Hi, the brightest bearers of reason!
If you (like me) have children then probably you are going to dedicate lots of time contemplating which kind of education they need to be in tune with most important things in the world. Right?
I believe that the subject which the community is involved in is most important one and actually is definitive for the future itself.
I am personally a big fan of HTML-school. But is there any a formal education similar to it already, which may be taken into account for getting a job or to be used in some research area in the nearest future? I mean something on the intersection of neuroscience and IT. I mean something that is considered like a right way towards the building AI accordingly to HTM approach.
Have you any ideas?
Hi, the brightest bearers of reason!
HTM is an interesting approach and one in which I am greatly interested. I agree that the biological evolved neocortex should serve as a fundamental basis on which to build future AI. That said, I think that there needs to be an investigation of other approaches as well.
So, if you are interested in providing an academic path to future students, it is best to let them study the area and choose topics and class in each of those. Most well rounded universities have a wide enough selection of classes in neuroscience, brain anatomy, and computer science to provide a sufficient background to get you started. Maybe fields like mathematics and philosophy might bring an additional perspective.
Given that computational neuroscience is such an active and new science, a student has to be constantly learning both what is happening with the HTM approach, as well as alternate formulations, even though doing so might be counter-productive in the short term. Learning about DeepMind, Allen Institute, European Brain Initiative, and CBMM might yield some fruitful applications too. It could happen that the HTM approach, while as biologically accurate, might be too difficult to implement at larger scales within the next several lifetimes.
I wonder if HTM is enough. Should the model also encompass parts of the thalamus and hippocampus in order to be useful? I don’t think that we know enough at this point, so it will probably take a lot of investigation before we converge on a system design of the brain.
Personally, I am trying to learn HTM and computational neuroscience late in my life. Jeff Hawkins ignited an intellectual curiosity in this field that was not as previously active (as my focus was in computational bioinformatics, mathematics and physics). Back in grad school, my adviser (Arwel Evans) said that the brain will be the next frontier of mathematical research, although I did not think that that would happen within my lifetime. I run a Machine Learning Meetup which helps keep me intellectually active and focused on NuPIC.
I’m not an expert on this at all, but here’s my advice to someone entering college. Knowing how to program is required to build AI, and it ensures you can get a job fairly well. Neuroscience is a broad field, so make sure you’re okay with spending a lot of time learning about chemistry, doing lab work, and other things not directly related to AI, at least in college. My college requires 6 semesters of lab work for the neuroscience major, and you can’t start neuroscience classes until completing the prerequisites, which takes until the second semester of the sophomore year. Basically, just make sure you don’t dislike pure neuroscience if you go that route. The same goes for computer science.
Since you’ll have a specific interest in AI, and your college might not have much related to AI, you should find people who are working on AI or something closely related. For example, you could do independent study with the support of a professor, get an internship, or maybe help a graduate student. Again, I’m not an expert, so you’ll want to look into this more. You just want to make sure that, if you self-educate outside of class, you have some sort of guidance.
My plan is to encourage my kids to be curious and teach them to be resourceful (not require information to be spoonfed to them). Hopefully that will be a good foundation for them in whichever career direction takes their interest.