A5036470848- Parallel Computing and Optimization Techniques
- Security and Verification in Computing
- Interconnection Networks and Systems
- Advanced Memory and Neural Computing
- Cryptographic Implementations and Security
- Radiation Effects in Electronics
- Physical Unclonable Functions (PUFs) and Hardware Security
- Microbial Natural Products and Biosynthesis
- Stochastic Gradient Optimization Techniques
- Auction Theory and Applications
- Engineering Applied Research
- Liver Disease Diagnosis and Treatment
- Biochemical and Molecular Research
- Energy and Environmental Systems
- EEG and Brain-Computer Interfaces
- Cloud Computing and Resource Management
- Mosquito-borne diseases and control
- Network Packet Processing and Optimization
- Technology and Data Analysis
- Power Systems and Technologies
- Foreign Body Medical Cases
- Advanced Neural Network Applications
- Distributed and Parallel Computing Systems
- Business Process Modeling and Analysis
- Technology Assessment and Management
Samsung (South Korea)
2025
University of Florida
2024
American Society For Engineering Education
2024
Texas A&M University
2019-2024
Yonsei University
2024
Mitchell Institute
2019-2024
Uijeongbu St. Mary's Hospital
2018
Catholic University of Korea
2018
Abstract Mosquitos can be dangerous because they transfer viruses and parasites to animals humans. These harmful agents include West Nile, dengue, Zika, such as Malaria. For mosquito research, climate data environmental conditions used in studying the effect on breeding sites of mosquitoes population control. Microclimate monitored for better resolution analysis deeper understanding human lives. In order obtain microclimate data, a custom monitoring system has been development. This is based...
The explosion of data availability and the demand for faster analysis have led to emergence applications exhibiting large memory footprint low reuse rate. These workloads, ranging from neural networks graph processing, expose compute kernels that operate over myriads data. Significant movement requirements these impose heavy stress on modern subsystems communication fabrics. To mitigate worsening gap between high CPU computation density deficient bandwidth, solutions like near-data...
Ever increasing performance demand and shrinking in the transistor size together result complex dense packing large chips. That motivates designers to opt for many small specialized hardware modules a chip extract maximum benefits with relatively lower complexity cost. These altogether opens up new directions heterogeneous modular System-on-Chip (SoC) research, where system is built by assembling independently designed chiplets (small chips). We focus on communication aspect of such SoCs,...
Large-scale distributed deep learning training has enabled developments of more complex neural network models to learn from larger datasets for sophisticated tasks. In particular, stochastic gradient descent intensively invokes all-reduce operations update, which dominates communication time during iterative epochs. this work, we identify the inefficiency in widely used algorithms, and opportunity algorithm-architecture co-design. We propose MultiTree algorithm with topology resource...
Spectre and Meltdown attacks reveal the perils of speculative execution, a prevalent technique used in modern processors. This paper proposes ReViCe, hardware to mitigate speculation based attacks. ReViCe allows loads update caches early but keeps any replaced line victim cache. In case misspeculation, lines from cache are restore caches, thereby preventing cache-based Moreover, injects jitter conceal timing difference due lines. Together restoration injection allow make execution secure. We...
Mosquitos might have caused approximately seven hundred thousand deaths per year.Monitoring environmental factors for potential breeding sites of mosquitoes is an important and effective method in mosquito control.Several networks systems research been developed academic settings.For the this paper, developmental progress a low-cost data server network form IoT (Internet Things) system with undergraduate engineering students presented.A was proposed as one implementations research.This...
With the widespread adoption of Deep Learning (DL) models, demand for DL accelerator hardware has risen. On top that, models are becoming massive in size. To accommodate those multi-chip-module (MCM) emerges as an effective approach implementing large-scale accelerators. While MCMs have shown promising results inference, its potential Training remains largely unexplored. Current approaches fail to fully utilize available links a mesh interconnection network MCM accelerator. address this...
The data explosion and faster analysis demand have spawned emerging applications that operate over myriads of exhibit large memory footprints with low reuse rate. Such characteristics lead to enormous movements across the hierarchy pose significant pressure on modern communication fabrics subsystems. To mitigate worsening gap between high processor computation density deficient bandwidth, networks, near-data processing techniques are proposed keep improving system performance energy...
Memory safety invariants extracted from a program can help defend and detect against both software hardware memory violations. For instance, by allowing only specific instructions to access certain locations, system out-of-bound or illegal pointer dereferences that lead correctness security issues. In this paper, we propose CPU abstractions, called, specify check provide defense mechanism violations at runtime. ensures the must be satisfied every accesses. We present fast invariant address...
For a distributed last-level cache (LLC) in large multicore chip, the access time to one LLC bank can significantly differ from that another due difference physical distance. In this paper, we successfully demonstrate new distance-based side-channel attack by timing AES decryption operation and extracting part of an secret key on Intel Knights Landing CPU. We introduce several techniques overcome challenges attack, including use multiple threads ensure hits, detect vulnerable memory...
Although surgical resection is a curative treatment option for solitary hepatocellular carcinoma, high recurrence rate contributes to dismal long-term prognosis after resection. Early within 2 years surgery associated with intrahepatic metastasis of primary tumor. Liver regeneration hepatic can accelerate tumorigenesis in remnant liver. Treatment strategies include salvage transplantation, repeated resection, local ablation, and transarterial chemoembolization (TACE). Here, we report...
The increase in design cost and complexity have motivated designers to adopt modular of System on Chip (SoC) by integrating independently designed small chiplets. However, it introduces new challenges for correctness validation, increasing chances forming deadlock the system involving multiple Although there been many solutions available freedom flat networks, study issue chiplet-based systems is still its infancy. A recent suggests adding extra turn restrictions as a viable solution this...
Agent-Based Approach for Intelligent Automated Landfill Operation System Sung-Keun Kim, Kyo-Jin Koo, Jeffrey S. Russell Pages 1-6 (2000 Proceedings of the 17th ISARC, Taipei, Taiwan, ISBN 9789570266986, ISSN 2413-5844) Abstract: Workers and operators in sanitary landfill sites face health risk due to constant exposure toxic solid, liquid, gaseous by-products waste materials. One solution this problem is automation technology that can increase productivity quality as well safety. But,...
The resurgence of near-memory processing (NMP) with the advent big data has shifted computation paradigm from processor-centric to memory-centric computing. To meet bandwidth and capacity demands computing, 3D memory been adopted form a scalable memory-cube network. Along NMP system development, mapping for placing guiding in network become crucial driving performance improvement NMP. However, it is very challenging design universal optimal all applications due unique application behavior...
For a distributed last-level cache (LLC) in large multicore chip, the access time to one LLC bank can significantly differ from that another due difference physical distance. In this paper, we successfully demonstrated new distance-based side-channel attack by timing AES decryption operation and extracting part of an secret key on Intel Knights Landing CPU. We introduce several techniques overcome challenges attack, including use multiple threads ensure hits, detect vulnerable memory...