advanced Mathematical Modeling in Engineering
- Web Data Mining and Analysis
- Advanced Wireless Communication Techniques
- Sparse and Compressive Sensing Techniques
- Stochastic processes and financial applications
- Complex Network Analysis Techniques
- Forensic and Genetic Research
- DNA and Biological Computing
- Genetic diversity and population structure
- Data Management and Algorithms
- Machine Learning and Algorithms
- Model Reduction and Neural Networks
- Nonlinear Partial Differential Equations
- Privacy-Preserving Technologies in Data
- Advanced Bandit Algorithms Research
- Machine Learning and ELM
- Multimodal Machine Learning Applications
Chinese Academy of Sciences
2015-2024
Academy of Mathematics and Systems Science
2015-2024
Shenyang Ligong University
2024
University of Chinese Academy of Sciences
2018-2024
Qingdao Institute of Bioenergy and Bioprocess Technology
2019-2024
Ningxia University
2024
Mudanjiang Normal University
2024
Xinjiang Normal University
2021-2023
Huawei Technologies (China)
2021
Microsoft Research (United Kingdom)
2021
Tuning the geometric and electronic structure of single-metal-atom catalysts via simultaneous presence metal nanoparticles or nanoclusters (NCs) offers an alternative avenue to improving their catalytic performance. Herein, we demonstrate that coexistence Fe NCs in proximity single atoms on N-doped porous carbon can significantly improve performance aerobic oxidation primary amines imines using air as oxidant. A broad spectrum aromatic, heterocyclic, aliphatic was efficiently selectively...
A bstract Deep learning methods have been increasingly adopted to study jets in particle physics. Since symmetry-preserving behavior has shown be an important factor for improving the performance of deep many applications, Lorentz group equivariance — a fundamental spacetime symmetry elementary particles recently incorporated into model jet tagging. However, design is computationally costly due analytic construction high-order tensors. In this article, we introduce LorentzNet, new The...
This paper proposes a new method for computing page importance, referred to as BrowseRank. The conventional approach compute importance is exploit the link graph of web and build model based on that graph. For instance, PageRank such an algorithm, which employs discrete-time Markov process model. Unfortunately, might be incomplete inaccurate with respect data determining because links can easily added deleted by content creators. In this paper, we propose using 'user browsing graph' created...
This paper is concerned with rank aggregation, the task of combining ranking results individual rankers at meta-search. Previously, aggregation was performed mainly by means unsupervised learning. To further enhance accuracies, we propose employing supervised learning to perform task, using labeled data. We refer approach as Supervised Rank Aggregation. set up a general framework for conducting Aggregation, in which formalized an optimization minimizes disagreements between and As case...
With the fast development of deep learning, it has become common to learn big neural networks using massive training data. Asynchronous Stochastic Gradient Descent (ASGD) is widely adopted fulfill this task for its efficiency, which is, however, known suffer from problem delayed gradients. That when a local worker adds gradient global model, model may have been updated by other workers and becomes "delayed". We propose novel technology compensate delay, so as make optimization behavior ASGD...
A heterogeneous nanocomposite of Fe–Fe<sub>3</sub>C nanoparticles and Fe–N<sub>x</sub> sites on N-doped porous carbon allows for efficient synthesis quinolines quinazolinones <italic>via</italic> oxidative coupling amines aldehydes in aq. solution using H<sub>2</sub>O<sub>2</sub> as the oxidant.
The utilization and development of biomass resources is an efficient solution to mitigate the fossil energy crisis. Based on advantages mild reaction conditions, rapid reaction, high conversion, synthesis 2,5-furandicarboxylic acid (FDCA) by electrocatalytic oxidation 5-hydroxymethylfurfural (HMFOR) has attracted considerable attention. This review will summarize recent advances HMFOR FDCA, including pathway mechanism, as well catalytic performance various heterogeneous electrocatalysts....
Decision tree (and its extensions such as Gradient Boosting Trees and Random Forest) is a widely used machine learning algorithm, due to practical effectiveness model interpretability. With the emergence of big data, there an increasing need parallelize training process decision tree. However, most existing attempts along this line suffer from high communication costs. In paper, we propose new called \emph{Parallel Voting Tree (PV-Tree)}, tackle challenge. After partitioning data onto number...
Aerobic oxidative cross-dehydrogenative coupling represents one of the most straightforward and atom-economic methods for construction C–C C–X (X = N, O, S, or P) bonds, especially when environmentally friendly air is used as oxidant. Herein, we report development an inexpensive, stable, highly dispersed ultrafine Ni2P nanoparticles with narrow size distribution supported on N,P-codoped biomass-derived porous carbon. The as-prepared catalyst active stable synthesis pharmaceutically important...
We herein report the fabrication of a bifunctional iron nanocomposite catalyst, in which two catalytically active sites Fe–Nx and Fe phosphate, as oxidation Lewis acid sites, were simultaneously integrated into hierarchical N,P-dual doped porous carbon. As it exhibited high efficiency for direct oxidative cleavage alkenes ketones or their 1,2-diketones with broad substrate scope functional group tolerance using TBHP oxidant water under mild reaction conditions. Furthermore, could be easily...
By making use of the reflected α-stable process on a closed domain ℝ n and its killed subprocess part domain, in this paper we study boundary value problem for Schrödinger type equation fractional Laplacian. The condition is imposed partly follow Dirichlet Neuman condition. We obtain existence uniqueness resutls. solution expressed as functional process.
With the rapid development of Large Language Models (LLMs), numerous Reinforcement Learning from Human Feedback (RLHF) algorithms have been introduced to improve model safety and alignment with human preferences. These can be divided into two main frameworks based on whether they require an explicit reward (or value) function for training: actor-critic-based Proximal Policy Optimization (PPO) alignment-based Direct Preference (DPO). The mismatch between DPO PPO, such as DPO's use a...
As a powerful all-weather Earth observation tool, synthetic aperture radar (SAR) remote sensing enables critical military reconnaissance, maritime surveillance, and infrastructure monitoring. Although Vision language models (VLMs) have made remarkable progress in natural processing image understanding, their applications remain limited professional domains due to insufficient domain expertise. This paper innovatively proposes the first large-scale multimodal dialogue dataset for SAR images,...
This paper presents a theoretical framework for ranking, and demonstrates how to perform generalization analysis of listwise ranking algorithms using the framework. Many learning-to-rank have been proposed in recent years. Among them, approach has shown higher empirical performance when compared other approaches. However, there is no study on as far we know. In this paper, propose which can naturally describe various algorithms. With framework, prove theorem gives bound algorithm, basis...
Abstract Herein, we report highly chemoselective hydrogenation of α,β‐unsaturated carbonyls to saturated catalyzed by cobalt nanoparticles supported on the biomass‐derived carbon from bamboo shoots with molecular hydrogen in water, which is first prototype using a heterogeneous non‐noble metal catalyst for such organic transformation as far know. The optimal nanocatalyst, CoO x @NC‐800, manifested remarkable activity and selectivity C=C under mild conditions. A broad set α,β‐aromatic...