Publications

Serving DNNs like Clockwork: Performance Predictability from the Bottom Up

Published in 14th USENIX Symposium on Operating Systems Design and Implementation (OSDI), 2020

In this paper, starting with the predictable execution times of individual DNN inferences, we adopt a principled design methodology to successively build a fully distributed model serving system that achieves predictable end-to-end performance.

Recommended citation: Arpan Gujarati, Reza Karimi, Safya Alzayat, Wei Hao, Antoine Kaufmann, Ymir Vigfusson, Jonathan Mace, "Serving DNNs like Clockwork: Performance Predictability from the Bottom Up", 14th USENIX Symposium on Operating Systems Design and Implementation (OSDI), 2020. https://arxiv.org/abs/2006.02464

Learning Amyloid Pathology Progression from Longitudinal PiB-PET Images in Preclinical Alzheimer’s Disease

Published in 2020 IEEE 17th International Symposium on Biomedical Imaging (ISBI), 2020

This paper talks about a novel trainable network diffusion model that infers the propagation dynamics of amyloid pathology, which conditioned on individual-level structural connectivity network.

Recommended citation: Wei Hao, Nicholas M. Vogt, Zihang Meng, Seong Jae Hwang, Rebecca L. Koscik, Sterling C. Johnson, Barbara B. Bendlin, and Vikas Singh, "Learning Amyloid Pathology Progression from Longitudinal PiB-PET Images in Preclinical Alzheimer’s Disease", International Symposium on Biomedical Imaging (ISBI), 2020. https://ieeexplore.ieee.org/abstract/document/9098571

Sparse Convolutions for Faster Object Recognition

Published in AI Systems Workshop at SOSP, 2019

This paper presents a preliminary technique for accelerating ML inference on sparse inputs by modifying the convolution operator to be sparsity-aware.

Recommended citation: Wei Hao and Shivaram Venkataraman, "Sparse Convolutions for Faster Object Recognition", AI Systems Workshop at Symposium on Operating Systems Principles(SOSP), 2019. http://learningsys.org/sosp19/assets/papers/21_CameraReadySubmission_sparse_conv_aisys19_final.pdf