Sign In

Carlos Florensa

Research Scientist at Covariant

Carlos Florensa is a Research Scientist at Covariant, a company focused on developing AI for robotic applications.23 He joined Covariant in February 2020 after completing his PhD in Artificial Intelligence at UC Berkeley.23

Florensa's research focuses on enabling robots to operate in variable environments with minimal supervision. He specializes in Deep Reinforcement Learning algorithms, incorporating policy hierarchy, few-shot learning, and automatic curriculum generation to solve complex locomotion and manipulation tasks.3

Education and Career Highlights

  • PhD in Artificial Intelligence from UC Berkeley (2015-2020)23
  • Double degree in Mathematics and Industrial Engineering from Polytechnic University of Catalonia (2010-2014)2
  • Research internships at DeepMind (2018) and NVIDIA (2019)2
  • Visiting Research Scholar at Carnegie Mellon University (2014-2015)2

Research Contributions

Florensa has made significant contributions to the field of AI and robotics:

  • Authored papers on automatic goal generation, reverse curriculum generation, and stochastic neural networks for hierarchical reinforcement learning1
  • Developed the Guided Uncertainty-Aware Policy Optimization (GUAPO) algorithm, which received the best paper award at the Robot Learning workshop at NeurIPS 20192
  • Worked on self-supervised learning of image embedding for continuous control at DeepMind2

At Covariant, Florensa continues to work on developing universal AI that allows robots to see, reason, and act on the world around them, with a focus on material-handling robots in warehouses.3

Highlights

Mar 11 · sites.google.com
Carlos Florensa - Google Sites
‪Carlos Florensa‬ - ‪Google Scholar‬
Feb 26 · x.com
Carlos Florensa (@CarlosFlorensa) / X
Carlos Florensa (@CarlosFlorensa) / X

Related Questions

What are some of Carlos Florensa's most notable research contributions?
How did Carlos Florensa's work at NVIDIA influence his career?
What is the Covariant Brain, and how does it work?
Can you explain the concept of Guided Uncertainty-Aware Policy Optimization (GUAPO)?
What challenges does Carlos Florensa face in variable robotics environments?
Carlos Florensa
Carlos Florensa, photo 1
Carlos Florensa, photo 2
Add to my network

Location

Berkeley, California, United States