272 citations · 2.5k across the 136 of their papers we have counts for
15 papers · 1 filter
Video Action Understanding
Matthew Hutchinson, Vijay Gadepally
Many believe that the successes of deep learning on image understanding problems can be replicated in the realm of video understanding. However, due to the scale and temporal natur…
Survey of Machine Learning Accelerators
Albert Reuther, Peter Michaleas, Michael Jones +3
New machine learning accelerators are being announced and released each month for a variety of applications from speech recognition, video object detection, assisted driving, and m…
Layer-Parallel Training with GPU Concurrency of Deep Residual Neural Networks via Nonlinear Multigrid
Andrew C. Kirby, Siddharth Samsi, Michael Jones +3
A Multigrid Full Approximation Storage algorithm for solving Deep Residual Networks is developed to enable neural network parallelized layer-wise training and concurrent computatio…
Accuracy and Performance Comparison of Video Action Recognition Approaches
Matthew Hutchinson, Siddharth Samsi, William Arcand +16
Over the past few years, there has been significant interest in video action recognition systems and models. However, direct comparison of accuracy and computational performance re…
Benchmarking network fabrics for data distributed training of deep neural networks
Siddharth Samsi, Andrew Prout, Michael Jones +16
Artificial Intelligence/Machine Learning applications require the training of complex models on large amounts of labelled data. The large computational requirements for training de…
Best of Both Worlds: High Performance Interactive and Batch Launching
Chansup Byun, Jeremy Kepner, William Arcand +16
Rapid launch of thousands of jobs is essential for effective interactive supercomputing, big data analysis, and AI algorithm development. Achieving thousands of launches per second…
Multi-Temporal Analysis and Scaling Relations of 100,000,000,000 Network Packets
Jeremy Kepner, Chad Meiners, Chansup Byun +23
Our society has never been more dependent on computer networks. Effective utilization of networks requires a detailed understanding of the normal background behaviors of network tr…
DBOS: A Proposal for a Data-Centric Operating System
Michael Cafarella, David DeWitt, Vijay Gadepally +5
Current operating systems are complex systems that were designed before today's computing environments. This makes it difficult for them to meet the scalability, heterogeneity, ava…
ExSample: Efficient Searches on Video Repositories through Adaptive Sampling
Oscar Moll, Favyen Bastani, Sam Madden +3
Capturing and processing video is increasingly common as cameras become cheaper to deploy. At the same time, rich video understanding methods have progressed greatly in the last de…
Fast Mapping onto Census Blocks
Jeremy Kepner, Andreas Kipf, Darren Engwirda +21
Pandemic measures such as social distancing and contact tracing can be enhanced by rapidly integrating dynamic location data and demographic data. Projecting billions of longitude…
Technical Report: Developing a Working Data Hub
Vijay Gadepally, Jeremy Kepner
Data forms a key component of any enterprise. The need for high quality and easy access to data is further amplified by organizations wishing to leverage machine learning or artifi…
GraphChallenge.org Sparse Deep Neural Network Performance
Jeremy Kepner, Simon Alford, Vijay Gadepally +5
The MIT/IEEE/Amazon GraphChallenge.org encourages community approaches to developing new solutions for analyzing graphs and sparse data. Sparse AI analytics present unique scalabil…
GraphChallenge.org Triangle Counting Performance
Siddharth Samsi, Jeremy Kepner, Vijay Gadepally +9
The rise of graph analytic systems has created a need for new ways to measure and compare the capabilities of graph processing systems. The MIT/Amazon/IEEE Graph Challenge has been…
AI Data Wrangling with Associative Arrays
Jeremy Kepner, Vijay Gadepally, Hayden Jananthan +2
The AI revolution is data driven. AI "data wrangling" is the process by which unusable data is transformed to support AI algorithm development (training) and deployment (inference)…
75,000,000,000 Streaming Inserts/Second Using Hierarchical Hypersparse GraphBLAS Matrices
Jeremy Kepner, Tim Davis, Chansup Byun +16
The SuiteSparse GraphBLAS C-library implements high performance hypersparse matrices with bindings to a variety of languages (Python, Julia, and Matlab/Octave). GraphBLAS provides…