Changing the Way We Talk to Machines

As our mobile devices, wearables and cities become more powerful and feature-rich, a better understanding and dialogue between humans and machines is becoming key to the success of emerging technologies. Achieving human-machine dialogue is often considered the very definition of artificial intelligence. Deep learning has resulted in large improvements in speech recognition, but there are still barriers to overcome before true conversation is realised. A conversational system must also understand what was said, adjust for context, decide what to say and decide how to say it. Blaise Thomson is CEO of VocalIQ, a company that builds technology to solve all of these tasks with machine learning. Their aim is to augment and impact our experiences with machine interaction - changing every aspect of everyday lives. We caught up with Blaise ahead of his talk at the Deep Learning Summit in London this September.What are the key factors that have enabled recent advancements in deep learning? Data, computational power and community interest.  What are the main types of problems now being addressed in the deep learning space?   All of the standard machine learning problems: image processing, speech, NLP tasks, etc. More recently people have been moving to the problems in traditional artificial intelligence: dialogue, logic and knowledge  What are the practical applications of your work and what sectors are most likely to be affected? Around 20 billion smart devices will soon be produced each year. For most of them, spoken dialogue is the natural interface. A machine-learned approach means that dialogue will actually work, improving interactions in all sorts sectors. Examples include cars, smart homes, TVs, wearables, smart phones, robots and language learning.  What will VocalIQ revolutionise our interaction with machines? Current voice interfaces are almost always restricted to one search or command. VocalIQ allows a system which has a full dialogue and that can learn for itself how to behave.  What sectors will have been disrupted by deep learning in the next 10 years? Every one. I can’t think of a sector that wouldn’t be.The Deep Learning Summit is taking place in London on 24-25 September, alongside the Future Technology Summit. For more information and to register, please visit the event website here.