What is he best known for?
The fields of study he is best known for:
- Operating system
- Central processing unit
- Parallel computing
His scientific interests lie mostly in Parallel computing, Embedded system, Interleaved memory, Memory management and Shared memory.
He frequently studies issues relating to Reliability and Parallel computing.
His work in Reliability tackles topics such as Node which are related to areas like Supercomputer.
His Embedded system study integrates concerns from other disciplines, such as Virtual memory, Fault tolerance, Interconnection and Memory controller.
In most of his Memory management studies, his work intersects topics such as Memory map.
Mattan Erez combines subjects such as Registered memory, Conventional memory, Physical address and Cache-only memory architecture with his study of Non-uniform memory access.
His most cited work include:
- Sequoia: programming the memory hierarchy (439 citations)
- Merrimac: Supercomputing with Streams (302 citations)
- Addressing failures in exascale computing (277 citations)
What are the main themes of his work throughout his whole career to date?
Mattan Erez mainly focuses on Parallel computing, Embedded system, Memory bandwidth, Memory management and Fault tolerance.
Many of his studies involve connections with topics such as Bandwidth and Parallel computing.
The study incorporates disciplines such as Dram, Random access memory, Interconnection and Redundant array of independent memory in addition to Embedded system.
His research integrates issues of Deep learning, Artificial intelligence and Central processing unit in his study of Memory bandwidth.
Mattan Erez focuses mostly in the field of Fault tolerance, narrowing it down to topics relating to Resilience and, in certain cases, State and Software.
His Flat memory model study combines topics in areas such as Memory map and Distributed memory.
He most often published in these fields:
- Parallel computing (34.58%)
- Embedded system (18.69%)
- Memory bandwidth (14.95%)
What were the highlights of his more recent work (between 2018-2021)?
- Artificial intelligence (7.48%)
- Deep learning (6.54%)
- Parallel computing (34.58%)
In recent papers he was focusing on the following fields of study:
Mattan Erez focuses on Artificial intelligence, Deep learning, Parallel computing, Fault tolerance and Overhead.
His Parallel computing study focuses mostly on Memory hierarchy and Memory bandwidth.
His Memory bandwidth research incorporates themes from Regularization, Network model and Inference.
He has researched Fault tolerance in several fields, including Interconnection, Embedded system, Snapshot and Node level.
The Embedded system study combines topics in areas such as Resistive touchscreen and Wear leveling.
His Overhead research includes themes of Software, Distributed computing, State and Failure rate.
Between 2018 and 2021, his most popular works were:
- PruneTrain: fast neural network training by dynamic sparse model reconfiguration (20 citations)
- Mini-batch Serialization: CNN Training with Inter-layer Data Reuse (10 citations)
- PruneTrain: Gradual Structured Pruning from Scratch for Faster Neural Network Training. (9 citations)
In his most recent research, the most cited papers focused on:
- Operating system
- Central processing unit
- Programming language
Convolutional neural network, Memory bandwidth, Artificial intelligence, Regularization and Artificial neural network are his primary areas of study.
His research on Convolutional neural network also deals with topics like
- Computation which is related to area like Dram, Bandwidth and Serialization,
- High memory and related Computer architecture.
His studies in Memory bandwidth integrate themes in fields like Network model and Computer engineering.
His research on Artificial intelligence often connects related areas such as Parallel computing.
Mattan Erez works on Parallel computing which deals in particular with Memory hierarchy.
Mattan Erez combines subjects such as Inference and Pruning with his study of Regularization.
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.
