@Cuchullain. thank you for your feed-back. Some answerI browsed. Some constructive feedback:
1 It's quite chatty/bloggy style (not necessarily bad) and might appeal to certain reader group.
2. No maths
3. Many 3rd party links (some are broken, again)
4. Well-written technical articles with begin, middle and end + CONCRETE EXAMPLES!
5. No cats, please.
My personal preference is 70-80% maths and 30-20% text/bla bla.
That is a very good initiative Daniel. Some quick thoughts:I like RKHS for several reasons(hard [$]\epsilon,\delta[$] analysis, Cauchy sequences, applied Functional Analysis and not much hand-waving and hot air and applications to ML.)
I am thinking of a 'compact' project for a student who is taking this course.
https://www.datasim.nl/onlinecourses/10 ... oundations
https://www.datasim.nl/application/file ... ations.pdf
Once he gets to Part F he could start on a RKHS-meets-ML project.
Any ideas, JohnLeM?
I see. So you do work with a University to give these lectures ?Sure, you have my email address. It sounds very interesting.
Actually we take the critical path B->C->D->F. I customise to suit the end game:-)
// Actually, this is the first 3 years of a maths undergrad degree program, slightly watered down while keeping the punch lines in place.
I do work with several unis' students who take my courses that I built with my own 2 hands (the courses, not the students).
True Ok, I'll try this. Lets go private now.Sounds good. I found you
I sent you an aswer at your datasim address !Sounds good. I found you
BTW I wonder how many DSers know all these spaces?