Artificial Intelligence: Cédric Villani, Mathematician, 2010 Fields Medal winner and Rand Hindi, founder of the start-up Snips shared their insights into the challenges of AI and its applications in the energy sector.
Read the transcript:
Artificial Intelligence & Data
Rand HINDI // Not embracing artificial intelligence today would be like not embracing computers in the 1980s. It’s not a niche technology. It’s an entirely new approach to computerization and digitization for the future.
HOW ARTIFICIAL INTELLIGENCE IS CHANGING THE WORLD
Founder of SNIPS,
an artificial intelligence start-up
Rand HINDI // Artificial intelligence is a field of research that involves endowing machines with autonomous behavior. A branch of artificial intelligence, known as Machine Learning, seeks to provide the machine with examples of behavior, which enable it to learn to recreate such behavior itself. Within the branch of Machine Learning, there is a specific area called Deep Learning.
The computer is programmed to learn increasingly abstract concepts, and to combine these different concepts with one another. For example, if I were to input the picture of a cat, the computer would process the outlines of its shape, learn the concept of an outline and then combine the outlines to recreate the image of a cat's face.
Using data to improve lives
Rand HINDI // Basically, artificial intelligence only needs one thing – data. Data is the raw material used to produce artificial intelligence. When I write a letter to someone, I create data about my feelings for that person at that time. When my cell phone tracks my position, it is collecting data about my movements. If we start to capture everything that is going on in the world and in our lives, we’ll be able to use artificial intelligence to improve those things too.
The beginning of a new era
Mathematician, 2010 Fields Medal winner
Director of the Henri Poincaré Institute
Cédric VILLANI // We have entered an era in which all, or almost all, phenomena can be measured and quantified using all kinds of sensors and models. One after another, we have seen numerous fields open up to the possibilities of modeling. In finance, for example, the revolution has done away with traders in the traditional sense, whose tactical battles of intuition and strategy have been replaced by battles of algorithms. Transportation is now dominated by questions regarding traffic optimization, resource deployment and ticket price calculations. Even medicine is slowly being transformed by algorithms. While mankind develops the broad strategy, machines takes care of the calculations.
A world of opportunities for the energy industry
Rand HINDI // For starters, there is anything related to smart grids. Sooner or later, we are going to have to channel energy consumption to the right places and anticipate needs. But it's also very easy to imagine using artificial intelligence to improve energy use in buildings. Google has succeeded in harnessing artificial intelligence to lower data center power consumption by 40%.
The future of artificial intelligence
Rand HINDI // The future of AI first lies in understanding our natural speech patterns. It's one of the most complicated things for a machine to learn, simply because we call upon a great deal of contextual data to "decode" the true meaning of what is being said.
Today, we are able to create AI with very vertical abilities, meaning that they are able to carry out one task, and carry it out better than human beings. But artificial intelligence is very far from thinking independently or understanding abstract and general concepts like we, as humans, can do.