The simulation of human intelligence by machines, better known as AI, is no longer a chapter in sci-fi novellas. AI is here, and it’s here to stay.
When caught up with AWS customers at the re:Invent expo, all were unanimous that AI was reducing the dependence on data scientists. It wasn’t just about Internet of Things (IoT), but a convergence of all forms of data on to a single platform.
With data from social networks, personal devices and work converging for a single view of understanding how things work, the times, they are a’changin. Similarly, data in business is converging from silo-ed machines to everything. Consumer data, supply chain data and manufacturing data are being prepped on to a single platform to enable businesses to make faster decisions.
recently spoke to Matt Fryer, the Head of Data Science at travel company Expedia; Walter Scott, CTO and Founder of Digital Globe, a company that provides high-resolution earth imagery; and Mati Kochavi, the Founder of AGT International, an IoT company; to know more about how they are using AI in their businesses. Here are excerpts of the discussion:
YS: Matt, Expedia crunches TBs of data. How do you view the cloud and why should organisations move into it today?
Matt Fryer: Machine Learning along with data science offers scale and provides new capabilities and insights. We, at Expedia, are helping the customer gather insights from millions of price changes in the travel industry and offering him the best choices.
In the 1960s, travel agents had power; then the internet offered price discovery, and now you can navigate the complexities of the best price with AI and Machine Learning. The core is the cool basics like compute, elasticity and cost economics, which have all fallen thanks to the cloud. It transforms how we tackle problems. We have a culture of experimentation and AWS helps us with its new AI and ML tools.
A good example of AI is to get the right photo of the place and help the customer find the image that best describes the place they are looking for. We have hundreds of photographs pulled in seconds. Crunching images takes so much time; now AI helps us. AI will allow us to access huge numbers of options of flights and hotels and bring it to customers on the fly.
It will no longer be like a recommendation engine built on rules; it will be one that is constantly learning about tastes and preferences. You will be empowered to find the right options and that is a game changer. Machine Learning and people have come together. You need creativity and this is what data science combined with AI is about. Mining millions of data points and bringing the right imagery to customers is what increases revenues. Machine Learning has made it possible to teach software about patterns; machine intelligence makes it faster.
YS: How can AI save lives?
Walter Scott: Digital Globe provides high-quality satellite imagery, which can be used to analyse geospatial data. Companies like Uber use our data and every year we consume volumes of data – around 100 PB of data. Our satellite gives us 80 TB of data and generates 4 million km of data, which will take 85 years to analyse if we employ a single data scientist.
Now, what do our images and our platform tell you? We predict wildfires, for example, and the path the fire is taking. I can give you instances where wildfires burnt houses with solar panels. But those of you who know science will also know these solar panels are now fully charged and open to electrocution risk.
Meaning, if anyone, say a firefighter, walked on to the site he could be electrocuted although he is there to fight the fire. He may not know that the building had a solar panel and that it is dangerous to walk on that path. So, AI helps firefighters know this with quick image processing. You don’t know what the fire has burnt, what is live, and what is not.
AI can crunch data fast and guide people in a firestorm. It takes the guess work out in many industries and makes data more relevant and meaningful. You also need to know where buildings are and trees are. Trees block signals and having AI helps understand an entire layout faster.
So what we are doing is converting structured and unstructured data by leveraging Machine Learning. What took two months for data scientists to figure out can now be done in a week.
Data can help drive necessary change. By crunching satellite imagery we can also ensure vaccine delivery in crises-affected areas in Africa, for example. We have mapped populations from space to deliver the right vaccine. The industry is moving from guesswork to answering queries better. Thanks to the cloud, large data sets are no longer the problem.
YS: Mati, is AI helping us tell stories better?
Mati Kochavi: AGT International established HEED 18 months ago to change the way we view live events. HEED crunches live data collected from sensors in the sports arena and reads player emotion. It then matches it to the emotions of the fans as visualisations and avatars. IoT and sensors can change the way data is measured live by broadcasters and how it is visualised to people. With AI, companies will be able to understand a problem and solve it in the most efficient ways. We are investing in data scientists and they understand Machine Learning and people. We look at IOT as a great concept of storytelling.
Data is telling you stories all the time and it can solve complicated problems. Data can predict anything. Image-processing technologies generated from sensors can look at emotional status of people and, in sports broadcasting the game can be experienced on the couch. Ten years from now, data will determine the future.
HEED is blending the art and science of reading human emotion.
The data talks about the physical aspect of different manoeuvres and the damage they can inflict on the adversary. Sport is about human spirit and there are reasons why people win. The answers are very different from what people thought when we measure emotions. We want to tell this in a very artistic way by using live data. Athlete avatars can be played to broadcasters with the data of live sport and use cool visualisations. Imagine how it will change the sports business when 1 billion people are watching an event.
While we do all this at HEED, the pilot of crunching volumes of data was what AWS helped us with. Data tells you different stories about different people, but it is tailor-made with AI.