Archive | March, 2018

114. Brain, AI, and Behavior (3rd Revision)

13 Mar

This blog was stimulated by a Ray Kurzweil newsletter topic:
               Will artificial intelligence [AI] become conscious?
It reminded me that I have been wanting to explain this and related topics more thoroughly. Before continuing, I must describe my qualifications related to the conclusions that I will draw. I have had considerable formal training and professional experience in the following areas:
1. Behavioral Science
2. Neuro-science
3. Computers and control systems
4. Advanced computer programming

All of these topics are related and the relationships are illuminating. I have divided my ideas into several topics:

1. “Consciousness” is a layman’s term but is also used by scientists outside the field of behavioral science. It is most often used in a vague way without clear definition. And when defined, the definition is often made with vague statements. Many years ago, P.W. Bridgman (The Logic of Modern Physics, 1959) advocated “operational definitions.” One should use terms that can be defined in terms of specific procedures. For example, “hunger” could be defined as 24 hours of food deprivation. Another, “meter” is the length of the path traveled by light in vacuum during a very small, specified time interval: 1 over 299,792,458 sec.

Instead of the vague “consciousness” we should use terms like:
“Aware:” meaning there are measurable responses to specific type of stimuli.
Asleep:” defined by measurable patterns of EEG, and breathing patterns.
“Coma:” lack of responsiveness, but not asleep or under drug influence.
There are many similar terms that can have precise definitions. One can find numerous discussions of “consciousness” that go nowhere because the terminology used is not precise or “operational.”

So, here is my answer to the above question: Will AI become conscious?
My answer is that there will be amazing developments and uses for AI, but it will never exactly duplicate the capabilities of the human brain. Our brains developed over millions of years of evolution and have abilities that are not likely to be completely imitated. Throughout his lifetime and responding to all his experiences, a person’s brain develops by adding new structures, new neurons, and billions of new interconnections. Could this changing, adaptive system, with many trillions of connections and chemical operations, ever be duplicated by humans. (See below for details.)

2. How does the Brain work? Using all the knowledge areas mentioned above (behavior, neurophysiology, computers, etc) I will make the following description. First, the processing ability of the person primarily depends on the brain, but also includes other parts of the nervous system, and other systems, such as hormonal, sensory, and muscular.
The overall system is much like an ordinary computer, with keyboard and scanner inputs, a central processor, memory, and outputs such as a screen, printer, and speakers. In humans, a wide variety of sensory cells (receptors), such as cells in the retina, provide inputs, the nervous system (mostly brain) provides processing and memory, and outputs are complex behaviors, reflexes, hormone production, vocalizations, etc. An interesting fact is that even spurious factors like viruses work in the body and computers, in very much the same way. In both cases, they use the normal processing features to reproduce themselves and to cause damage.

Manufactured processing systems are pretty familiar. Most interactions are based on wires that carry electrical charges (+ and -). In humans, the wires are nerves that transmit over distances using the motions of ions in a wave process, much like a fuse. Ions are tiny charged (+ or -) particles composed of elements such as sodium, chlorine, calcium, potassium, etc. The ions move sideways to the direction of information flow, much like a tsunami moves in a wave without transferring the water itself. The moving wave that transmits info is called an action potential.

Our nervous system is composed of billions of nerves with around 150 trillion interconnections called synapses, and other connection variations. Further, each synapse (which functions like a transistor)  has a complicated and variable structure.  The nerve cells, their branching structures, and connections, provide all of our simple and complex behaviors. Frequently used connections associated with “learning” often expand and acquire new protein components. Functioning of these cells can also be modulated by various hormones, chemicals and drugs. So, our brains are a gigantic system with a number of control points so large as to be incomprehensible, that evolved in several billion years (also incomprehensible) to a structure that can create abstractions like, Einstein’s Relativity, and can ask where did I come from? It is also important to note that although the brain is complex almost beyond comprehension, it is still composed of chemicals and processes governed by the man-made laws of physics and chemistry. It is very unlikely that these totally “deterministic” components can produce any “free will.” In support of this conclusion, we know that computers (unquestionably deterministic) can produce amazing “behaviors” and can be programmed to imitate something like the assumed human “free will.”

We understand and know how the brain and spine produce simple reflexes using the input, output and processing systems described above. Not yet described here are more complex functions like memory retrieval; logic and reasoning; “creative” actions; and “emotions” like love and anger. It is clear that our brains can do a wide variety of things and has specially evolved to implement those most related to survival and the achievement of reproduction.
We know, for example, that special parts of the brain are devoted to facial recognition, to strong emotions, sex, visual memories, and the fight/flight response. We know that the brain can group together a series of actions or things and can rapidly produce a whole learned series without separately retrieving the components. There are experiments in “learning to learn” where if one learning process is similar to another, there is a facilitation. Really good brains can produce valuable associations and retrieve deeply “buried” little used, but relevant info. Brains have a remarkable ability to search, summarize, and draw conclusions. We do have some idea how these remarkable processes can take place, but much of this is purely speculative. Yet, the fact that computers can be programmed to do much of this abstract work, supports the idea the even the most amazing actions are “deterministic” and ultimately predictable. Also supporting determinism is that the huge number of anatomical and functional studies of the brain have never disclosed any super-natural “free-will” elements. The argument that free-will could “emerge” from deterministic elements, seems unlikely to me, but in the end, determinism forces us towards certain conclusions. Personally, when I really examine my life, I see that all my current behaviors are the result of a life-time of experiences.  I must ask free-will advocates: if your current behavior does not come from your DNA and past experiences (learned, imitated, stored, etc), where does it come from?

3. Thinking
There is one more topic that should be mentioned: “thought.” What is thought? Is it a behavior? Does it precede all overt behaviors? Is it “neuronal” like other actions? What is its function? Etc. Based upon some behavioral science studies and my own intuition, I propose the following.

First of all, most behavior just occurs without any thinking or planning. Second, thoughts can be words, pictures, or even “feelings.” Thoughts are studied scientifically by using a subject’s verbal responses, which ARE observable.
Thought is a covert brain output that does not reach the status of observable. An interesting facet of this idea is that some people “think out-loud” and what should be covert isn’t. I have known several people who do this. The most likely and useful aspect of thinking, is to produce a sub-threshold behavior to test its effect before causing the thinker any problems. For example, you ask your boss for a raise in your head, with different wordings, to find the best version. Or, you imagine yourself climbing a mountain and you note the fear that it generates. Thinking allows you to try things out before you actually do them, and serves as a safeguard.

Under the heading of thinking, one could imagine advanced retrieval processes that would be important for developing a theory or concept. A thought could be stimulated by an event in the environment. You see a stranger that looks like a past friend and a thought about the friend emerges. Clearly, there are environmental events that elicit related thoughts, but maybe there is also a thought generator, based upon the relative importance of stored info. Do we have some sort of scanner that finds important or otherwise significant items to think about?

Final thought: Even though our brains are extremely complicated and likely can never be duplicated, downloaded, or fully understood, brain research can still be productive. Studies of brain inputs and outputs, small systems of nerves, and comparisons with computers and other control systems,  have yielded valuable insights as to how higher functioning is accomplished.



113. Do Good Candidates Make Good Officials?

13 Mar

For several years, I have been concerned about the nature of our USA election process. The traits that promote election, are not necessarily the same as those required for good governing. Successful candidates can have any or all of the following:  1. rich donors,  2. a powerful appearance,  3. an ability to convincingly lie and exaggerate,  4. an ability to say what people want to hear, such as subtle racism or anti-science sentiment.
Emerging from this process may be leaders lacking necessary experience, temperament, intellect, and/or decision-making skills. Focus is often on pleasing wealthy donors, instead of making life better for our citizens. The recent “tax reform” bill is an example of this.
Our current election process was formed by legislation, judicial decisions, and by questionable traditions. It needs revision.  We could start with election campaign laws such as those giving rich donors almost complete control over the election process. It would also help to have more voter education. There are no easy ways to reform elections, but we should at least make every effort to do this.