Luca Moschella

Luca Moschella

Machine Learning Researcher

Apple

About me

Hey there! I’m a Machine Learning Researcher at Apple. Previously, I was an intern at NVIDIA and NNAISENSE, and an ELLIS Ph.D. student at Sapienza University of Rome, co-advised by Prof. Emanuele Rodolà and Prof. Francesco Locatello.

My research explores latent communication between neural networks, a foundational concept we introduced. This key observation is fundamental in understanding the universality and reusability of neural representations. Latent communication, a central theme at the UniReps Workshop @ NeurIPS 2023, has inspired other researchers to pursue innovative and state of the art work in fields such as natural language processing, computer vision, reinforcement learning, multimodal learning, and in supervised, weakly-supervised, and unsupervised learning settings.

As a long-time Linux user and open-source enthusiast, I take great satisfaction in contributing to the community whenever possible. I find it to be an immensely rewarding experience, and I believe that collaborative efforts can help foster an environment of knowledge-sharing and innovation.

Interests
  • Latent Communication
  • Representation Learning
  • Geometric Deep Learning
  • Reasoning
  • Artificial Intelligence
Education

Research Experience

 
 
 
 
 
Apple
Machine Learning Researcher
Aug 2024 – Present Zurich, Switzerland
 
 
 
 
 
Institute of Science and Technology Austria (ISTA)
International Research Visit
Oct 2023 – Jan 2024 Klosterneuburg, Austria
  • Research visit in the Causal Learning and Artificial Intelligence laboratory led by Prof. Francesco Locatello.
 
 
 
 
 
École polytechnique
International Research Visit
Jun 2023 – Jun 2023 Paris, France
  • Research visit in the GeoViC group at the LIX research laboratory of Ecole Polytechnique led by Prof. Maks Ovsjanikov.
 
 
 
 
 
NVIDIA
Research Engineer Intern
Sep 2022 – May 2023 Toronto, Canada
  • Research on methods to apply diffusion models for fast generation of 3D scenes with only image-space supervision
 
 
 
 
 
NNAISENSE
Research Intern
Jul 2021 – Dec 2021 Lugano, Switzerland
  • Designed and implemented a generic deep learning template to bootstrap industrial projects and automate the hyperparameters tuning.
  • Improved the company infrastructure to streamline the development and deployment of deep learning solutions.
 
 
 
 
 
Deep Learning & Applied AI @Sapienza
Teaching Assistant
Jan 2020 – Present Rome, Italy
  • Designed the course lab sessions to explain and show to the students fundamentals and cutting-edge techniques in many areas of deep learning.
  • In the three years, lectured and mentored 150+ students complessively.
  • Latest course material with interactive notebooks available here.

Achievements

Unifying Representations in Neural Models (UniReps)
Program Chair at UniReps Workshop @ NeurIPS
  • Supported the ideation and organization of the Unifying Representations in Neural Models Workshop
  • Coordinated 150+ reviewers over 90+ submissions, ensuring at least two reviews for all papers with 95% having three.
Conference on Neural Information Processing Systems (NeurIPS)
Financial Assistance NeurIPS 2023
  • Selected and awarded to support the participation at the NeurIPS 2023 conference
International Conference on Computer Vision (ICCV)
Outstanding Reviewer
International Conference on Learning Representations (ICLR)
ICLR notable top 5%
  • First-author paper selected as a notable top 5% paper at ICLR 2023 and offered an oral presentation
  • Introduce the concept of latent communication, a core focus of the UniReps Workshop @ NeurIPS 2023
  • Joint project on multimodal 3D and NLP applications between our research group GLADIA at Sapienza and Maks Ovsjanikov’s group at Ecole Polytechnique. PI: Simone Melzi, Maks Ovsjanikov. Budget: 10,000€
Sapienza University of Rome
“Kickstarting Research” Grant
  • Selected among first and second year Ph.D. students to support the development of an international career. 2020, Jun

Timeline

*
Eleventh International Conference on Learning Representations (ICLR 2023)
Eleventh International Conference on Learning Representations (ICLR 2023)
Lightning Developer Conference 2022
The 43rd annual conference of the european association for computer graphics (Eurographics 2022)
Mediterranean Machine Learning summer school
ELLIS Workshop on Geometric and Relational Deep Learning

Contact