MS
Michael B. Sullivan
cs.DCcs.LGcs.ARcs.SEstat.MLcs.AIcs.CRcs.CV
On Valency
published · living versionsW_9kxxyt47·v1 · currentpublished
Making Convolutions Resilient via Algorithm-Based Error Detection Techniques
with Siva Kumar Sastry Hari, Timothy Tsai, Stephen W. Keckler
1 version
Preprints & journals
8 papers in the corpus · 2018–2025Cerium: A Multi-GPU Framework for Terabyte-Scale Encrypted Inference2512.11269v2 · Siddharth Jayashankar, Joshua Kim, Michael B. Sullivan et al.2025 · 1 citationACM SIGOPS 32nd Symposium on Operating Systems Principles (SOSP '26), September 29-October 02, 2026, Prague, Czech Republic
Making Convolutions Resilient via Algorithm-Based Error Detection Techniques2006.04984v1 · Siva Kumar Sastry Hari, Michael B. Sullivan, Timothy Tsai et al.2020 · 79 citationsarXivon Valency
Estimating Silent Data Corruption Rates Using a Two-Level Model2005.01445v1 · Siva Kumar Sastry Hari, Paolo Rech, Timothy Tsai et al.2020 · 7 citationsarXiv
HarDNN: Feature Map Vulnerability Evaluation in CNNs2002.09786v2 · Abdulrahman Mahmoud, Siva Kumar Sastry Hari, Christopher W. Fletcher et al.2020 · 33 citationsarXiv
Kayotee: A Fault Injection-based System to Assess the Safety and Reliability of Autonomous Vehicles to Faults and Errors1907.01024v1 · Saurabh Jha, Timothy Tsai, Siva Hari et al.2019 · 34 citationsarXiv
ML-based Fault Injection for Autonomous Vehicles: A Case for Bayesian Fault Injection1907.01051v1 · Saurabh Jha, Subho S. Banerjee, Timothy Tsai et al.2019 · 120 citationsarXiv
Buddy Compression: Enabling Larger Memory for Deep Learning and HPC Workloads on GPUs1903.02596v2 · Esha Choukse, Michael Sullivan, Mike O'Connor et al.2019 · 58 citationsarXiv
CRUM: Checkpoint-Restart Support for CUDA's Unified Memory1808.00117v1 · Rohan Garg, Apoore Mohan, Michael Sullivan et al.2018 · 32 citationsarXiv
Career total: 57 works. 8 are in this corpus.
Profile built from the corpus for this byline.
Author records are still filling in while the Hub is in alpha. If this is your page, you'll be able to claim it soon. Spot a mistake? Tell us.