Federico Baldassarre

Federico Baldassarre

PhD Student in Deep Learning

KTH Stockholm

Biography

I am a PhD student at the Royal Institute of Technology (KTH) in Stockholm, Sweden. My doctoral studies began in September 2018 at the division of Robotics, Perception and Learning, under the supervision of Prof. Hossein Azizpour.

My research focuses on the explainability of deep learning, improving our understanding of modern learning systems from their internal representations to the way they determine a prediction.
The long-term ambition is to be able to approach AI agents as peers, to whom we’d be able to ask questions and observe the reasoning process. This, in turn, would allow developers to easily debug issues and biases, and the general public to trust AI more in their everyday life.

As part of my PhD experience, I volunteer as sysadmin of our in-house compute cluster. In a team of three, we helped procure servers with 70+ GPUs and actively manage a SLURM cluster for 100+ users.

Also, I supervise students in their master’s theses and work as teacher assistant for the following courses:

Beforehand, I studied a master’s in Machine Learning at KTH, during which I interned at a foodtech startup in Kuala Lumpur, Malaysia, and joined the Zalando Research team for my thesis in Berlin, Germany.

Interests
  • Explainability and reasoning
  • Self-supervised representation learning
  • Computer vision on natural images/videos
  • Graph networks and transformers
Education
  • PhD in Deep Learning, 2018 - Present

    KTH - Royal Institute of Technology

  • MSc in Machine Learning, 2016 - 2018

    KTH - Royal Institute of Technology

  • BSc in Computer Engineering, 2013 - 2016

    UNIBO - University of Bologna

Experience

 
 
 
 
 
FAIR (Meta AI)
Research Scientist Intern
Jun 2021 – Sep 2021 Paris, France
 
 
 
 
 
Huawei Research
Research Scientist Intern
Jun 2021 – Sep 2021 Dublin, Ireland
 
 
 
 
 
Zalando Research
Research Scientist Intern
Jan 2018 – Jul 2018 Berlin, Germany
 
 
 
 
 
Pop Meals (formerly Dahmakan)
Data Scientist
Jun 2017 – Aug 2017 Kuala Lumpur, Malaysia

Accomplish­ments

Travel Grant
Awarded a travel grant from the Karl Engvers Foundation to attend the BMVC 2022 conference and present the paper “Quantitative Metrics for Evaluating Explanations of Video DeepFake Detectors”.
Outstanding Reviewer
Ranked among the top 12 reviewers of the ECCV 2020 conference, according to the quality of the reviews.
See certificate
Travel Grant
Awarded a travel grant from the Karl Engvers Foundation to attend the ICLR 2020 conference.
Best Course Project
Popular project: “Grayscale image colorization using a CNN conditione on Inception-ResNet-v2”.
350+ GitHub stars, endorsed by François Chollet, several spinoffs, scientific citations, and blog posts.

Education

 
 
 
 
 
KTH - Royal Institute of Technology
PhD Student
Sep 2018 – Present Stockholm, Sweden
 
 
 
 
 
KTH - Royal Institute of Technology
Master’s degree
Aug 2016 – Jul 2018 Stockholm, Sweden
 
 
 
 
 
Unibo - University of Bologna
Bachelor’s degree
Sep 2013 – Jul 2016 Bologna, Italy

Other activities

Climbing
Cooking
Photography

Curriculum Vitae