How to connect perception, learning and action in a machine that interacts with the physical world.

Artificial intelligence, vision, embedded systems, mechatronics, bio-inspiration, and soft robotics.
Cross-cutting technologies and methods, linked to application domains rather than presented as exclusive uses.
Bio-inspiration borrows principles from living organisms; soft robotics studies in particular deformable structures. They overlap without being synonymous. Explore their cross-cutting report →
Robotics combines movement, perception and control. Inria’s educational path helps understand this articulation. A conversational capability alone is not enough to establish that a robot can manipulate objects or act reliably in a real environment.
The work presented by the Flowers team explores in particular learning and interactions between humans and robots. To judge a result, one must look at the task tested, the conditions of the experiment and its limits, rather than attributing general intelligence to the machine.
When personal data is processed, the requirements of the GDPR remain relevant. The CNIL publishes guidelines for aligning AI development and data protection. The European regulation on AI adds a framework whose obligations depend on the system and its use.
From electronic turtles to vision-language-action models, discover how machines learned to perceive, plan and act. A documented timeline traces the major milestones of AI and cognition, distinguishing scientific foundations, prototypes and commercial applications.
Explore the history of AI and cognitive robotics →
Find this branch in the general tree →Our article examines NEURA Robotics’ announcement and the criteria that will make it possible to assess its real capabilities.
The NASA ASTRA trial explores how tasks are distributed among robots according to scientific interest and risks. This research work helps understand collective autonomy and the role of goals set by humans.

The AI Act applies in stages and depending on uses. The Commission notably indicates transparency obligations since August 2026; the deadlines for high-risk systems differ according to their category. The GDPR remains relevant when a robot processes personal data.
Choose a path by length of study, then explore the training courses and their institutions. These thematic groupings do not constitute job offers.
References consulted on 4 October 2026.