The Turing Test

1. Is the Turing Test a sufficient test? That is, if a machine passes the test, would you agree it is intelligent?

It depends on how we define intelligence. If intelligence is all about making judgments based on logical deduction, then passing the test might serve as an evidence of the machine being intelligent. However, human intelligence is supposed to be much more inclusive than this. For one thing at least, it should be able to cope with variations on its own when an encountered situation is not preprogrammed. An interesting example would be, on the “talk with God” website, when I typed in “Did you have breakfast?”, it replied as “no, I don’t have any breakfast, but I do have some gossip”. Apparently, the phrase “did you have something” is misunderstood here as “did you possess something”. While programmed perfectly for many other responses, when it comes to the transformation of a verb under certain context, the program could not generate appropriate responses for this variation. Although a machine or program similar to this might pass the Turing test satisfactorily, it does not meet the broader requirements of being intelligent, such as self-generated adaptation, etc.

Also, it seems to me that the analysis skills the test requires resemble what the many data analysis software can do nowadays. There is one major difference that some preprogrammed questions are asked. Other than that, the way it collects the responses, analyses them and provides conclusions is just similar to some statistics software. But we are not considering those software as intelligent.

2. Is the Turing Test a necessary test? That is, does a machine have to pass this test in order to be intelligent?

If a machine can pass such test, it is reasonable to believe that it has certain ability in information processing related tasks. But if the machine fails, it is also unfair to conclude that it is not intelligent, since even humans can make mistakes in tasks like this. In addition, I don’t see this test as a perfect simulation of the problems we come across in the real world. Therefore, this test need not be viewed as the criteria for intelligence.

3. Will a machine ever pass the Turing Test? Why or why not?

The test is not requiring highly generic information processing skills. So if the machine is equipped with programming that enables it to analyze the responses it encounters, passing the test should not be a hard thing.

4. Will a machine ever be intelligent? Why or why not? (This may or may not be the same answer to the previous question).

I’m convinced that machines will be intelligent eventually. While I can’t really think of any sci-fi like and inspirational ideas to support this argument, I would like to point out a very fundamental common character machine and human mind share, that is, despite all the other differences, they both run on electricity. The imitation of the passage of electrical signals should be successfully accomplished in the end.

Also, breakthroughs in at least two areas are necessary for this dream come true. One is the mechanism of how human brain is functioning. This area has been vexing scientists for ages, but humans should be able to know all about themselves finally. The other could be the appropriate medium or programming methods used by machines. It is highly possible that chips are not the most appropriate material that machines should be using when it tries to perform intelligently, or that the programming methods we are using right now has already limited some possibilities of machines’ being intelligent.

week 2 How the mind works differently from computer

I still remember how surprised I was when I learned about the fact that the information processing theories originate partly from the comparison of human mind to the way computers operate. To me, human mind works in a way that is much more complicated than computers do. And this is probably why no software or programs so far can simulate perfectly the thinking process.
One of the most distinguished and not likely to be imitated features of human mind is that it is capable of analyzing and synthesizing available information and then make justifications to approach to the more desirable results. Computers on the other hand, do not seem to have such synthetic function of appropriating operations and results on demand.
Taking the task of searching for journal articles on the library website for instance, human’s role of monitoring is still indispensible in spite of the advanced technology of the search engine. First, one needs to decide the appropriate key words to use. In doing this, the mind has to speculate the possible representatives for the targeted topic, such as “story reading” “child” “literacy” for the subject of how reading books to children influence literacy development. Numerous results might be generated after these entries are typed in. Normally, only a portion of the returned results are useful while the rest merely share the same words that were typed in. Thus, the mind is relied upon to filter out the unwanted ones. This is indeed a tricky thing to do, since the search results are involved with a vast quantity of information. But as it appears, within a short amount of time, humans are able to make the decisions regarding which entries are useful and if nothing works, how the key words should be modified and changed to get better results.
I suspect that this process is done in a sequence similar to this: the mind retrieves from memory the concepts, theories frequently discussed about on a certain topic, and a concept map is formulated, afterwards, the search results are compared to the map to see if they match up. One thing noteworthy is that, the contrast is not done at the very detailed word by word level, but is at the level of more structural comparison of the general meanings conveyed beyond the text itself. Not surprisingly, even the fanciest computer nowadays is not able to perform at such a highly abstract level. What enables human mind to carry out such operation is not quite clear yet. However, if I am allowed to have a wild guess, I would suggest that the human mind stores and retrieve information by chunk. So when an operation is under way, our mind would retrieve and utilize the grouped information in chunks, compared to the digit by digit calculation used by computers. This chunking feature of human mind reduces the necessary processing procedurals to the least, saving up large amount of spaces in attention and memory. Such simplification of operations is highly important, since we have limited space for attention and working memory. However, such structural simplicity is at the expense of accuracy. Consequently, the human mind might not be perfectly accurate but is capable of doing complicated synthetic work, such as filtering out the unwanted search results based on contextual or syntactic clues, without much difficulty. Computers on the contrary, can only provide word by word search results despite of its ability of doing large amount of calculations.