Presented By O'Reilly and Cloudera
Make Data Work
31 May–1 June 2016: Training
1 June–3 June 2016: Conference
London, UK

Visual data analysis for intelligent machines

Francesca Odone (University of Genova)
11:00–11:30 Wednesday, 1/06/2016
Hardcore data science
Location: Capital Suite 4 Level: Advanced
Tags: ai
Average rating: ***..
(3.50, 2 ratings)

Prerequisite knowledge

Attendees should be familiar with visualization and data mining.

Description

Human perception is one of the keys to intelligence and one of the most complex abilities to emulate in machines. In recent years, for the first time, we have been able to develop artificial intelligence systems able to “see”: modern cameras recognize faces, cars can detect the presence of pedestrians, and mobile apps read text and recognize logos. In most cases, at the root of these success stories are machine-learning algorithms, which allow us to design programs trained rather than programmed to solve a task.

Francesca Odone explores analyzing visual data (images and videos) with the purpose of extracting meaningful information to solve different scene-understanding tasks. Francesca addresses the problem of learning adaptive data representations trying to incorporate domain prior knowledge and specific requirements, referring in particular to regularization methods for machine learning—one of the better established approaches that play key roles in high-dimensional learning—and covers different application scenarios, including human-robot interaction, activity recognition, and object categorization.

Photo of Francesca Odone

Francesca Odone

University of Genova

Francesca Odone is an associate professor of computer science at the University of Genova, Italy. Francesca’s research interests are in the fields of computer vision and machine learning. In particular, most of her research activity in recent years has been devoted to finding good visual representations able to capture the complexity of a problem, while allowing for the design of systems with the ability to perform their visual tasks in real time. In this respect, she has been involved in learning representations for high-dimensional data, (structured) feature selection, dimensionality reduction, support set estimation, visual recognition pipelines for object detection, retrieval, and recognition in images and image sequences, algorithms for behavior understanding, and action recognition. Francesca received a laurea degree in information sciences and a PhD in computer science, both from the University of Genova. She was a visiting student at Heriot-Watt University, Edinburgh, UK, with a EU Marie Curie research grant, as well as a researcher at the Italian National Institute for Solid State Physics. Besides theory and algorithms, Francesca also enjoys playing with real-world applications. Over the years, she has been a scientific coordinator of technology transfer and applied research projects.