Deep Learning for Fashion Search
Only 4 to 6% of users that visit an online store actually make a purchase. One key reason for low conversion rates is the quality of the search functionality. Currently most online stores restrict consumers to search only with text or category filters, but visual search is not allowed. However, the fashion e-commerce world is largely dominated by visual cues and consumers want to find what they seek fast with the right colors, shapes and patterns. In this talk, we will discuss a series of techniques that can be implemented in your store to improve the quality of search results, enabling visual search and using state-of-the-art techniques in computer vision and natural language processing. Implementing these techniques will enable stores to capture insights into what customers really want to buy and ultimately increase sales.
Susana received a PhD in Computer Science in Dec. 2016. Her research interests lie at the intersection of computer vision and natural language processing, and include deep learning, topic modeling and graphical models. Specifically, she focuses on developing end-to-end learning architectures to jointly detect fine-grained attributes on both images and text. She has worked for NASA as an Artificial Intelligence Researcher to automatically search for long-period comets that might impact Earth. She has worked for Microsoft Research, where she developed machine learning tools for optimizing environments for large scale software development. Additionally she holds two Masters degrees, one in Mechanical Engineering, where her research focused on human-robot interaction technologies, and one in Mathematical Physics, where she focused on gravitational fluctuations in Domain Wall Spacetimes. In 2014, she was awarded a Google Anita Borg Scholarship.