 Dyna yn ymdyn nhw'n dweud y potensiwch ar aion a ardyn ni i gael ei wneud yn creu newidol i gael i'r cyffredin ar gyfer y gwerth gweithio. Mae cyfnodd yn ymryd yn ymryd yn ymgyrch. Mae'n dweud ymweld ar y ddweud â'r cyffredin. Ac mae'n dweud o'ch ei ei ddwyf yn symud. Mae'n ddweud bod yn ymryd yn ymryd. Mae'n ddweud â'r cyffredin sy'n dweud â'r cyffredin sydd yn ymryd. My name is L.J. Rich. Everything I am going to do is live. Nothing has been pre-recorded. Okay, okay. Something strange just happened. I didn't tell you anything. You all knew what to do. How did you know to do that? Now you've got the idea. Let me try another song. Okay, so something. Something's going on. There are some rules. We use these rules to create moods. Like magic and mystery. We use those rules to create strength and certainty. Or just something to keep us calm while we're on hold. These are our unwritten rules. How we respond to music. How our body moves. How our thinking changes. These are rules that we know so well. And yet, they are really hard to teach a machine. How do we teach a machine musicality? So in my day job, I'm a TV presenter. I'm an inventor. I present on the BBC talking about music and technology trends. I'm also a composer and a sound designer. I build musical installations that make people feel good. These are paper mandalas by touching this. That lady is triggering sounds from NASA's Discovery Space Mission. On the left here, those are people touching flowers and triggering tropical forest sounds. And just there is the Science Museum, who hosted an idea I had about creating dance music based on how much people were moving. We picked up sensor data from their mobile phones, and my plan was to derive the wiggle index and therefore get the ultimate dance anthem out of data. I make devices that make people experience music differently. Some of my inventions have been, well, in the National Press. I've lectured at Harvard, DJed in Tokyo, and now I'm leading a project with NASA on creative approaches to data sonification. So just a few years ago, I found out that my brain is wired differently than yours. It's a neurological condition called synesthesia. My senses are completely mixed together. When I first tell people about this, many think it sounds a bit crazy. Some find it funny, but most are quite fascinated and would like to know more about it. Around three to four percent of the population have this. You might see colours and numbers together, or letters or days of the week. My version is particularly strong. For me, these sentences make complete sense. David Bowie, Life on Mars. It's like a beautiful dark blue colour with purple. Aber. That's actually, pleasingly, the same colour as the Swedish flag, blue and yellow. Sushi is this kind of beautifully clear acoustic D and A. It's that perfect fifths. And coffee. Oh, beautiful coffee. It's like a beautiful coffee. It's actually like Gershwin. That's why I don't drink bad coffee. Making unexpected connections is pretty much how everyone generates new ideas, going from one modality to another. And it's a practice easily seen in consumer electronics before the worldwide web, the smart revolution and the internet of things, phones did not come with cameras. Wow. Speakers didn't come with built-in AI assistants. Central heating didn't come with internet connectivity, but remixing technology is now commonplace. So it makes perfect sense to me that AI will change music just as it will every other aspect of our lives. So going back to teaching a machine musicality, first we need to think about what's useful to know about how music is built. Music is a repetitive phenomenon. It's made up of a set of repeatable notes. Everything is made up of the same small bass elements. Now western music has one set of notes. Other styles have different intervals, but in general all music is considered repeatable sounds. And notes are combined in different ways. They're used to create different feelings. Perhaps we can get music and machines to understand each other if we extract its essence, its DNA. A machine could recombine it, work out how to compose music to make us feel specific things. Maybe machine learning could create the tune that gives us that hairs on the back of our neck feeling. Well, we're halfway there. No, I can't do that, honestly. OK, so with a bit of service space, you can train an AI to create music based on your favourite band or artist, which is exactly what some people have done. Would you like to hear some music generated by machines? Yes, awesome. OK, here is Bach by David Cope. He's a composer. Bach wrote over 1100 organ works and he decided to just throw this all into the computer quite a while ago, actually. When it first started throwing music out, it was boring. He had to add a random element so that the machine would break the rules just like Bach did. It's pretty good. What about this? This is the Beatles. The Beatles, I should say. This is by Flow Machines. They got a band to play the song that was composed by an AI. This song is called Daddy's Car. The lyrics were generated by machines. The idea was generated, but it was played by humans, so it does sound more human. OK. What about black metal? Ready? It's a CJ Carrs project and he has a 24-hour black metal AI streaming channel on YouTube. It's gone viral. It's generated completely by an AI using a recurrent neural network. That's enough of that. OK. There are issues with AI musicians. I know this because I was jamming with an AI in Croatia as part of a consortium of music technologists. There's a sentence you don't see every day. An AI machine is terrible at listening and yet attempting to play along with it was quite inspiring. You have a conversation with another person. You find places to connect with them. You learn their character. How fast they talk. It's the same when you play music with another person, but in this case I was learning the personality of a machine. Machines don't understand when to break the rules. They don't understand social cues when you empty a band. There are lots of these unwritten rules that we follow together and the AI can do low-level work. What is the next note after the previous note? But it struggles with high-level concepts. Right now, I can't make AIs that cross genres. A bit like playing an easy listening version of a rock classic like Iron Maiden singing Two Minutes to Midnight in the style of Mozart Opera. It's currently something that no machine can do very convincingly, but luckily my synesthesia allows me to do this and I'm now at your service. So if you would like to have a go, please can you call out some artists? What was that? Motley Crue. Prince Daft Punk? OK. Can I have some genres, please? Pope Blue's Reggae? OK. So we've got Daft Punk and Reggae I picked out. OK. So what I do is I kind of take the elements of these things and I mix them together. So Reggae is interesting because it has the accent on the second and fourth beat of the bar. OK. And then you've got like a Daft Punk song which has a really driving beat. OK. So how do you mix those together? You do this. So that's the idea. Normally I would do one or two more because you're such a fabulous audience but I'm aware our time together is coming to an end. So maybe I'll do this later if there is a piano somewhere. So let me leave you with a few thoughts about why I think music and technology could give us all clues on what the future holds. Voice is music in motion and tonality. In motion recognition could be a very interesting and powerful development. If your bank, for example, detects stress in your voice it could know how much you really need to borrow that money and then adjust the interest rates accordingly. And mechanical devices, they make music. The sounds of an engine could be analysed and this could allow a system to work out the likely points of failure way before we hear that big bang there'll be micro sounds that you could learn to pick up. I believe all of us have the potential to think like someone with synesthesia in the same way you feel when something is right. Inspiration is everywhere but it's how we channel it that makes a difference. Those unexpected connections could be the key to you solving your specific problem creatively and elegantly. And I believe every one of you can access this more easily than you think. Inspiration is ours right now. Doesn't have to be elusive. It could be as simple as brushing your teeth with the other hand or ordering something different. So please start making your own unexpected connections and I hope that you have a truly inspirational evening and beyond. Thank you very much.