Robotics and Biology Laboratory

Vito Mengers

Wiss. Mitarbeiter_in

v.mengers@tu-berlin.de

Gebäude MAR
Raum MAR 5.065

Office Hours

Only by appointment.

Projects

© RBO

​Differentiable Interconnected Recursive Estimation as a Principle of Intelligence

We propose a computational principle for mapping sensory inputs to suitable actions, consisting of three building blocks: recursive estimators, interconnections, and differentiable programming. This integrated system can extract task-relevant information from the sensory input and generate suitable actions to achieve complex goals. We seek to study it as a model for different intelligent behaviors, thereby proposing it as a more general principle of intelligence.

Research Interests

While most of my experience lies within robotics, I am interested in the broader field of intelligence research. The central question of my research is how the mechanisms work that enable humans and animals to behave in such a robust, adaptive, and goal-directed -- we might just say intelligent -- way. Many research fields need to contribute to answer this question, and thus I like to work as interdisciplinary as possible. However, the central part of my contribution lies in implementing different models of the mechanisms of intelligent behavior in synthetic agents, usually robots that act in the real world. Using these synthetic agents, we can quickly see how changes in the model affect behavior, a route towards insight which is usually unavailable when working with biological subjects.
Currently, I am investigating a computational principle from robotics as a candidate for a more general principle of intelligence. The candidate principle links perception directly with action: We use multiple interconnected recursive estimators to extract relevant information from sensory inputs and generate actions directly from these estimators, leveraging their differentiable implementation. In order to validate it as a more general principle, we apply it to a wide spectrum of intelligence research in collaboration with different research projects within the cluster of excellence Science of Intelligence.

Short CV

  • Since 2021
    • PhD-student at Robotics and Biology Laboratory (Oliver Brock), TU Berlin
  • Since 2021
  • 2018 2021
    • M.Sc. Computer Engineering, TU Berlin
  • 2015 2019
    • B.Sc. Computer Engineering, TU Berlin

Publications

2023

Mengers, Vito; Battaje, Aravind; Baum, Manuel; Brock, Oliver
Combining Motion and Appearance for Robust Probabilistic Object Segmentation in Real Time
2023 IEEE International Conference on Robotics and Automation (ICRA)
IEEE
2023