Sciweavers

Share
FPL
2009
Springer

A multi-FPGA architecture for stochastic Restricted Boltzmann Machines

9 years 4 months ago
A multi-FPGA architecture for stochastic Restricted Boltzmann Machines
Although there are many neural network FPGA architectures, there is no framework for designing large, high-performance neural networks suitable for the real world. In this paper, we present two concepts to support a multi-FPGA architecture for stochastic Restricted Boltzmann Machines (RBM), a popular type of neural network. First, a hardware core, called the kth Stage Piecewise Linear Interpolator, is used to implement a high-precision, pipelined function generator. The interpolator increases the resolution of a Look Up Table implementation, guaranteeing an additional bit of precision for every pipeline stage. This function generator is used to implement a sigmoid function required in stochastic node selection. Next, a partitioning algorithm is used to efficiently divide a RBM amongst multiple FPGAs. The partitioning algorithm optimizes performance by minimizing the inter-FPGA communication. The architecture is tested on the Berkeley Emulation Engine 2 running at 100MHz. One board su...
Daniel L. Ly, Paul Chow
Added 24 Jul 2010
Updated 24 Jul 2010
Type Conference
Year 2009
Where FPL
Authors Daniel L. Ly, Paul Chow
Comments (0)
books