Figure - uploaded by Cassey Lee
Content may be subject to copyright.
Source publication
The goal of this essay is to reexamine the nature and relationship between bounded rationality and AI in the context of recent developments in the application of AI to two-player zero sum games with perfect information such as Go. This is undertaken by examining the evolution of AI programs for playing Go. Given that bounded rationality is inextric...
Contexts in source publication
Context 1
... like today, was fairly heterogeneous methodologically then. This can be seen in the four categories of definitions of AI proposed by Russell and Norvig (2010), which continue to be valid descriptions to this day ( Table 1). ...Context 2
... computer programs have clearly evolved over time with advances in computer software (algorithms) and hardware. Appendix Table 1 summarizes some of milestones in the application of AI to games. The progress made in AI-application to games have been possible in three key elements, namely: (i) knowledge; (ii) search; and (iii) learning. ...Similar publications
This essay presents some aspects of the Gromov-Witten theory from the point of view of symplectic topology. Symplectic manifolds are smooth even dimensional manifolds admitting a symplectic struc- ture: a non-degenerate closed 2-form. There is not much one can say about the global structure of a symplectic manifold by working locally, which is in c...
Citations
... Algorithms work with finite computational resources which in practice means that they cannot achieve Turing completeness and are limited to linear bounded automation." The reference here to Turing Completeness is to the theoretical possibility that AI can be globally rational as the mythical agent of neoclassical economics (Lee, 2019). A Universal Turing Machine (UTM) is a "computing machine", proposed by Turing (1936) that can "be used to compute any computable sequence" (Turing, 1936, p.241). ...
... It is Turing Complete. However, it is subject to the Halting Problem, which is how to determine if and when the UTM will find a solution (Lee, 2019). Turing (1936) proved that there is no general algorithm for solving this problem in all cases. ...
... The pioneering contributions by Simon to the disciplinary fields of both economics and AI are well-known and the concept of bounded rationality was in his view applicable to both contexts, even if in recent decades it has been pursued far more extensively in economics, staying instead rather overlooked in AI (just as AI was substantially unrelated to the developments of the Simon-inspired branch of behavioural economics). Recently, however, the theme of explicitly addressing the boundaries of the rationality criteria applicable in tasks performed by artificial agents is becoming increasingly relevant, even in the context of narrow AI [7]. For example, in the games of chess and Go, computers have undoubtedly surpassed humans, but they have not done so by identifying moves that are demonstrably better on strictly formal (logical and mathematical) grounds. ...
The successes of Artificial Intelligence in recent years in areas such as image analysis, natural language understanding and strategy games have sparked interest from the world of finance. Specifically, there are high expectations, and ongoing engineering projects, regarding the creation of artificial agents, known as robotic traders, capable of juggling the financial markets with the skill of experienced human traders. Obvious economic implications aside, this is certainly an area of great scientific interest, due to the challenges that such a real context poses to the use of AI techniques. Precisely for this reason, we must be aware that artificial agents capable of operating at such levels are not just round the corner, and that there will be no simple answers, but rather a concurrence of various technologies and methods to the success of the effort. In the course of this article, we review the issues inherent in the design of effective robotic traders as well as the consequently applicable solutions, having in view the general objective of bringing the current state of the art of robo-trading up to the next level of intelligence, which we refer to as Cognitive Trading. Key to our approach is the joining of two methodological and technological directions which, although both deeply rooted in the disciplinary field of artificial intelligence, have so far gone their separate ways: heuristics and learning.
... The game starts with an empty board, to capture opponent piece(s). The size of the Go board contributes to its complexity, and paves for being the master challenge after chess has been solved (Lee, 2019). A board with a smaller size exists like 15 X 15, which was used for playing Gomoku games as reflected in (Tang et al., 2017). ...
Popular games, such as Chess, Tic-Tac-Toe, and Go, have been formulated as search problems, and consequently many research works have been conducted concerning these games with numerous artificial intelligence (AI) search algorithms been used as the underlying search algorithms for solving such games. Despite the popularity of the ancient Daragame within African communities and its two distinct actions of Go (positioning) and Chess (movement) the game attracts less research efforts; as such no search algorithm was evaluated using the Dara game. The work adopts a mathematical formulation of Shiva Dara (Dara) using a five tuples Aisearch problem formulation model. Minimax and greedy with the same heuristic function are generic search algorithms used to develop agents that play the implemented model (game). The agents were set to play against one another to evaluatethe performance of the underlying search algorithms. The results show that the proposed model is biasedto the opponent player, and the greedy agent defeated the minimax agent and is more optimal. We consider evaluating other search algorithms or hybridization of the two algorithms in subsequent work.
