Computational Problems & Cognitive Functions: Modeling, Characterizations, and Solutions

点击翻译按钮获取中文摘要

Due to rapidly advancing technology, many internet of things devices and apps now extend human capabilities, ranging from support for experts to everyday activities. As highlighted in IBM's News Room, David Stokes, Chief Executive for IBM, UK and Ireland, posits that "We’re at the dawn of a new era of cognitive computing, during which advanced data-centric computing models and open innovation approaches will allow technology to greatly augment decision-making capabilities for business and government." Research on this topic is commonly referred to as human problem solving (HPS). To advance research in this field, an increased understanding of the how, when and what of HPS is required. Fundamental to this endeavour is the characterization of cognitive functions. My research program investigates fundamental questions: What problems can people solve successfully? Which ones cannot be solved? What are the properties of the ones that people are interested in solving? What are successful solving strategies? We might find answers to some of these questions in the properties of games and puzzles, such as Sudoku. To identify and characterize cognitive functions, we draw on methods in classical and parameterized complexity, the development of exact algorithms, as well as empirical studies and their analyses. As an effective data collection and problem solving platform for our empirical studies, we intend to utilize Lively, an open source infrastructure for Web applications programming. An improved understanding of human problem solving impacts many different application areas directly, including but not limited to: (a) The game-design industry. Methods for identifying properties of the cognitive functions used when playing games or puzzles that many people are attracted to and intrigued by while having a satisfactory playing experience will provide the industry with strategies and tools to create new games involving these properties. (b) The design of software in general and the quality of self-adaptive systems in particular. For example to design self-driving cars that are safer than human operated cars, it is beneficial to understand how humans make decisions on the road. This will aid the design software functions that have to react to human decision making in other cars. (c) Education. The knowledge gained will be informative for educators at all levels, including teaching problem solving skills. (d) Training. Highly qualified personnel (HQP) trained in this area and in my research group in particular acquire an interdisciplinary education based on solid foundations in theoretical computer science, and learn how to conduct empirical studies. These HQP are highly sought after by industry in general, and in Canada in particular.