@conference {Ko2020, title = {A 3mm2 Programmable Bayesian Inference Accelerator for Unsupervised Machine Perception using Parallel Gibbs Sampling in 16nm}, booktitle = {IEEE Symposium on VLSI Circuits (VLSI)}, year = {2020}, abstract = {This paper describes a 16nm programmable accelerator for unsupervised probabilistic machine perception tasks that performs Bayesian inference on probabilistic models mapped onto a 2D Markov Random Field, using MCMC. Exploiting two degrees of parallelism, it performs Gibbs sampling inference at up to 1380{\texttimes} faster with 1965{\texttimes} less energy than an Arm Cortex-A53 on the same SoC, and 1.5{\texttimes} faster with 6.3{\texttimes} less energy than an embedded FPGA in the same technology. At 0.8V, it runs at 450MHz, producing 44.6 MSamples/s at 0.88 nJ/sample.}, keywords = {accelerators, bayesian inference, deep learning}, url = {https://doi.org/10.1109/VLSICircuits18222.2020.9162784}, author = {Ko, Glenn and Chai, Yuji and Marco Donato and Paul Whatmough and Tambe, Thierry and Rob Rutenbar and David Brooks and Gu-Yeon Wei} }