Mirko Nava

Making robots understand the world using deep learning

Connect
Mirko Nava on a snowy mountain

About Me

I'm Mirko, an Artificial Intelligence researcher specialized in Self-supervised Learning and Computer Vision for mobile robots.

As a postdoctoral researcher, my time is divided between working on applied research projects, where I focus on solving challenging vision problems using AI, and advancing my research on Self-supervised autonomous robots.

I hold a PhD in AI from Univeristà della Svizzera italiana, where I developed Deep Learning techniques for autonomous robots capable of interacting with the world, collecting their own experiences, and learning from them.

Outside of work, I enjoy cooking, hiking, and working on quirky programming projects.

Research

Self-Supervised Learning

Leverage data to create meaningful training signals such that learning from them forces robots to capture structures, essential features, or relationships in the data itself.
— Louis Bouchard

Spatial Perception

Enabling robots to to perceive and visually understand outside spatial information such as features, properties, measurement, shapes, position and motion of objects in space.
— Alison Simmons

News

LEDs classification for pose estimation (CoRL)

LEDs classification for orientation estimation (IROS)

Postdoc researcher at IDSIA

LEDs classification as a Pretext task (RA-L)

Officially a Doctor

Internship at Magic Leap