hfng - Fotolia
Tired of hearing about AI tools and techniques? Well, get ready to hear more, because AI is likely to continue to be a hot topic throughout 2018 and beyond.
That's according to Deep Varma, vice president of engineering at online real estate listings provider Trulia LLC. The San Francisco-based organization, which is part of Zillow Group Inc., has seen promising results from investments it made in AI over the past couple years. Now, Varma thinks many more companies will follow in the year ahead.
"We are going to see AI become more and more pervasive," Varma said. "Businesses that are not able to join the bandwagon are going to miss out on the opportunity."
Trulia uses a number of AI tools and techniques, including recommendation engines based on machine learning functionality, personalized website experiences and natural-language-generating algorithms that automatically create descriptions of neighborhoods across the country. Varma said Trulia has seen substantial value from all of these, and he predicted that more businesses are likely to launch similar initiatives as the barriers to entry continue to be lowered.
Using AI tools and techniques gets easier
One reason AI is getting easier to deploy is that enterprises are creating and managing more data than ever before, driven in part by plummeting storage costs. This data is a key ingredient for training machine learning algorithms.
At the same time, advanced algorithms are becoming standardized and made available to anyone via open source software and programming languages like Python. Varma said these trends make it possible for virtually any business to get started with AI.
Recommendation engines and personalized website experiences are the first areas in which Varma expects businesses to get involved. They're among the easiest implementations of AI tools and techniques, involving tried-and-true machine learning algorithms. Additionally, they tend to have quick and obvious payoffs via improved customer engagement.
Visual search is another big opportunity Varma is eyeing. He said applying deep learning algorithms to develop natural language descriptions of the content in images will enable people to search for items in a more intuitive way. In Trulia's context, users could search for homes with granite countertops and find results even if the listings don't specify the counter types in any structured-text form.
Varma also expects augmented reality to be a big part of the data landscape in 2018. For Trulia, that means letting people see what a home would look like with different wall colors or landscaping.
"We can use augmented reality before [people] buy a home," Varma said. "They can see what the home is going to look like so it's not speculative."
Other industries, including retail, will also adopt augmented reality, Varma said. He sees shoppers being able to virtually try on clothes or get a 3D view of items. Machine learning algorithms can play a role, as well, in creating personalized experiences for shoppers by remembering their preferences and finding products that fit those tastes, he added.
Watch out for data security problems
One area that Varma said concerns him is data security. As more enterprises increase their data stores to feed machine learning algorithms, the repositories are going to become bigger targets for hackers. However, Varma said he doesn't see many businesses taking active steps to protect their data.
"I'm not seeing enough momentum building up around that," Varma said. "What we are seeing now is businesses are trying to capitalize on an opportunity with AI. But as they capitalize on this, consumers are going to demand more protection of their data."
He said blockchain offers some opportunities to improve data protection. But he views the security protocol as still being immature, and he wants to see more regulations develop around it before Trulia starts using blockchain throughout its operations. He expects many businesses to take the same approach.