Transactions on Machine Learning and Data Mining (Print ISSN: 1865-6781) (ONLINE-ISSN: 2509-9337)
Volume 12 - Number 1 - July 2019 - Pages 23-29
Metacognitive Architectures for Human Roles in Machine Learning for Analyzing Multimedia Data
Piet Kommers
Professor UNESCO Learning Technologies Emeritus
University of Twente, The Netherlands
Abstract
Machine Learning is often regarded as antagonist to human cognition and deci-sion making. This article shows how schematic diagrams offer a unique mitiga-tion between the human- / machine partnership. Machine Learning grows fast in power and speed; it demands a carefully-chosen interface in order to benefit from the typically human mental faculties like intuitive-, imaginative-, moral- and exis-tential. For such diverse and unpredictable symbiosis, the term ‘interface’ is an understatement. Similar to the ambition of the ‘5th generation expert systems in the late 80ties’ we face an underestimated challenge to exploit metaphoric repre-sentations like the early transition from map into circuit scheme of the London Underground Map.
Keywords: Machine Learning, Human Cognition, Intuition, Imagination, Schematic Repre-sentation, Circuit Representations.
Download Paper
(379 KB)