Pat, we all agree that the AI today is totally different than the AI in earlier dates. In earlier dates, for example, AI had to use LISP to build Expert Systems to do anything useful. Expert Systems were hard-coded and were not learned, and no one is using that at all. Today, we want everything to be learned, particularly by neural networks (deep learning, as function approximators) which were not available before. Of course as I pointed, the infrastructure for doing AI is not Machine Learning, but we don’t call the infrastructure (such as cloud) as part of AI. I think you agree that If someone does not learn Machine Learning, there is no way for him/her to understand the essence of computer vision, autonomous vehicles, Deep  Reinforcement Learning(optimal control with robotics) , etc., as all models are learned today. As you know, Deep Learning has been the center of AI since 2012. The concept of DL is based on ML, with the only difference being the function approximator. The point is that Machine Learning is the core concept to understand if students want to learn AI.

Sent from myiPhone

> On Sep 24, 2019, at 4:10 AM, Pat Langley <xxxxxx@gmail.com> wrote:
>
> Frank,
>
>> The difference between traditional programming and AI is that, in
>> traditional programming approach, we manually set all parameters
>> and actions by imaging all possibilities. In AI, the machine “learns”
>> parameters by finding the patterns in datasets (numerical data,
>> images). Here the parameters refer to the coefficients or weights
>> which determine a “model”. A Model is for us to make predictions
>> after it has been well trained with datasets.
>
> Thanks for your instructional efforts, but you may want to read a bit
> more about the relationship between AI and machine learning. Patrick
> Winston's AI text would be a good place to start:
>
> https://www.amazon.com/Artificial-Intelligence-3rd-Winston/dp/0201533774
>
> This has chapters on learning but also on many other aspects of AI.
> You may even want to read my 1995 textbook on machine learning,
> which covers more paradigms than you seem to know about:
>
> https://www.amazon.com/Elements-Machine-Learning-Morgan-Kaufmann/dp/1558603018
>
>> It is common for non-AI professionals to think anything with automatic
>> functions (such as automations, robots) is AI. Right now, we have
>> formed many branches in AI such as Supervised Learning, Unsupervised
>> Learning, Reinforcement Learning, but the core concept is always
>> machine learning. Without machine learning for various models, this
>> is NO concepts of AI. We need to make sure what we are doing.
>
> I assure you that I'm not confusing traditional programming with AI,
> but also that not all AI is machine learning.
>
> I helped launch the field of machine learning in the 1980s and I served
> as first editor of the journal Machine Learning. I think you may want
> to read more about the history of the field.
>
> Regards,
>
> Pat Langley
> http://www.isle.org/~langley/
> http://www.isle.org/~langley/cv.pdf
> https://scholar.google.com/citations?user=5JcRw-sAAAAJ
> To unsubscribe from this list please go to http://www.simplelists.com/confirm.php?u=91ZLBc54XJq6iN3fCxNKGCcxZ2J9Qrpy