In this FiveMinuteFriday episode of the Super Data Science Podcast, we talk about how emerging techs like machine learning, AI or deep learning are embedded in almost everything around you.
Are you even aware of it? Today, we find out how beneficial are they to you!
Let me guess, most of you always feel the urge to reach for your mobile phones for more than a few times in a single day. Did I get it right? Well, no shame in that. Phones are becoming more and more a necessity because of its features and advanced technology.
Today, I talk about mobile applications that use machine learning or artificial intelligence. There’s really nothing sophisticated about these but what’s astounding is they are used by a lot of active users. So if we focus on the serviceability to a large audience, then that’s a very good feedback alone for the incorporation of machine learning in these apps.
It’s amazing that there’s a bunch of applications out there heavily utilize machine learning, AI, and deep learning. Apps like Gmail, Google Maps, Waze, Spotify, Netflix, etc., analyze your usage pattern and give recommendations based on it. Running out of activities to do? Tailored suggestions will be given to you. Replying has become tedious? Autocomplete is there for you. All our tasks, may it be personal or work related, can be made easier and more convenient using these apps in your phone.
It’s nice to see that people are becoming aware of how they can benefit from these innovations. But, we as data scientists should level up how we look at this. Since we know that these kinds of developments are happening, then we should be leading the way on how to improve them and create more for the benefit of the greater community. Get inspiration, consolidate the best ideas, and figure out how to avoid future flaws. Always aim to make something better out of what’s already in the market. Teach yourself and learn more if need be.
But mobile applications are just a small part of a bigger picture. Explore everything you could improve with tech. Look into the biotechnology, assistive technologies, etc. and be more valuable. The possibility is endless.
ITEMS MENTIONED IN THIS PODCAST:
DID YOU ENJOY THE PODCAST?
- What technology features have amazed you lately?
- Download The Transcript
- Music Credit: Candyland by Tobu [NCS Release]
This is Five Minute Friday, episode number 206, Machine Learning is All Around You. Welcome back to the Super Data Science podcast. Ladies and gentlemen, very excited to have you here for this Five Minute Friday episode. Title today is Machine Learning is All Around You. Really it should actually be Machine Learning, Deep Learning, Artificial Intelligence, and Data Science are All Around You. Why is that? And how did this episode actually come to be? Well, I have a few apps on my phone that I'm sure we all do, which quite heavily utilize these technologies, whether it's data science, machine learning, deep learning, or AI.
We can all think of one or two, but what I found fascinating is only a couple of them, only like one or two of them actually trigger in my head this sense of amazement, the sense of awe at how, how cool they actually are, how they're leveraging these technologies and they're really, like it really stands out that this is deep learning. Wow, this is really cool.
But if you think about it, there's plenty, plenty of apps on your phones that are indeed using this. And I think, I really think that it's a good idea, it's a very good idea for us as data scientists to actually observe that, to be aware of when and how these applications are using one of these technologies, whether it's data science, or deep learning, or AI, machine learning, doesn't matter. The point is that by being aware of it, by being aware of machine learning around us in use, we can actually come up with better ideas, we can think of better applications and maybe dig deeper into certain algorithms, and basically the point is use what's around you to get inspired.
I'm only using a phone as an example because your phone is always in your hands, and you're looking at it on average like 115 times per day. It's become an integral part of us. It's become like an extension of our mind. So it's the easiest example too. But surely there are other examples of technologies that use around you during the course of your day or week that also use these, use deep learning or AI or something like that. What I'm saying is I'm calling on you to be aware of when that happens and try to pause yourself for a second and look into it further.
As an example, I put a little list together of the technologies that amaze me, and at the very end I've got like the two that I was talking about, at the end of this I'll show the two apps that really like blow my mind every time I use them. They're nothing crazy sophisticated. They're just daily kind or like weekly usage apps, but they're really cool. But let's go through the list.
First one that I thought of was like a Tesla car. If you have a Tesla, you're a legend. I've been in a Tesla once. No, I actually, I don't think I've even been inside a Tesla, like driven, being driving or driven in a Tesla. I know friends who have, and I hear it's an amazing experience, especially when you can let go of the steering wheel and it goes in the highway, and it just drives itself, changes lanes and all those things. That's really cool. That would blow my mind. I'm really looking forward to that. If you already had that experience, that's awesome.
One thing is just to enjoy it and to have fun, but the other thing is to think through how is it doing it, like what kind of ... I know we can't really guess all these proprietary algorithms, and it's much more complicated and everything, but is maybe a good idea to think what do I really know in deep learning or machine learning that it could be using to make these decisions, or what is it that I could learn from how it's doing it so that to inform my further education in this space?
Another one is when you type something on your phone like a message and it gets autocompleted by your phone. That's a clear example of natural language processing that uses deep learning most likely to figure out what you're going to say next. And it learns as well. It learns from how you typed your previous 100 messages or what not and gives you better recommendations. In fact, you also right now have autoreplies suggestions in Gmail. If you get an email and then instead of replying a whole text to it, you get these three options. Like somebody invites you to a party and you can say, "Yes, I'm in," or, "Thank you for the invitation, I will be there," or it'll be like, "No, unfortunately I can't make it," and you can just click one of them and send. That makes life easier and that's also another example of an NLP in action.
Another case is when autocomplete ... when you have ... Now they actually have in Gmail when you're typing the message, it actually in this faint gray it will finish the sentence for you. You can just click Tab and you don't have to finish typing the sentence. I've already used that quite a lot, so that's another example of an NLP in action. Those are simple applications, simple use cases, but they're really powerful and they do use some powerful mechanics or algorithms in the background.
Another cool one, which we use probably on a daily basis, is Google Maps. How do Google Maps know that there's a traffic jam somewhere along the road, how do they, how does it know that your route is going to be delayed by 5 or 15 minutes? Well, we probably all know this by now but nevertheless Google just analyzes how the phones are moving across the street. All these phones in the cars, it can see like how fast they're traveling as opposed to how fast they should be traveling, and so then it decides whether there's a traffic jam, another example of machine learning in action.
An even cooler example is an app called Waze, and it's quite popular in the US, is getting more popular in Australia as well. I was present at an interview with the founder of Waze, and he described how they actually created Waze, and this is really fascinating. Waze unlike Google Maps or Apple Maps, it actually started with no map at all, like it doesn't actually have a map integrated as a starting point in the app, application. It started off with just like a blank black screen. Then people who signed up for Waze, they had the app on the phones and they would move around on the street. Waze, the app itself, would decide, all right, so like 100 people went in this direction from A along this straight line, that must be a road, and they went with this speed, and then here they went 60 miles an hour, so that must be a highway or a motorway. They could decide.
The app itself it builds up this world representation with all these roads, not based on any kind of map initially. That was all based on the activity of the users. And I think that's mind blowing. How cool is that? It took them many years to get it to a level where there was enough data for actually people to like to ... for it to be massively adopted, but now it's a very successful app, is one of the most successful map applications in the world, and all started with a black screen, blank black screen and just collecting data, machine learning, algorithms, analyzing it, and deciding what's a one-way street, what's a two-way street, it's a highway, what's a turn, what's a bicycle way I guess. I'm not that familiar. I don't use that app much myself, so I'm not sure the complexity is, but I'm pretty confident that has got some very, very complex algorithms in the background.
Another one that we probably all know as well is Spotify and how it thinks, it goes, analyzes your music consumption to give you suggestions, so the recommender engine. That goes for recommender engines on Amazon, recommender engines on Netflix and all these other platforms as well. Those are really cool to think about, like how did it know that I would like this music. If you think about it, 10 years ago, that would've been insane to tell somebody that an app can predict what kind of music you're going to like, even though, you've never heard that music before. I think that's pretty cool.
Okay, so we're getting close. Third last one, which I really like, an app that I enjoy a lot is called Foursquare. Foursquare, if you're using Foursquare, you're probably smiling now because it saved me plenty of time, like at least a couple of dozen times. It's an app that allows you to find places to eat, and if you're traveling to a new city or even in your own city, if you haven't used it yet, highly recommend checking it out.
You open it up and you say you want dinner or lunch, you want Indian, you want vegan, you want, I don't know, a steakhouse, and you put that in the search and then it will pop up a map showing you how users of the app have already rated the restaurants around your place. Every single time like if you find a restaurant that's eight stars or above, or especially if you find something like a nine stars or 8.9, once I think I saw a 9.2 restaurant, you have to go there. That means all these people, you can see how many ratings, but usually if it's at least 100, it means there's at least 100 people have been there and how that's how they rated it.
The power of the crowd is always much more powerful or much stronger than just one expert or a couple of experts saying, "This is a good restaurant." When you have 100 or a 1000 people saying it's a great restaurant, you got to go there. Every single time because as you guys probably know I'm vegan, so when I go to a new place I just type in vegan and brings up all these places that are vegan or serve vegan options and then I pick the one with the best stars and usually like 99% of the time it's a very good guess. They have a lot of users, very powerful platform. I think we already talked on the podcast about how they use their foot traffic data to like one time they predicted how the annual statements of Chipotle will ... what will they come out as even before the annual statements came out.
It's a long story. You can check it out. There's an article I think on Medium about that, Foursquare and Chipotle, but other than that, great app, uses of course data science quite extensively, so another thing to think about when you're using it.
Finally, my two top two apps that really blow my mind. As I said, they're quite simple ones, nothing sophisticated, not like a Tesla car, but really, really cool to watch them at work. Number two, or second top one is Scannable. Scannable is an app by Evernote. So if you don't use notebooks by Evernote highly recommend those as well. You can replace all your notebooks and all your notes with Evernote and take really cool ... just keep a lot of notes in one place, and it's very, very powerful, but then you can download Scannable as well, and instead of using a scanner to scan documents, you use Scannable and it's actually better.
There's lots of apps or you can take photos with your phone or there's other apps that you can take a picture, and it will recognize it as a document and save it as a PDF, but Scannable is by far the best one I've ever seen and the way it uses, I think it uses deep learning. I'm not sure. It might be some kind of machine learning or something like that, but to me it looks like deep learning, but it's just mesmerizing. When you put it, like put your phone over a page, even before you finish bringing your phone over, it's already taken a photo. This is blue screen that adapts to the size of the page. If you guys have a chance to check it out, you will really feel the power of these technologies in this simple application. Once again, it's called Scannable.
Finally, one of my favorite ones is Vivino. I was recommended this app by a friend and colleague and is if you drink wine, then this is a lifesaver app because you can go to a store and you can take a photo of a bottle, of a label of a bottle of wine, and just like with Foursquare you will get a rating, well, how have people rated it before, how have people who've drank this wine, how do they rate it. You'll see if it's like a three star wine, a two star wine, a four star wine, whatever it is, and then you decide whether you want to buy or not. Amazing. It's really cool to see it in action.
I especially like the feature when you can set it to not scan one bottle but scan multiple bottles, and you just go through the whole rack taking photos one after the other like in a row, like continuously, and as you're doing it, it's analyzing those photos in the cloud and popping up ratings right on your screen, so you can see like 4.3, 3.1, 3.9, 2.1, 4.4, and so on. Just as you go and then when you see like acceptable rating like a 4 or 4.5 stars or 4.3 stars for a wine, then you can just stop. It's really cool, also mesmerizing too [inaudible 00:13:53] in action. So even if you're not a wine drinker and you want to see some cool, I guess, and once again, I think that's deep learning in action, then highly recommend checking out Vivino.
There we go. Those are some apps that can get you inspired, and I'm sure you've been already using quite a few of them already, but it's like it's about changing the perspective. I've said this on the podcast before, a really cool quote from my coach and it's, "When we change the way we look at things, the things we look at change." We've been using these apps, but now if you change the way look at them, the perspective you have, what you think about when you use them, then all of a sudden you'll be looking at different things, all of a sudden you'll broaden your horizons, you'll find more opportunities for your education, for the application of the knowledge that you already have in the space of deep learning, AI, machine learning, and data science.
On that note, thank you so much for being here today, and I can't wait to see you next time. Until then, happy analyzing.