7. Machine Learning Tasks and Types

Machine learning is typically broken up into 4 types: supervised, unsupervised, semi-supervised, and reinforcement learning. But is this all?
In this video, start by defining artificial intelligence, machine learning, and deep learning. We then cover the 14 tasks and types of machine learning including: supervised learning, unsupervised learning, semi-supervised, reinforcement learning, self-supervised, multi-instance learning, inductive learning, deductive learning, transductive learning, multi-task learning, active learning, online learning, transfer learning, and ensemble learning.
0:00 AI vs ML vs deep learning
3:45 4 basic types of ML
9:15 unsupervised learning
12:41 reinforcement learning
16:38 semi-supervised learning
18:34 self-supervised learning
20:35 multi-instance learning
22:30 inductive vs deductive vs transductive inference
27:20 multi-task learning
29:37 active learning
33:45 online learning
37:46 transfer learning
39:50 ensemble learning
Check out the whole materials informatics series at • Materials Informatics with workbooks and course notes available at github.com/sp8rks/MaterialsIn...

[@martinmckee5333]

Possibly the best introduction to various machine learning techniques I've seen. There were even a couple I hadn't seen before. I always love learning new approaches!

[@TaylorSparks]

Dude, feedback like this makes me so happy! I will have new videos in this series to release very soon. Just waiting on my students to publish a pip install version of the composition-based feature vector

[@martinmckee5333]

@@TaylorSparks Absolutely, credit where credit's due.
I'm coming from AI and control systems in Computer Science, so I've got a grounding in the machine learning aspects, but I'm loving the material informatics focused content. And, as I said, even the basics are a better introduction than you find elsewhere.
Cheers!

[@TaylorSparks]

@@martinmckee5333 rad. Spread the word!

[@lidiyapriyadarsinik2815]

Awesome resource. Your videos are becoming my primary source of learning by helping me get over my programing phobias. Thank you so much 🙏.
Could you please mention some supporting reading materials or add them in description of the videos.

[@TaylorSparks]

Great idea! Will do.

[@rishabhsharma8743]

This is an amazing initiative Dr. Sparks. Thank you for sharing such stuff with us.

[@ericterrywang]

As a PhD student trying to come up with some informatics project in nanotech, your videos truly inspire me.

[@TaylorSparks]

That makes me so happy to hear!

[@TheMitchyevans]

Good stuff Dr. Sparks. While watching, I was considering what types of learning I was practicing.

[@TaylorSparks]

Online learning, albeit of a different type than what we discussed here!

[@richardaversa7128]

Hi Dr. Sparks, I am working my way through your ML series and it is incredibly valuable. The data visualization you built for thermoelectrics around 11:52 is very nice. Did you construct that in Python? Would it be possible by any chance to share some of the code? Thank you very much for the high-quality content you share with the world.

[@TaylorSparks]

I wrote it originally in matlab, *shudder* and then we ported it to JavaScript. I've never redeployed it in pythons since

[@richardaversa7128]

@@TaylorSparks Oof, thanks for your reply.

[@RD-fs7en]

you'r allowing ads on a lecture series or its youtube invasion?

[@TaylorSparks]

I allow them. I hope you don't mind. I put a lot of time and effort into this series so if you're willing to watch an ad then that helps me get reimbursed for my effort. Online learning resources, like UA-cam are sort of like a modern-day textbook. Instead of getting royalties for the sales of textbooks academics can now get royalties through ads on their videos.

[@RD-fs7en]

@@TaylorSparks hi there, well, i think differently --not that i am here to fight, just have my reasons. Maybe we discuss once we meet 😊
Despite our disagree, very nice work, i enjoyed watching 👍

[@TaylorSparks]

@@RD-fs7en hope we meet soon ;)