Minds and Computers : An Introduction to the Philosophy of Artificial Intelligence
Download 1.05 Mb. Pdf ko'rish
|
document (2)
- Bu sahifa navigatsiya:
- Exercise 10.3
participants in very short order. This is not, however, ipso facto an indictment of the Turing test, but rather of the current state of play in artificial intelligence. Perhaps a behavioural test of this kind, while a good indicator under certain strong conditions, simply does not resolve the question satisfactorily one way or another in problem cases. In such instances it is not obvious what other methods we might employ to determine the presence or absence of a mind. Exercise 10.3 Would a version of the Turing test involving a number of judges and unrestricted in duration serve as a reliable indicator of mentality? Can you think of any other conditions 111 or constraints we might build in to the test to increase its reliability? Let’s pause now and think about where we are and where we are going. So far in this book, we have introduced various philosophical problems associated with understanding the mind and examined a number of theories which propose answers to these questions. We have also introduced some rudimentary neuroanatomy to get a feel for how brains work, given that there is clearly some important rela- tion between brains and minds. Of the philosophical theories of mind we have covered, function- alism was the least problematic and most satisfactory. However, it left us wanting to know more about the relevant functions in question. In order that we might be well prepared to understand a particular way of answering this, we then rigorously developed a precise account of computation. Armed with this understanding, we have now examined a particu- lar way of fleshing out functionalism which supposes that the func- tions in question are computations – namely computationalism. We have also lent some consideration to the question of the appropriate conditions under which we should attribute mentality. There will be much more to say about this as we progress. In the coming chapters we are going to see how a commitment to computationalism confers a methodology for artificial intelligence research. We will be concentrating on various computational prob- lems which must be solved in order to equip a device with the capacity to reason and use language. Along the way we will be gathering evidence which bears on the tenability of computationalism – an issue we will explicitly return to at the beginning of Chapter 17 when we begin to mount serious chal- lenges and sophisticated philosophical objections to the theory. Now it is time to begin telling the story of artificial intelligence – a story which begins with the concept of search. 112 C H A P T E R 1 1 SEARCH Given the computationalist hypothesis as we have described it, inves- tigating mentality involves investigating the operations of the formal system [MIND] and its constituents. The classical or symbol systems approach to artificial intelligence involves trying to determine the algorithms for the cognitive functions involved in [MIND] and investigating their associated formal sys- tems. Many of the problems in the classical artificial intelligence tradition reduce to determining whether or not a certain state can be generated in a particular formal system, or to finding a particular generated state of a system. In other words, many artificial intelligence problems reduce to the problem of searching for a particular state. In this chapter, we are going to briefly run through various methods for searching the generation tree of a formal system. You may wish to refer to section 7.5 to refresh your understanding of gen- eration trees and the associated terminology. 11.1 TOP DOWN, BOTTOM UP One method of determining whether a state of a formal system is gen- erated, or finding a derivation for a particular state of the system, is to construct the entire generation tree for the system. If the state we are interested appears on the tree it is a generated state and we can read o ff its derivation(s) by following the branches back up to the root node. For instance, suppose we are investigating the system [BIN] from section 7.5 and we are interested in whether the state 01 is generated. We can construct a generation tree for [BIN] beginning from the initial state and working our way down through all the possible ways in which rules can be applied to each node. Figure 11.1 gives a generation tree for [BIN] which shows that (and how) the state 01 is generated. 113 A search of this kind is known as a top-down search. Often, however, it will be more e fficient to begin with the state we are trying to derive and work backwards through the rules to see if we can get back to the initial state. This kind of search is called a bottom up search. Figure 11.2 depicts a bottom up search in [BIN] for the state 01. One advantage of using a top-down search is that it is comprehen- sive. Every generated state appears on a completed top-down tree. Bottom-up searches, however, have the advantage of delivering all and only the possible derivations for the solution state, if there are any. A major consideration in determining whether or not a top-down search or a bottom-up search will be more e ffective is the associated branching factor. The branching factor of a tree – the average number of descendants below each node – determines the complexity of the tree so ideally we want to minimise it. As Figures 11.1 and 11.2 demon- strate, the branching factor of a top-down search in a system will often be distinct from the branching factor of a bottom-up search. In the case of [BIN], a bottom-up search proves to be more e fficient. 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