Abstract: 

Convolutional Neural Networks (CNNs) are responsible for unprecedented advances in the field of Computer Vision since they have achieved impressive performance in challenging tasks such as image classification, attribute recognition, object detection and segmentation, among others. In this talk, we will take a deep dive of the inner workings of CNNs. We will answer questions like: What are CNNs? What exactly do they compute? Why and when do they work well? How do they learn? We will explore common architectures and we will showcase an application on multi-label attribute recognition.

Bio: 

Susana Zoghbi, Co-Founder & CEO, Macty. Susana is a researcher and entrepreneur in a quest to help businesses grow with Artificial Intelligence. She received a PhD in Computer Science and her research focused on cross-modal processing of textual and visual Information. She designed deep neural network architectures and probabilistic graphical models to understand visual and textual content from e-commerce and social media. Her work has been published in top conferences and journals in Artificial Intelligence. She has worked for NASA’s Frontier Development Lab as a Deep Learning Researcher to automatically search for long-period comets that might impact Earth. She has also worked for Microsoft Research in Cambridge, where she focused on machine learning for optimizing environments for large scale software development. Before her PhD, she obtained two Masters degrees, one in Mechanical Engineering from the University of British Columbia, 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 granted a Google Anita Borg award for her contributions in Computer Science and her community.