Digital electronics is based on stable bits that can have one of two values, 0 and 1. At the other extreme we have quantum computing using using q-bits that can be in superposition states that are 0 and 1 at the same time. In our recent work we have introduced a concept that is intermediate between bits and q-bits, namely a probabilistic bit or a p-bit that fluctuates randomly between 0 and 1.
In this talk, we will argue that within MRAM technology the magnetic tunnel junctions (MTJ’s) that store bits, with minor modifications, can represent p-bits as well. These can be used to build p-circuits for (1) combinatorial optimization, (2) Bayesian inference, (3) invertible Boolean logic that not only provides the output for given inputs, but also all inputs consistent with a given output, (4) image recognition and other machine learning applications.
- Bayesian Inference and Optimization
- Invertible Boolean
- Non-magnetic implementation
-  A.Z.Pervaiz, L.A.Ghantasala, K.Y. Camsari and S.Datta, “Hardware Emulation of Stochastic p-bits for Invertible Logic,” Scientific Reports, 7, 10994 (2017).
-  A.Z.Pervaiz, L.A.Ghantasala, B.M.Sutton and K.Y. Camsari “Weighted p-bits for FPGA implementation of probabilistic circuits,” arxiv.org/abs/1712.04166.
- Embedded MRAM based implementation
-  K.Y. Camsari, S. Salahuddin and S.Datta “Implementing p-Bits with Embedded MTJ’s,” IEEE Electron Device Letters, 38, 1767 (2017).
-  O.Hassan, K.Y.Camsari and S.Datta ”Voltage-driven Building Block for Hardware Belief Networks,” arxiv.org/abs/1801.09026.
-  R.Faria, K.Y.Camsari and S.Datta ”Implementing Bayesian Networks with Embedded MRAM,” AIP Advances, 8, 045101.
Cite this work
Researchers should cite this work as follows:
Purdue University, West Lafayette, IN