I develop machine learning and statistical tools to study cooperative phenomena occurring in brain activity at multiple scales.
I explore programmable self-assembly as an active platform for high throughput, parallel, and distributed (swarm-like) micro- and nano- robotic operations and for high volume manufacturing of multi-component micro- and nano-robots.
My research focuses on causal inference and large-scale machine learning and how these fields can benefit from each other. More efficient computation and approximate inference enables novel methods for causality and establishing causal relations inevitably leads to more robust predictions of effects in changing environments.
My research is focused on the intelligent control of technical systems by combining traditional approaches with advances from machine learning. Specifically, I am working on data driven model improvement for the use in model predictive control (MPC).
I am working on functionalizing magnetic microrobots for bio-medical application, with special emphasis on tetherless energy transfer to actuate microrobot for applications such as hyperthermia and shape transformation
My research interests are computer vision, discrete and continuous optimization, and machine learning for physics based simulations.
I have a strong interest in optimization, probabilistic graphical models and deep learning. I am also interested in developing new machine learning models for text and image understanding.
My current research focuses on analyzing dynamic visual scenes and aims to develop efficient and robust methods for the geometric and semantic analysis of scenes captured by multiple videos.
My research focuses on micro-robots that propel through complex biological fluids. I am working on new micro/nano fabrication methods, actuation principles in low Reynolds number hydrodynamics, and exploring potential biomedical applications.
My current research focuses on reconstructing dense 3-d surface geometry, appearance and motion from image sequences.
I am working on reinforcement learning and stochastic optimal control for robotics. In particular, I am interested in finding bounds for approximate solutions, in the sense of probably approximately correct learning.
My research goal is the design and development of bio-hybrid micro-robots and micro-actuators powered by programmed muscle cell contraction
I am interested in trial-and-error processes of reinforcement learning and how a single rendition of a specific action can shape and improve future renditions. To this aim we study birdsong learning and develop dynamical system models to unravel the influence of reward, variability and sensory target on song learning trajectories.
My research interests are in the development of flexible and soft sensors and actuators as well as the design of soft-bodied robots that can couple sensing and actuation.
My research interests include large-scale learning, classification of data to large output spaces aka extreme classification, deep learning, and optimization.
My research focuses on perception for autonomous robotic manipulation and grasping. I am specifically interested in developing methods that are goal-directed, real-time and multi-modal such that they can provide meaningful feedback for execution and learning.
My current research interests involve using and extending statistical models to combine heterogeneous biological data sources in order to understand cellular regulation in cancer.
My research concerns bio-inspired construction by collectives of robots. I have designed a system of robots autonomously assembling three-dimensional structures. Now, I aim to extend the potential for long term autonomy by introducing soft durable robots to the field of autonomous construction.
I am fascinated by the basic physical learning processes that take place in nature at the micro-scale (e.g. the adaptive emerging feature of metachronal wave in ciliary propulsion); my research focuses on implementing this kind of features in artificial micro-actuators and advanced micro-robots.
My scientific work focuses on causal inference from empirical data: I develop both theory and methodology that allow us to learn causal relationships from observational (and interventional) data.
I am interested in contact free manipulation based on controlling magnetic interactions between a magnetic field source and a magnetic device. I explore how to model and control devices with an eight-coil magnetic manipulation system designed for human-scale catheter guidance.