A5101527887- Coding theory and cryptography
- Algebraic Geometry and Number Theory
- Privacy-Preserving Technologies in Data
- Limits and Structures in Graph Theory
- Hepatitis C virus research
- Analytic Number Theory Research
- Neural Networks and Applications
- Monoclonal and Polyclonal Antibodies Research
- Finite Group Theory Research
- Cloud Computing and Resource Management
- Quantum Computing Algorithms and Architecture
- Anomaly Detection Techniques and Applications
- Machine Learning in Materials Science
- Building Energy and Comfort Optimization
- IoT and Edge/Fog Computing
- Computational Drug Discovery Methods
- Gait Recognition and Analysis
- Domain Adaptation and Few-Shot Learning
- Machine Learning and Algorithms
- Human Motion and Animation
- Stock Market Forecasting Methods
- Polynomial and algebraic computation
- Commutative Algebra and Its Applications
- Markov Chains and Monte Carlo Methods
- Traffic Prediction and Management Techniques
China University of Petroleum, East China
2023-2025
Nagase ChemteX (Japan)
2024
Xi'an Conservatory of Music
2023
Beijing University of Posts and Telecommunications
2021-2023
Beijing Jiaotong University
2023
University of Shanghai for Science and Technology
2022
China University of Petroleum, Beijing
2021
University of Waterloo
2004-2017
Central China Normal University
2013
Naval Aeronautical and Astronautical University
2010
Drug-Drug interactions (DDIs) are a challenging problem in drug research. Drug combination therapy is an effective solution to treat diseases, but it can also cause serious side effects. Therefore, DDIs prediction critical pharmacology. Recently, researchers have been using deep learning techniques predict DDIs. However, these methods only consider single information of the and shortcomings robustness scalability.In this paper, we propose multi-type feature fusion based on graph neural...
Heating, ventilation, and air conditioning (HVAC) energy consumption now accounts for a major portion of use buildings. Therefore, finding the optimal energy-saving control strategy HVAC systems to optimize has become crucial in realizing savings, emission reductions, green Traditional methods parameter require complex physical model calculation; continuous coupled parameters are handled poorly. Developing deep reinforcement learning (DRL) provides new ideas optimization. Herein, DRL-based...
Federated learning can collaboratively train AI models while protecting data privacy. In practical industry environment, non-independent and identically distributed (Non-IID) characteristics of affect the effectiveness federated learning. Personalized help resolve this, but it cannot adapt to unknown data. addition, applications also call for trusted training environment remain stable when there are security threats. this article, we propose a credible self-learning (CFSL), based on idea...
Federated learning is a distributed machine approach that achieves collaborative training while protecting data privacy. However, in scenarios, the operational of industrial equipment dynamic and non-independently identically (non-IID). This situation leads to poor performance federated algorithms anomaly detection tasks. Personalized viable solution non-IID problem, but it not effective responding environmental changes. Implementing directed updates model, thereby effectively maintaining...
Predicting organic reaction feasibility and robustness against environmental factors is challenging. We address this issue by integrating high throughput experimentation (HTE) Bayesian deep learning. Diverging from existing HTE studies focused on niche chemical spaces, in work, our in-house platform conducted 11,669 distinct acid amine coupling reactions 156 working hours, yielding the most extensive single dataset at a volumetric scale for industrial delivery. Our neural network model...
With the development of internet-related technology, more and consumers are trading on second-hand platforms. However, due to uncertainty products, there a lot transaction disputes platform. It is important efficiently solve Based comment data social media, this paper systematically studies public opinion towards online dispute resolution Xianyu, China’s largest goods-trading platform, using word-frequency analysis, theme analysis sentiment analysis. The study found that mainly focus return...
Abstract We axiomatize the main properties of classical Erdös-Kac Theorem in order to apply it a general context. provide applications cases number fields, function and geometrically irreducible varieties over finite field.
Abstract Let 𝔽 q [ t ] denote the polynomial ring over finite field . We employ Wooley's new efficient congruencing method to prove certain multidimensional Vinogradov-type estimates in ]. These results allow us apply a variant of circle obtain asymptotic formulas for system connected problem about linear spaces lying on hypersurfaces defined
Let ${\mathbb F}_q[t]$ be the polynomial ring over finite field F}_q$, and let G_{N}}$ subset of containing all polynomials degree strictly less than $N$. Define $D(N)$ to maximal cardinality
Let $\mathbb{J}_q^k[t]$ denote the additive closure of set $k$th powers in polynomial ring $\mathbb{F}_q[t]$, defined over finite field $\mathbb{F}_q$ having $q$ elements. We show that when $s\ge k+1$ and $q \ge k^{2k+2}$, then every is sum at most $s$ polynomials from $\mathbb{F}_q[t]$. When $k$ large $s (\frac{4}{3}+o(1)) k\log k$, same conclusion holds without restriction on $q$. Refinements are offered depend characteristic $\mathbb{F}_q$.
Let A = 𝔽 q [T] be the polynomial ring over finite field , k (T) rational function field, and K a extension of k. ϕ Drinfeld A-module rank r. For place 𝔓 good reduction, write [Formula: see text], where text] is valuation its maximal ideal. P 𝔓, (X) characteristic Frobenius automorphism acting on Tate module ϕ. χ (𝔓) (1), let ν(χ (𝔓)) number distinct primes dividing (𝔓). If 2 with we prove that there exists normal distribution for quantity r ≥ 3, show same result holds under open image...
Investment as an important issue in daily life is accompanied by the occurrence of various financial assets, such stocks, bonds, and mutual funds. However, risk tolerances vary across individuals. Individual investors have to select corresponding personalized investment portfolios satisfy their own needs. Moreover, it difficult for ordinary people a portfolio themselves, too expensive inefficient look professional consultation. Therefore, objective this research propose recommendation model,...
Federated learning allows for collaborative training of distributed deep networks, tackling the issue data silos while ensuring privacy. However, in setting industrial equipment, collected is dynamic, non-independent, and identically distributed(i.e., Non-IID), which negatively impacts performance federated learning. While personalized solution can address Non-IID problem, it falls short effectively handling dynamic environment poses a significant challenge to model stability. Model...
Let 𝔽 q [t] denote the polynomial ring over finite field , and let [Formula: see text] subset of containing all polynomials degree strictly less than N. For non-zero elements r 1 …, s satisfying + ⋯ = 0, maximal cardinality a set which contains no non-trivial solution x 0 with i ∈ A (1 ≤ s). We prove that text].
Abstract Recently, graph convolutional networks have shown excellent results in skeleton‐based action recognition. This paper presents a multi‐stage part‐aware network for the problems of model over complication, parameter redundancy and lack long‐dependence feature information. The structure this has multi‐stream input two‐stream output, which can greatly reduce complexity improve accuracy without losing sequence two branches same backbone, includes 6 multi‐order extraction blocks 3...
Abstract We axiomatize the main properties of classical Turán Theorem in order to apply it a general context. provide applications cases number fields, function and geometrically irreducible varieties over finite field.
Air conditioning contributes a high percentage of energy consumption over the world. The efficient prediction can help to reduce consumption. Traditionally, multidimensional air data could only be processed sequentially for each dimension, thus resulting in inefficient feature extraction. Furthermore, due reasons such as implicit correlations between hyperparameters, automatic hyperparameter optimization (HPO) approaches not easily achieved. In this paper, we propose an auto-optimization...
We present a scheme of quantum repeater for infrared communication system. It is to connect string entangled pairs photons. The can be applied protocol. used long distance and realized with current or near future experimental technology.