Speaker:
Prof. Marek Perkowski
Department of Electrical and Computer Engineering
Portland State University
Oregon, USA.


BACK TO MAIN PAGE OF Professor Marek Perkowski


Title: EVOLVABLE HARDWARE OR LEARNING HARDWARE?

INDUCTION OF STATE MACHINES FROM TEMPORAL LOGIC CONSTRAINTS

April 2nd, Friday, 4-5 p.m., room PCAT 28.

Abstract

Evolvable Hardware is Genetic Algorithm (GA) plus reconfigurable hardware. One may ask: "Why Genetic Algorithm"? Based on our experience, we question the usefulness of GA as a sole learning method to reconfigure binary FPGAs. Instead, we propose the "Learning Hardware" approach, which consists of creating a sequential network based on feedback from the environment (for instance, positive and negative examples from the trainer), and realizing this network in an array of Field Programmable Gate Arrays (FPGAs).

Here we advocate the approach to Learning Hardware based on Induction of State Machines from Temporal Logic Constraints. The method involves training on examples, constraints solving, determinization, FSM minimization, structural mapping, functional decomposition of multi-valued logic functions and relations, and FPGA mapping. Thus learning occurs on the level of constraints acquisition and functional decomposition rather than on the low level of programming binary switches. Ours is an Occam's Razor learning that allows for generalization and discovery.

Our software algorithms require fast operations on complex logic expressions and solving NP-complete problems such as satisfiability. They should be realized in hardware to obtain the necessary speed-ups. Using a fast prototyping tool, the DEC-PERLE-1 board based on an array of Xilinx FPGAs, we are developing software/configware processors that accelerate the acquisition, synthesis, and optimization of Reactive State Machines.

PLAN:

1. Introduction of the ideas of Evolvable Hardware.

2. Evolvable Hardware or Learning Hardware.

3. Reconfigurable Computing for Learning.

4. Use of evolutionary ideas in robotics.

5. Robot-Cat, Robot-Dog, Robot-Human-Torso, or, what I have seen in Japan and what should we do.