This week, the The Deep Learning Summit and AI Assistant Summit saw over 450 DL and AI experts and enthusiasts come together to learn from each other and explore the most recent research and progressions in the space. Over the past two days we’ve heard from the likes of Amazon, Facebook, Google, Vodafone, as well as Universities such as Cambridge, Warwick, UCL, Imperial, and exciting new startups like Jukedeck and Echobox. Topics have been incredibly diverse covering NLP, space exploration, ML for music composition, and many more.

We’ve collected some of our favourite takeaways from both tracks over the last two days, as well as hearing what our attendees thought.

What did we hear at the AI Assistant Summit?

I’m driving in France and google translate is automatically translating all the french road signs for me and directing me to my location, telling me my time of arrival - this is the future. There is no interface there is no screen.
Adi Chhabra, Evolution of AI & Machine Learning in Customer Experience - Beyond Interfaces, Vodafone

Interactive technology is getting ubiquitous, it’s all around us. You can talk to your car, tv, even your home. These systems can understand what we are saying but not how, and this is what we need to solve.
Fabien Ringeval, Behavioural Computing for Artificial Intelligence, University of Grenoble Alpes

Even if you mix up what you said, Google Assistant can work out what you meant to say and correct your question for you and answer the question you meant to say. It has natural language - you can ask ‘what’s that movie where there's a beef sandwich and Bob in a car’ – it replies 'easy peasy! – its XX film'.
Yariv Adan, The Age of Assistant, Google Assistant

People often volunteer very intimate secrets that they wouldn’t usually share with friends or family to bots - people feel more comfortable talking to AI than people.
Andrew McStay, All Smiles? Emotional AI, ethics, citizens & ubiquitous feeling-intro, Bangor University

If there’s a chatbot it doesn’t matter how stupid the question is, you don’t feel silly asking.
Gregor Jarisch, Best Practices & Challenges of Designing a Virtual Teaching Assistant, Differ

We are at the beginning of the era of assistance. In the future every employee will have an assistant to help him with decision making.
Christophe Bourguignat, Deep Learning for Conversational Intelligence on Analytics Data, Zelros

There’s no common practices or framework of building chatbots, and our framework creates that and ensures that we’re creating bots that deliver a consistently high user experience for our customers.
Alex Fawcett, Pegg: Building Conversational UI at Scale, Sage

Brands spend money convincing people to buy a product through advertising and not to create a better future. What if we could create a better future? Advertising should be innovation and the key is AI and our relation with it.
Rob McInerney, Using Artificial Intelligence to Make Everyday Products Better, Intelligent Layer

At the Deep Learning Summit we heard..

The goal of this event is to bring together people with diverse backgrounds and to foster discussions and interactions around DL, not just to listen.
Fabrizio Silvestri, Facebook

I managed to give my colleague a lazy eye even though he doesn’t have a lazy eye, but I managed to find the data to alter that - it’s the way of letting data imagine something based on what you already know that’s important to drive these progressions.
Andreas Damianou, Probability & Uncertainty in Deep Learning, Amazon

If we show a picture of cubes to a machine, normal machines would simply see and identify a square. However, what we want to do is constrain the model by taking individual pixels and teach it to assume that every grid is an object. We can then analyse everything separately and force the model to compare.
Marta Garnelo, Representations for Deep Learning, DeepMind

Q: How does the AI know which hashtag to use? A: We predict hashtags by how often they are used, we also auto generate our own hashtags appropriate to article. We also exclude hashtags that are generic or inappropriate to the article.
Antoine Amann, Using Deep Learning to Understand the Meaning of Content, Echobox

AI is changing everything around us and impacting every field. It's unavoidable.
Jonas Lööf, Considerations for Multi GPU Deep Learning Model Training, NVIDIA

I don’t know to what extent machines will be able to touch emotion, and it’s definitely not there yet, and that kind of thing is a challenge we’ve not yet faced. We’re literally just training these systems on music. When I write music I don’t just look at Bach’s music and think how I can create it, but when I write music there are so many aspects that go into it - maybe my girlfriend just broke up with me, or maybe there’s turmoil in the US - a machine can’t capture that...yet.
Ed Newton-Rex, AI that can Compose Original Music, Jukedeck

It’s always great to get feedback from attendees to see what their highlights of the event are, from the people they’ve met, to any hiring or investing opportunities, and what everyone’s learned over the two days. It’s a lot of information to take in, and everyone takes away something different, so let’s take a look at what our attendees have to say:

We’ve already met some potential customers and people interested in our technology. The event is also great for potential hires as there are lots of skilled and very interesting attendees. Guillaume Bouchard, Bloomsbury AI

Luca from DigitalGlobe’s talk on speech to video and video speech in DL was so good it was crazy! Trym Moller, Liace

My favourites yesterday have got to be Marta from DeepMind & Yariv from Google Assistant. Ed today talking about machine learning for music was fabulous - so diverse. Ahmed Razed, BBC

I've been to many conferences but here I enjoy the mix of technology and the soft side of application i.e. Ethic, emotional intelligence and the business aspects. Krishna Dubba, Nokia

If you were at the summit, we’d love to hear from you, so get in touch at [email protected]. If you missed out this time, check out our events calendar and see where we’re heading on our Global Deep Learning Summit tour.

Our largest event of the year in San Francisco is coming up on January 25 & 26, , register now to guarantee your spot at the event where we will hear from Google, Netflix, Facebook, DeepMind, OpenAI and many more.