AI Meets Blockchain! | Learning Intelligence 5

Hey learners, what's going on? Welcome to Learning Intelligence Episode 5. Today, well, this week is all about me going through Project 2 of the Udacity Artificial Intelligence Nanodegree. I'm actually planning on submitting it tomorrow. So stay tuned for whatever the rest of the week will hold. So you see over here, we've got this little test result summary and see all these dots, that means the tests are passing. I've got 2 more things to do here. I've got a heuristic analysis PDF to do and a research paper review. And if we go up here, I'll show you this, my custom function, 2 is winning. It wins the game 72% or 73% of the time. So that's what I'll be doing tomorrow. I'll be finishing up the final section of part 1, which is to write an analysis on why I chose my heuristic functions. And I'll put all my code and analysis and whatnot on GitHub as well, so you can check it out. I'll put the link in the description. And also part 2, which is writing a summary on a research paper. And I chose a research paper that DeepMind and Blizzard put out. I'll put the link in the description as well. And I spoke about in the last episode how they released a StarCraft Graph2Environment to practice reinforcement learning. And I'll publish that on my blog and whatnot, and you'll be able to see that. Just a one-page review on why they did it. By the way, reinforcement learning is amazing. I'm loving it. The ability to sort of let an artificial— it's like raising a child or training a pet, right? It's like letting an artificial intelligence agent into this environment so it can learn on its own. That's really exciting. I wanna show you one more thing that I think is— I'm incredibly excited about as well. I think has a potential to change up a lot of things. It put power back into the hands of developers and learners like you and I. This is OpenMined, openmined.org, and I can't explain it fully, so I'll put a link to Siraj's video in the description. But essentially what I gather it is, is it's decentralized data combined with artificial intelligence and blockchain. So right now, all of the data in the world is sort of Facebook and Google and all these big, big internet companies, they have control of all the data, right? And what OpenMined hopes to do is decentralize the data so that you and I have full control of our data Because right now we sort of sign up to these companies and we trade off our data for free and they use it to build amazing products. I'm not saying they're evil, like Google's motto is "Do no evil," but we essentially lose control of our data and we can't gain value from that. So this is what OpenMined is targeting to do, is bring control back to individuals' data and allow developers and whatnot to access these mines of data so they can build products as well and bring value to the world. But as I said, I can't explain it fully yet because I'm still learning about it. But I'll put Siraj's video somewhere, maybe here or in the description or something like that. So Siraj does a much better job than I do at explaining it. We did it! Check it out, officially submitted the second project in the Udacity AI Nanodegree. I just went for a workout to like reward myself because I've been studying all day inside. Actually, it's raining outside, so it was like a perfect day to sit inside and do some learning. I submitted my artificial intelligence game-playing agent to play a game of Isolation, and now I could have probably improved on my evaluation heuristics. Mostly The reason why I didn't sort of keep going on them is because first of all, the project was due today. And second of all, I'm having trouble understanding the concept of evaluation heuristics. So I've got to keep learning more about that. Otherwise, part 2 of my— I'm pretty proud of this part actually. This was part 2. So a review on a research paper. I ended up going a bit over, right? Like it said to be 1 page and I did 2, 2.5 pages. I was excited, but I'm pretty proud of that work. That'll be up on my blog or something. I'll put the link in the description, by the way. Otherwise, I was practicing machine learning on DataCamp and I'm getting a whole new perspective of it. I know I've done some machine learning courses in the past in the 100 Days of Code. It was Andrew Ng's machine learning course on Coursera, like the, the course that started Coursera, and that was amazing. It was kind of advanced for where I was, and it's good to revisit some fundamental concepts in machine learning like clustering, supervised learning. I'm learning more about scikit-learn and pandas and things like that, which are Python libraries for machine learning. I'm really loving it. I want to show you something. So I submitted my project last night, submission successful, and look, I don't know if you can see it, but it's 6:33 PM when I submitted it, and if we go here Look at this, review available 7:25 PM. So that means within an hour my project was fully reviewed. A project that I spent probably close to 15, 20 hours on is got fully reviewed by Udacity reviewer. These guys are legit. They know their stuff. And I'll show you the review. Look at this feedback that I got. So I got brilliant learner. Yeah, a little boost to the ego there. The project submission was very impressive. In fact, it's almost perfect, but I've got some few, a few things that I need to change. Nonetheless, you got some awesome, awesome code review as well. So this is all the code I submitted. We go down here. So I need a change to this section of the code. And I got two awesomes. Look at that, thumbs up emoji. And another one. So that'll be my goal for today. Redo or fix up the changes that I need to implement into my Udacity Project 2 assignment, or AI Nanodegree Project 2, and upload that. And then hopefully I'll pass from there. Oh yeah, we did it. Second submission of the Udacity Artificial Intelligence Nanodegree Project 2 Lots of 2s here. Submitted the second project for the second time. Yes, that's it. It took me almost all day because I was sort of going above and beyond. I want this project to— this submission to pass. I was a little bit disappointed in myself that the first one didn't because I thought it would. I guess I got a bit lazy for the first submission because I'd worked so hard and I was like, yes, I'm going to submit it now without sort of going back over it, making sure it was bulletproof. I do a lot of that in my works actually, like with my writing and whatnot. I really have trouble going back over what I've actually done. That's a weak point for me. So note to future self, go over it again. I think this is the title of my, one of my last videos, go over it again. It's like I get this feeling that once I've done something, like that's it. I don't sort of revisit it and think about how I can improve for the future. What else have I been up to today? Well, I donated blood. As you can see, I've got a bandage around my arm. Amazing cause, it's free, painless. Well, if you can withstand a mosquito bite, you can withstand donating blood. It takes 5 minutes and you have the potential to donate I mean, save 3 lives. My blood is O negative, which I think is universal donor. So I do that. I do it every 6 to 12 weeks or so. I can't really put a date on it. I get messages from the donation place that I donate to that say, "Your blood's been used at this hospital." And that's really amazing. It goes to people who need it. So life-saving operations and whatnot. That's a plug for donating blood. If you've never done it, give it a try. It's painless and you can help other people. While I was training my algorithms to, or while I was testing them, I was reading this book here. I'd highly recommend it. Tools of Titans by Tim Ferriss. I'll put the link in the description. The tactics, routines, habits of billionaires, icons, and world-class performers. So essentially Tim Ferriss has a podcast where he's interviewed over 250 of the world's best top performers, world-class people successful in their domains. He's turned all the sort of best bits from each of them into a book. And so I was reading that while I was testing my algorithms and it is just packed with knowledge. I highly recommend that book. Highly recommend donating blood. And we'll see if this project passes or not. I think it should. Yeah, we did it. Check it out. Passed the second project in artificial intelligence nano degree program. Meet specifications. And look, the reviewer even called me brilliant learner. Build a game playing agent. Project 2 in the bag. And now it's time to get onto project 3. What else did I do today? I worked on DataCamp. I did a little module there. It's, I'm in the machine learning course of DataCamp or machine learning track. It's proving to be a bit harder than what the Python intermediate track was. And that's understandable because a lot of this stuff I haven't really covered. Because using new libraries, I'm excited actually. We're using scikit-learn, using NumPy, using Pandas, finding more about this stuff here. And otherwise, I'm going to go play netball now. I'm going to come back tomorrow and get into the next set of classes on Udacity. I'm not entirely sure what it will be. If you're wondering why I'm a bit sweaty, it's because I just went for a run to finish up the day of study, finish up the day of learning. I still got some writing that I'm going to do, but as for studying artificial intelligence, I think that's, that's it for the week. What did I learn today? So I started on lesson 10 of Udacity's artificial intelligence nano degree, which is on SIRK. I'm about halfway through the class at the moment. It's been taught by the amazing Peter Norvig, who also authored this textbook that I'm using for the course. Actually, this book, must, must recommend, How to Win Friends and Influence People by Dale Carnegie. I've learned so much of interacting with people with that book, but Peter Norvig is the author of, or one of the authors of, Artificial Intelligence: A Modern Approach, as well as a researcher, one of the lead researchers at Google, as well as an instructor on this artificial intelligence nanodegree that I'm doing. From the search class, I learned about 3 main algorithms so far: breadth-first search, cheapest-first search, and depth-first search. Depth, depth-first. So breadth-first will go, it'll find the shortest amount of steps to the path. Cheapest will find the, uh, the pathway that takes the least amount of distance or steps or computing power. And depth-first will go for the deepest search first. So in terms of, like, let's use a map example. If you're wanting to plan a route or using artificial intelligence to plan a route from one location to the next, uh, you could either use breadth-first search, which will find the route that will take the least amount of steps, or cheapest first, uh, cheapest, cheapest first algorithm, which will find the one where we'll have the least distance, or depth-first search will go for the longest distance first. So it's not probably— depth-first is not optimal for a route-finding algorithm. I have one more thing to show you before this episode of Learning Intelligence is over. Where is it? I really got to remember to take off that zoom. But I stumbled across this the other day when I was going through reading about artificial intelligence on Medium. If I can find the article that I read that linked this, I'll put that in the description so you can read it too. But otherwise, it's the Artificial Intelligence Playbook by Andreessen Horowitz, which is a massive VC firm. And their motto is software is eating the world and machine learning is eating software according to Andrew Ng. And AI gives you software superpowers as well. So they've released this playbook to educate people like you and me who are learning about artificial intelligence as well as anyone who's involved in the world of technology who wants to know where AI can be used, how it's implemented, from the ground up essentially. It's not gonna be sort of the one-stop shop for learning things about AI, but it's certainly, I'll put a disclaimer here, I haven't looked into it yet, I haven't read it myself, but from what I gathered from it, it's gonna be a good, good starting point if you're looking to learn more about it. So I'll link that in the description. But otherwise, that's all I have for this week. Next week I'm gonna be continuing the Udacity artificial intelligence Nanodegree, as well as the machine learning course on DataCamp, and anything else that I manage to stumble upon. But thanks so much for watching, and if you wanna see anything in next week's video or in a future video, leave a comment below, and we'll see you then. Peace.