The Vanguard Project Overview

Chris Frueh Psychology 321 10.26.2008 Page 1 of 5 In the history of psychology, riddled as it is with vague or contradictory definitions espoused by one school of thought or another, few concepts have been less clear than what is generally called intuition. Osbeck, in particular, explored the historical blurriness of the topic in her 1999 paper. However, with the strong foundation set by Rosenblatt and Thickstun and built upon with Osbeck’s research, a consistent definition has arisen that bears exploring. I plan to show, in brief, how intuition, and specifically the subset of decision-making automatic thought, is a process strongly rooted in both memory and the pattern-recognition and comparison processes. Rosenblatt and Thickstun stated in their 1994 article “Intuition and Consciousness” that intuition functions “…as a form of unconscious pattern-matching cognition” (Rosenblatt 697). They elaborate further as they say that this ‘patternmatching’ functions as “…a currently perceived pattern is matched to a stored pattern. The result of this activity may become conscious… when a match is achieved.” (Rosenblatt 700) It is reasonable then to conclude that as an observer recognizes a situation similar to a previously encountered event, he might recall details of his decisions made and the results thereof. These recollections and comparisons may not even be fully conscious as one “… learn[s] from experience without awareness of doing so.” (Osbeck 231) In an attempt to explain in detail how this might work, one would need a mind just as capable of memory as a human’s would, if not more so, and be able to do a significantly large number of parallel comparative processes:

Chris Frueh Psychology 321 10.26.2008 Page 2 of 5 The computer and the human mind have different but complementary abilities. The computer excels in analysis and numerical computation, and human excels in pattern recognition, the assessment of complicated situations and intuitive leaps to new situations. (Gill 9) Given this recognition of the superior calculative ability of the computer, a revelation hardly new, one need only translate the identifying characteristics of a situation into symbols a computer can manipulate and design an adaptive protocol allowing the computer to somewhat mimic human memory and learning processes. To demonstrate in thought experiment form how such a monumental task would work, consider a digital computer game which I happen to pay in which a method for reading and recording the situational characteristics already existed. If one were set in a game such as this built around a simultaneously co-operative and counter-operative strategy and tactics framework, the foundation for this computer’s adaptation would nearly complete. To allow one to visualize this, a brief elaboration is required. While the game itself is unimportant to the workings of the intuition model, an understanding of the mechanics of the example might allow one a more vivid picture of the workings of the model itself. The game in question is a conflict between two teams. There are definable objectives, resources that may be expended, clear positional variables, a numerical health system, and an overt win condition. There is also a subtle hierarchy of importance between these various objectives and resources: momentum may be sacrifices to attain an objective or an objective may be surrendered to make victory certain. A human player

Chris Frueh Psychology 321 10.26.2008 Page 3 of 5 engaged in such a match examines and evaluates all of this at extremely rapid speeds, speeds beyond the scope of self reporting. Now the symbolic nature of a player’s thought may be understood in relation to Herbert Simon’s claim that “[human and computer] systems are symbol systems. They achieve their intelligence by symbolizing external and internal situations and events, and by manipulating these symbols.” (Norman 16) Now, in constructing an artificial intelligence to play this game using the memory-based intuition, the program needs to store the quantified characteristics of situations it encounters as well as the actions taken and the result. Only by recognizing future situations by their introductory characteristics will the computer be able to ‘remember’ previously successful tactics in the same way that I might watch match footage from previous games to learn tactics from players better than myself. Next, the computer would monitor the current situation of the game it is playing and compare it with previously encountered situations. Norman’s book elaborates this rather well: The principle mechanism of intelligence that we have observed… operating in problem environments is heuristic search. The “search” part of the heuristic search process is obvious enough. What are more subtle are the heuristic search devices that enable an intelligent system to carry on searches with great selectivity by (1) using information stored in memory to choose more promising over less promising paths, and (2) extracting from the problem environment new information about regularities in its structure that can similarly guide the search. (Norman 19)

Chris Frueh Psychology 321 10.26.2008 Page 4 of 5 These subtleties are exactly what the deliberative nature and hyper-accurate memory of the computer will bring forth as it scans through previous scenarios and compares the relevance of each historical scenario with the current in real time. In doing so, the computer is constantly discarding irrelevant tactics as the fluid situation changes. Thus, when a situation presents an opportunity to advance to a more advantageous resolution, the computer will work to maximize the minimum amount of gain like the MaxiMin protocol. The computer will enact the most successful tactic that most fits the current situation after analyzing the usefulness of the maneuver’s results to the strategic goals of the team. Given these two steps so far, the computer can watch every moment of any game and save the tactics used for future situations calling them up again as the situation warrants. All that is required is an adaptive coda. As each strategy is acted out over a theoretical infinite number of games, the computer reanalyzes each strategy’s successfulness and re-categorizes each appropriately. To relate this back to intuition, the human mind acts in a similar way, albeit with more rapid leaps and with less recognition of the process’ workings. While, admittedly, an inexperienced individual may have insight into the nature of a problem, an elder person has the experience to recall previously pitfalls or successes and act accordingly, a concept embodied in the phrase “seen the world”. This model would, as an experienced player would, learn from past mistakes even if he himself had not made them and would update his methods based on real-time results.

Chris Frueh Psychology 321 10.26.2008 Page 5 of 5 In summary, a memory-based intuition model is logically consistent and previous research supports the framework upon which it is based. A person acting in this intuitive fashion in a similar automatic decision would recognize the various pressures and choices and, recollecting a previous experience, would act in such a way to either avoid or promote the previous outcome. In this light, perhaps future developments in computer languages might allow further detailed enactments of decision-making processes that would confirm or deny the hypotheses presented here.