Why PhD?

Rethinking Values in Computer Vision Research and Practice
with Negar Rostamzadeh

“Why PhD?” aims at giving an overview of the opportunities offered by the DISI PhD programs in Information Engineering and Computer Science (IECS) and Industrial Innovation (IID) through an inspiring talk by our alumna Negar Rostamzadeh, now Senior Research Scientist at Google Responsible team.

Program

16:00 – 16:30 Rethinking Values in Computer Vision Research and Practice, with Negar Rostamzadeh (Google Research)

16:30 – 16:45 Presentation of Doctoral Programme in Information Engineering and Computer Science (IECS) by Coordinator prof. Elisa Ricci

16:45 – 17:00 Presentation of Doctorate Program in Industrial Innovation (IID) by Vice- Coordinator prof. Domenico Siracusa.

How to participate

It is possibile to participate both in presence and online on Zoom. Mandatory online registration.

Abstract

Computer Vision papers are often motivated by the importance of the research problem, and their difficulty. Broader impact of the technologies, assessment and mitigation of the potential harms of the research or products are often discussed in post development stages. In this talk, I argue the effectiveness of this approach and will be discussing the values that will help (1) emphasize important engineering and scientific practices that we should foster to increase the robustness of our research, and (2) lessons we can learn from communications between interdisciplinary fields of research.

About the speaker

Negar Rostamzadeh is a Senior Research Scientist at Google Responsible team, where she studies the social impact of machine learning technologies and evaluation systems. Prior to that, Negar was a research scientist at Element AI, where she worked on efficient sampling and labeling approaches in a variety of multimedia and computer vision problems. Negar obtained her PhD from the University of Trento, where she was advised by Dr. Nicu Sebe. Her main area of research during her PhD was on large scale video understanding problems. During her PhD, she spent two years at MILA, where she worked on attention mechanisms in videos, video captioning and generation under the supervision of Dr. Aaron Courville. Negar was a co-founder of Women in Deep Learning (WiDL) in 2016, and a co-organizer of WiML, WiCV and WiDL in 2017. She served as a DEI chair for ICCV 2021. She was also co-founder and chair of multiple workshop series in CVPR, such as Computer Vision for Fashion, Art and Design, Ethical Considerations in Creative Applications of Computer Vision, and Responsible Generative AI, as well as Science meets Engineering of Deep Learning workshop series at ICLR, NeurIPS and FAccT.