Distribute Inference to Support Inter-subjective Formalization and its Application to Sensor Networks

2016 Fiscal Year Final Research Report

• Project/Area Number: 25540101
• Research Category: Grant-in-Aid for Challenging Exploratory Research
• Allocation Type: Multi-year Fund
• Research Field: Intelligent informatics
• Research Institution: Osaka University
• Principal Investigator: Numao Masayuki 大阪大学, 産業科学研究所, 教授 (30198551)
• Project Period (FY): 2013-04-01 – 2017-03-31
• Keywords: 間主観 / 論理型言語 / 分散推論 / 帰納論理プログラミング / FPGA / センサーネットワーク

Outline of Final Research Achievements

We usually prove logical formulas by rewriting them step by step. Reduction machines have been proposed for functional programming languages to make an inference based on such a rewriting mechanism. However, it has not been efficient in distributed environment, since they rewrite a logical formula on a memory by using processors. A computer network has many switches, and transfers packets to their destinations. We propose to rewrite a formula in logic or algebra on distributed switches and state memories with higher-order meta-rules. Although such inference seems similar to one by a production rule in expert systems, it utilizes distributed working memories and self-optimizing properties in their inference with meta-rules. We show this mechanism is appropriate for weight-based learning for controlling its inference, and inter-subjective formalization for a sensor network in Empathic Computing.

Free Research Field

人工知能