| 104D6 |
|---|
| An LSI implementation of a finite physical quantity neural network model |
| AUTHORS |
| Minoru Watanabe , Fuminori Kobayashi (Kyushu Institute of Technology) |
| ABSTRACT |
| A standard-cell-circuit of a finite physical quantity neural network (FPQNN) model for LSIs is proposed to be able to easily add intelligence units to systems on the LSIs.The FPQNN model has a different feature to artificial neural network models, which keeps the summation of all the physical quantity in the network a finite when recalling a stored pattern but being given physical quantity by external works.The cell-circuit can precisely express the movement of the finite physical quantity between neurons and quickly execute the enormous iterative calculations in the FPQNN. In addition, the cell-circuits have an advantage to be able to be connected to other digital systems on the same chip directly without any interface. This paper presents estimation of the FPQNN cells designed for Standard-Cell-Layout and HSPICE simulation results for pattern recognition. |
Back to Program
Time Table
Top of This Site
Annual Conference 2002