What is he best known for?
The fields of study he is best known for:
- Artificial intelligence
- Machine learning
- Algorithm
Ji Liu spends much of his time researching Algorithm, Asynchronous communication, Speedup, Stochastic gradient descent and Coordinate descent.
His Algorithm research incorporates themes from Stochastic approximation, Saddle, Acceleration, Convex optimization and Pattern recognition.
His Convex optimization study incorporates themes from Artificial neural network, Artificial intelligence, Bounded function and Shared memory.
His work deals with themes such as Event, Machine learning and Data mining, which intersect with Artificial intelligence.
In his study, Support vector machine, Stochastic optimization, Convolutional neural network and Computation is inextricably linked to Bottleneck, which falls within the broad field of Stochastic gradient descent.
His Coordinate descent study improves the overall literature in Mathematical optimization.
His most cited work include:
- Tensor completion for estimating missing values in visual data (1038 citations)
- Sparse reconstruction cost for abnormal event detection (552 citations)
- Can Decentralized Algorithms Outperform Centralized Algorithms? A Case Study for Decentralized Parallel Stochastic Gradient Descent (365 citations)
What are the main themes of his work throughout his whole career to date?
His scientific interests lie mostly in Artificial intelligence, Algorithm, Mathematical optimization, Machine learning and Reinforcement learning.
His research ties Pattern recognition and Artificial intelligence together.
When carried out as part of a general Algorithm research project, his work on Coordinate descent is frequently linked to work in Asynchronous communication, therefore connecting diverse disciplines of study.
The various areas that he examines in his Mathematical optimization study include Regularization, Bounded function and Proximal Gradient Methods.
Ji Liu has included themes like Stochastic optimization and Variance reduction in his Reinforcement learning study.
His study in Rate of convergence is interdisciplinary in nature, drawing from both Stochastic gradient descent and Server.
He most often published in these fields:
- Artificial intelligence (36.89%)
- Algorithm (21.36%)
- Mathematical optimization (20.39%)
What were the highlights of his more recent work (between 2020-2021)?
- Artificial intelligence (36.89%)
- Regret (4.37%)
- Key (6.31%)
In recent papers he was focusing on the following fields of study:
His main research concerns Artificial intelligence, Regret, Key, Bottleneck and Scalability.
Artificial intelligence is frequently linked to Function in his study.
His Function research is multidisciplinary, incorporating elements of Stochastic gradient descent, Variance reduction and Reinforcement learning.
His Regret research includes elements of World Wide Web, Social network, Monte Carlo tree search and Federated learning.
His studies examine the connections between Bottleneck and genetics, as well as such issues in Data mining, with regards to Embedding.
His work focuses on many connections between Scalability and other disciplines, such as System model, that overlap with his field of interest in Feature.
Between 2020 and 2021, his most popular works were:
- Central Server Free Federated Learning over Single-sided Trust Social Networks (19 citations)
- Towards Statistically Provable Geometric 3D Human Pose Recovery (2 citations)
- Streaming Probabilistic Deep Tensor Factorization (1 citations)
This overview was generated by a machine learning system which analysed the scientist’s
body of work. If you have any feedback, you can
contact us
here.
