Sessions is temporarily moving to YouTube, check out all our new videos here.

Microsoft's view on AI

Martin Kearn speaking at HydraHack in March, 2017
43Views
 
Great talks, fired to your inbox 👌
No junk, no spam, just great talks. Unsubscribe any time.

About this talk

In this talk, we will see what Microsoft is doing product-wise in the AI space, and some of the cool things you can do using their technology.


Transcript


So, hi, everyone. My name is Martin. I work at Microsoft. I'm based local to Birmingham. I live in Worcester, so I try and get to as many Birmingham events as I can. I've got probably, what, three minutes now, maybe, to talk about the Microsoft AI story, which is going to be interesting. So what I'm going to try to go through is kind of Microsoft's view on AI, what we're doing in terms of products around the AI space, where we think it's important going forwards, awesome things you can do today around AI using Microsoft's technology. And obviously I'll be around for questions and that afterwards, because it's only a very high-level skim-over. So there's kind of an arms race going on at the moment with all the big tech companies around AI. This term, which is becoming like an overloaded term now, this idea of having artificial intelligence in applications and software. We're at it, IBM are at it, Google are at it, Apple. Everyone that's kind of doing big technology is really focusing on AI at the moment and providing services for developers especially to use around AI. And we're no different. We like to think we're definitely in the mix as one of the leaders in this space, and hopefully I'll explain why. So when you think of Microsoft and artificial intelligence, what do you think of? What's the first thing that pops into your mind? - [Man] Clippy! - Clippy! I'm glad you mentioned Clippy, actually. A lot of people don't think Clippy. They think evil Armageddon, you know, big corporations making robots that can think for themselves and all that kind of stuff. And actually, that's what a lot of people think of when they think of AI, the kind of Hollywood version of AI where robots and bad things happen and things take over the world. That's not really what we're about. We are actually more like Clippy. But not just Clippy, Clippy on steroids, so this is what that actually is. So it's more like that, and actually, that goes for AI across the whole industry at the moment. People are doing AI but they aren't really doing the kind of <i>Terminator</i> scenario. It's more about doing the kind of...helping people be more productive, that kind of stuff. And this really comes down to the difference between strong AI and weak AI. So strong AI is the <i>Terminator</i> scenario. It's where machines can develop...and this is a thing that exists. This isn't just in Hollywood. There are machines that can develop some kind of consciousness. There's machines that can do drawings and paintings, and really have what you might consider to be actual intelligent thoughts on their own. Mine is not one of them. And there's weak AI, and weak AI is all about assisting humans in doing their everyday tasks and productivity. And that's the space that Microsoft's doing, and that's the space we think is going to be interesting for the next kind of 10 years or so. And that's actually the space that most of the tech companies are in as well. So it's not about taking over the world, it's just about taking over your devices. So we've got this vision that we think kind of, humans plus machines equals ultimate intelligence. So it's about taking the things that humans are really good at, taking the things that machines are really good at, bringing them together, and then you've got something really good. That's basically the overall premise of what we think about with AI. And so, when you think about that, when you think about that, what are Microsoft's machines? What are the things that we've got in the market today that you can do AI with? And there's quite a few of them. And really, AI is at the core of everything that we're doing at Microsoft. So you think about Office, you think about Windows, you think about Azure and the Cloud, all of that stuff. AI is not my machine, my machine is not being intelligent at all at the moment, but AI is at the center of everything that we do. All of the product teams, though, looking at, "How can we build an AI into what we're doing with the products?" And it really breaks into three areas. The first one is cognitive computing. So this is this idea of having machines that kind of replicate the sorts of things that humans can do. So humans are very good at seeing a picture and recognizing the objects in the picture. Machines are generally not very good at doing that, but we've got technology that lets you do that. Humans are very good at reading emotions in people's faces, and machines generally can't do that very well, speech to text, that kind of stuff. Understanding sentiment in text, and that's cognitive computing. The second area is conversation as a platform. So this whole notion of being able to talk to a computer in a natural language way, whether you're typing or speaking, and having it, A, understand what you mean, and B, talk back to you in a natural language way. So actually having a conversational user interface for your computer or device, or whether that's a car or something in your kitchen that's touch and speak, whatever it might be, talking to an electronic device in a natural human way. That's what we call conversation as a platform. And finally, data and machine learning. So data and machine learning is really the kind of underpinnings of everything to do with AI. It's all about data, it's all about understanding data, whether that's understanding your own data or whether that's understanding general data. None of this would be possible without machine learning and data. So in terms of products, we have a thing called Microsoft Cognitive Services, which some of you may have heard of. A set of APIs that let you do some really clever things. I'll show you some of those in a second. In the conversation space, we have the Microsoft Bot Framework. So if anybody's ever looked at chat bots, you know, Facebook bots, Skype bots, those kinds of things, we have a tool called the Bot Framework which lets you write one bot and publish it to all of those different channels. That's called the Microsoft Bot Framework. We obviously have Cortana as well, which is our digital assistant. So any Windows device, you can talk to Cortana, you can have Cortana do things for you. And that's where we kind of...that's the interface of our AI vision, really, is Cortana. And in the data space we've got loads of stuff. We've got Microsoft Azure machine learning. Is that time already? Yeah. Okay. Fine. I'll go quick. So that's our service, we're building [inaudible] machine learning algorithms. We've got the Cortana Intelligence Suite and a thing called the Cognitive Toolkit, which is all about deep learning. And there's some really deep neural networks and that kind of thing. So in terms of cognitive services, there's 25 APIs, there's 66 things you can do. It's all about understanding pictures, understanding voice, understanding loads of stuff. A better way to explain it is to show this very quick video, which I think explains what cognitive does really well. If my speaker's on... - [Saqib] I'm Saqib Shaikh. I lost my sight when I was seven, and shortly after that I went to a school for the blind. And that's where I was introduced to talking computers. And that really opened up a whole new world of opportunities. I joined Microsoft 10 years ago as a software engineer. I love making things which improve people's lives, and one of the things I've always dreamt of since I was at university was this idea of something that could tell you at any moment what's going on around you. - [Cortana] I think it's a man jumping in the air doing a trick on a skateboard. - I teamed up with like-minded engineers to make an app which lets you know who and what is around you. It's based on top of the Microsoft intelligence APIs, which makes it so much easier to make this kind of thing. The app runs on smart phones, but also on the Pivothead smart glasses. When you're talking to a bigger group, sometimes you can talk and talk and there's no response, and you think, "Is everyone listening really well? Or are they half asleep?" And you never know. - I see two faces, 40-year-old man with a beard looking surprised, 20-year-old woman looking happy. - The app can describe the general age and gender of the people around me and what their emotions are, which is incredible. One of the things that's most useful about the app is the ability to read out text. - [Waitress] Hello. Good afternoon. Here's your menu. - Great. Thank you. I can use the app on my phone to take a picture of the menu, and it's going to guide me on how to take that correct photo. - Move camera to the bottom right and away from the document. - And then it'll recognize the text. Read me the headings. - I see appetizers, salads, paninis, pizzas, pastas. - Years ago, this was science fiction. I never thought it would be something that you could actually do, but artificial intelligence is improving at an ever faster rate, and I'm really excited to see where we can take this. Hey. - [Woman] Hi. As engineers, we're always standing on the shoulders of giants, building on top of what went on before. And in this case we've taken years of research from Microsoft Research to pull this off. - I think it's a young girl throwing an orange Frisbee in the park. - For me, it's about taking that far-off dream and building it one step at a time. I think this is just the beginning. - Okay, so that app that we saw there was built in about three days, and that guy actually built it himself as well. So it looks like it's really clever, but it's really not. It's just leaning on all of the services that we've got available as part of Cognitive Services and bringing it together in quite a useful scenario. So bots are important. If you want to know about bots, come and see me afterwards. And finally, we've got a load of stuff in Azure as well. So if you're not using Azure and you're interested in doing this stuff, then definitely take a look at that. But yeah, that's me done. So any questions, you can ask me directly now or we can go to the next speaker or whatever. I don't know what's next. - [Man] It's not a question, but I'm actually glad that you're focusing on making people more intelligent, in that sense. For me, like, the big fear of growth...it's not a fear but a concern that AI is trying to replace people too, driving for example, things like that. But the first thing, yeah, it can actually help people be better at certain tasks, or just have the extension of the mind that [inaudible] whatever I can do for them. [inaudible]... - Yeah, that's absolutely, yeah. - ...for me, so I'm glad Microsoft is focusing on that. - It's a kind of humans plus machines thing, so that's definitely where we're focusing at the moment, and then for the foreseeable as well. It's about...so things like with that video of Saqib there, it's helping him do things that he couldn't ordinarily do by using AI and using the ability to recognize objects, and images, and videos, and things like that. - Yeah. I mean, this is great, this video. For him, he [inaudible]. - Yeah, absolutely, yeah. And it is...it's kind of genuine. It looks like it's a big corporate video, but he works at the building next door to where I work. I know him really well, and he's actually a coder and he's been at Microsoft for 10 years. So it's like a real thing and makes a big difference for him especially. Cool. Okay. Well, thanks, guys. I'll hand it over to the next person.