Minds and Computers : An Introduction to the Philosophy of Artificial Intelligence
Download 1.05 Mb. Pdf ko'rish
|
document (2)
Figure 12.5
Pruning the tree. used to challenge chess grandmasters are able to evaluate around two hundred million states per second. Let’s compare what we’ve learned about automated chess play to what we intuitively know about human chess play. For starters, humans – at least those who play very well – also mem- orise large numbers of opening strategies and variations thereof. For seconds, humans are also very good at determining which moves are unlikely to be made by an opponent playing well. Those humans who are very experienced at chess also have a very good sense of how good a particular game state is for them. It seems then that there is a good fit between what we know of automated chess play and what we can intuit concerning the play of accomplished humans. There are, however, important distinctions between the two, the most obvious of these being the sheer quantity of search and evaluation carried out by chess computers. It is clear that chess grandmasters do not explicitly apply a heuris- tic function to some two hundred million states per second in order to evaluate the goodness of the current game state and determine a strategy. We need to be very careful here, however, as we might be tempted to countenance a very poor argument against computation- alism in light of this. The very poor argument runs as follows. Computers need to search two hundred million states per second in order to match the best human players. These human players, however, clearly do not. So human players are not enacting formal systems in playing chess. Therefore computationalism is false, since there is at least one cogni- tive task whose function cannot be accounted for in terms of the operations of formal systems. There are a number of reasons why this argument is bad. The most telling criticism of it is that there is a distinction between human per- formance and our best attempts at recreating it. It may well be the case – and will be if computationalism is true – that human players do actually employ these methods. It is just that humans have very good guiding principles that allow them to engage in significantly less search to determine winning strategies. The fact that these human players are unable to make explicit the principles they are guided by does not speak against the possibility that their cognitive functions are to be accounted for in terms of the opera- tions of formal systems. After all, much of our mental life is opaque to introspection. It merely speaks against the best results of our empirical investigation in seeking to determine the operations of these systems. 129 Another criticism of the argument against computationalism is that it equivocates between the cognitive functions we explicitly and consciously engage in and cognitive function per se. To say that a grandmaster does not explicitly consider two hundred million states per second is not equivalent to saying that their cognitive processes of evaluation are not implicitly tantamount to consideration of such a number of states. Relying on a claim of the latter when only the former is demonstrable is an equivocation which begs the question against the computationalist. While we clearly need to be careful about the conclusions we draw from perceived distinctions between computer chess play and human chess play, there are yet further di fferences that are worth mention- ing. For one thing, the minimax procedure assumes that one’s opponent will always make the move which is optimal for them. Very often, however, it will be the case that a human player will make a sub-optimal move. It may be the case that they fail to recognise the optimal move, or it may be the case that they deliberately play a sub-optimal move to try to force their opponent into following a particularly strategy. A computer may well, of course, also make a sub-optimal move, simply through failure to correctly evaluate. Deliberately making a sub-optimal move to achieve an e ffect in one’s opponent’s strategy is, however, another thing entirely. In order to actively seek to influence the strategy of one’s opponent by making an unusual move, one requires an understanding of how one’s opponent thinks about the game. In other words, one requires some appreciation of how one’s opponent is likely to move given certain circumstances. The best way to discern this is to play as much as possible against the opponent in question. During computer versus human tournament play, it is generally allowed for the programmers to modify the computer program between games in order to counter strategies deployed by the human player. It is important to note that where the computer requires human intervention to be able to respond to the idiosyncrasies of a given player’s strategy, rapid responsive adaptiveness to peculiarity seems to be something that humans naturally excel at. This leads us to the final point of distinction between human and computer game play we’ll consider here. Consider the fact that the principles which guide human evalua- tion of chess states are su fficiently superior to those employed by a computer as to require significantly less explicit evaluation of descen- dant states in order to play comparably well. Consider also the rapid 130 responsiveness to the idiosyncrasies of a given player’s strategy which is evidenced by human players but not computers. It appears that there is something subserving both these abilities which humans are very good at but which hasn’t been mentioned at all in this discussion of machine game play. Humans excel at extracting Download 1.05 Mb. Do'stlaringiz bilan baham: |
Ma'lumotlar bazasi mualliflik huquqi bilan himoyalangan ©fayllar.org 2024
ma'muriyatiga murojaat qiling
ma'muriyatiga murojaat qiling