From the widespread presence of facial recognition to the development of high-resolution photo-realistic images, 2018 saw a lot of growth for artificial intelligence, and 2019 will be no different. We asked our CEO and founder, Matt Zeiler for his thoughts on how AI developed last year and what we can look forward to!
Natalie: What was the most exciting development in AI and computer vision that happened in 2018?
Matt: For me, the most exciting development in AI last year was generative adversarial networks (or models that can generate target variables) now being able to produce nearly photorealistic videos of scenes. Nvidia showed some interesting results from this applied to self-driving cars.
N: What trends in AI can we expect to grow over the next year?
M: I think that we’ll see more companies fusing multiple data types to start making solutions that use multiple areas of AI. Today, for instance, computer vision only understands images and videos, but teaching computers to understand more data types (like sound and language) will help developers and AI companies like ourselves provide more value to customers.
N: What’s one thing you still wish people understood better about the technology?
M: We are at a point where our technology can learn new things as fast as a human, but then apply those learning to significantly more data instantly. This lets us do far more with the data we already have. West Elm discovered that with our computer vision API they could learn from customers’ Pinterest boards in seconds. This lets them build an app that uses the images from these boards to learn a consumer’s preferences, search through the entire West Elm product catalog, and quickly return suggestions that are tailored to the customer’s preferences. Adding to that though, when they don’t find matches within their catalog, this can help them to identify potential gaps in their inventory that they can improve on for future customers.
N: What’s one thing you’re happy to see people catching on to?
M: Computer vision’s versatility! It always makes me happy to see the developer community continuing to grow and innovate on top of Clarifai’s platform. At hackathons, we have seen many developers building awesome apps with Clarifai that can really have an impact. Recently, Box’s Director of Developer Advocacy, Jeff LeBlanc integrated our technology with Box’s platform, so their users can get the benefits of Clarifai by processing their data stored in Box.
N: Finally, what industry do you hope to see adopt or adopt more AI or CV AI in 2019?
M: AI’s versatility is having it gain a lot of traction across multiple industries, but I definitely think retail is ripe for more disruption. There are just too many instances there, both in the physical and digital context, where finding products is far more difficult than it needs to be in the age of AI. For example, the product specifications you searched for will be on the actual product page, but the search indexes just can’t find it. AI can change this.
Thanks, Matt!
Want to learn more about AI concepts and terms? Get the simplified and comprehensive glossary below.
