eBay is using machine learning and “deep learning” to help determine what a shopper is looking for when they enter a search term. eBay researcher Hassan Sawaf heads the team tasked with developing tools and services in language technology – and the main objective is to open up new geographic markets.
The Senior Director of Human Language Technology is featured in a video talking about the techniques. For example, the technology can be used to identify items based on single images from various angles.
The idea is to help a buyer in one country find relevant items in another country even when they don’t speak the same language – the technology uses semantic clues that are extracted using advanced machine learning techniques.
In the video, Sawaf talks about how eBay is using deep learning, and he hopes to speak on the topic at the South by Southwest show next year.
In the session, “Flip-Flops or Flats: Deep Learning for Commerce,” Sawaf would demonstrate how to utilize natural language processing and visual object recognition to understand users’ intent, and how this technology can distinguish whether users are buying or selling a pair of flip-flops or a pair of flats when simply typing “footwear.”
Last year, TechCrunch wrote about Sawaf and his team of data scientists and engineers. The publication said Russian users can search eBay in Russian and see listings in search results written in English that match their search.
TechCrunch said the new feature was resulting in greater numbers of Russians entering search queries in their native language.
eBay has been working to grow its business through cross border trade, and another eBay executive recently updated a patent dealing with geographic search algorithms.
