Tuesday, 31 January 2017

More on Machine Intelligence...




What is intelligence? Is it possible for a machine to have it?  Does it involve consciousness? Can there be thought without emotion? There are many more questions that can be asked.

If the Turing Test really is valid and reliable, how could we even tell if a machine has the qualities of sentience?  How would we know if it has self awareness? We might discover it doesn’t; that seems a lot easier than determining that it does.  This is a rare time when proving a negative seems more feasible than proving a positive. It seems to me that a machine could be programmed to imitate a human well enough to pass a Turing test.  Ask it about the feelings it has when listening to the Pastoral Symphony…..it could respond that it doesn’t enjoy music. How would you know from that answer that it is a machine? Suppose you asked for advice about a complex human relationship, with the idea that the answer would reveal whether the respondent is answering from a human understanding of human connections and motivations. Could you really tell from the response if the answer is human?  If it was a valid answer, maybe it was from a machine that has been programmed in behavior. If the answer was stupid, maybe the respondent just hasn’t had enough experience to formulate a wise response.

What is consciousness?  Awareness, yes. But what is that?  To me it has something to do with being able to reflect on and consider what is going on both outside and inside the one that is conscious. In a sense it is like being a mirror of those things, but if a machine was programmed to analyze such data and to act according to its analysis, is it really sentient? It is one thing to respond a certain way because of a programmed response that determines that “If X, then Y” and quite another to draw a conclusion based on understanding and then deciding that the understanding dictates that a certain response be made.

Do we ever make any points in a discussion without having emotion? Is it not the emotion that impels us to make the points? If we were completely without feeling, would we even bother?  Can a machine have emotions? 

Monday, 30 January 2017

4 Worms



A minister decided that a visual demonstration would add emphasis to his sermon.

Four worms were placed into four separate jars. 





The first worm was put into a container of alcohol. The second worm was put into a container of cigarette smoke. The third worm was put into a container of chocolate syrup. The fourth worm was put into a container of good clean soil.


At the conclusion of the sermon, the Minister reported the following results:


The first worm in alcohol - Dead. 





The second worm in cigarette smoke - Dead 






Third worm in chocolate syrup - Dead 






Fourth worm in good clean soil - Alive. 






So the Minister asked the congregation -



What can you learn from this demonstration?



Maxine was setting in the back, quickly raised her hand and said,




'As long as you drink, smoke and eat chocolate, you won't have worms!'


End of sermon.




Sunday, 29 January 2017

SRED Question 3 Continued



In a recent post (http://gordon-feil-sred.blogspot.ca/2017/01/question-3.html) we looked at the third question about a project to which CRA requires the answer to be Yes for the project to be considered as SR&ED.

Part of that question addresses experiment and analysis. Perhaps the easiest way is to quote directly from CRA’s website.

Experimentation and analysis are approaches used to investigate hypotheses. Experimentation involves structured and organized tests and studies to obtain information in order to address the hypotheses. Experimenting involves not only testing and analyzing but also exploring the relationships between tests, explaining the results as they relate to the hypothesis, drawing conclusions, proposing a new hypothesis, or conducting additional tests. Such experimentation can include work on the evolution of prototypes or models.

Analysis is the detailed examination of information to differentiate the various parts of a whole, determine their attributes, or explain their relationships. It is performed against the background of available knowledge and experience and it involves using tools such as models, graphs, statistics, tables, diagrams, mathematical formulas, and computer programs to express this knowledge or experience. Analysis is an integral part of a systematic investigation or search and it can be used to generate or test a hypothesis.

Experimentation and analysis are not to be confused with “trial and error”. The purpose of the systematic investigation involving experimentation and analysis of the outcomes is to obtain understanding of relationships between the factors involved. Trial and error addresses What.  SRED addresses Why and How.
CRA expects that this work has been done by people qualified, by reason of experience or formal training, to perform the experimentation and analysis.