Anca Dragan is an Assistant Professor in the EECS Department at UC Berkeley. Her goal is to enable robots to work with, around, and in support of people. She runs the InterACT Lab, where the focus is on algorithms for human-robot interaction -- algorithms that move beyond the robot's function in isolation, and generate robot behavior that also accounts for interaction and coordination with end-users. The lab works across different applications, from assistive robots, to manufacturing, to autonomous cars, and draw from optimal control, planning, estimation, learning, and cognitive science. She also helped found and serve on the steering committee for the Berkeley AI Research (BAIR) Lab, and is a co-PI of the Center for Human-Compatible AI. She has been honored by the Sloan Fellowship, MIT TR35, the Okawa award, and an NSF CAREER award.
Talk # 1
Optimal robot action for and around people
The traditional robotics problem is one of optimization. An engineer writes down a cost function and potentially a set of constraints, thereby specifying what it means for a robot to accomplish its task. The robot then is in charge of finding the behavior that is optimal for this specification. Thus, the focus in robotics is on how a robot can produce optimal (or even feasible) behavior despite the intricacies of operating in the real world. What drives my research is the realization that we are not building robots to work in some isolated universe, optimizing some exogenously specified cost function. We are building them to work in our universe, in order to help us. First, robots will not act in isolation. They will work with and around us. This makes optimal action in isolation far from sufficient – robots will need to choose actions that mesh well with ours. If there were no people on the roads, autonomous driving would be nearly solved. Instead, cars need to coordinate with us. So do quadrotors flying in our spaces, or assistive arms in our homes. My work formalizes the problem of optimal coordination with people, and introduces real-time solutions for continuous and high-dimensional state and action spaces. Second, robots will need to do what we want them to. This makes the notion of some exogenously specified cost function a myth. Cost functions don’t just fall from the sky and incentivize the robot behavior we want. Thus, figuring out how to optimize is only half the battle. The other half is figuring out what to optimize in the first place. And the key to that lies with us, people – what we want is the very definition of the cost function. My work casts the process of the robot acquiring its cost function as a human-robot collaboration, introducing theory and tools for aligning robot incentives with human preferences and ensuring that robots are resilient to changes in their environment.
Gerhard Neumann is a Professor of Robotics & Autonomous Systems at the University of Lincoln. Before coming to Lincoln, he has been an Assistant Professor at the TU Darmstadt from September 2014 to October 2016 and head of the Computational Learning for Autonomous Systems (CLAS) group. Before that, he was Post-Doc and Group Leader at the Intelligent Autonomous Systems Group
(IAS) also in Darmstadt under the guidance of Prof. Jan Peters. Gerhard obtained his Ph.D. under the supervision of Prof. Wolfgang Mass at the Graz University of Technology. Gerhard already authored 50+ peer reviewed papers, many of them in top ranked machine learning and robotics journals or conferences such as NIPS, ICML, ICRA, IROS, JMLR, Machine Learning and AURO. He is principle investigator for the National Center for Nuclear Robotics (NCNR) in Lincoln which is an EPSRC RAI Hub and also leading 1 Innovate UK project on Tomato Picking. In Darmstadt, he is principle investigator of the EU H2020 project Romans. He organized several workshops and is senior program committee for several conferences.
Talk # 1
Information-Geometric Policy Search for Learning Versatile, Reusable Skills
In the future, autonomous robots will be used for various applications such as autonomous farming, handling dangerous materials as for example decommissioning nuclear waste, health care or autonomous transportation. For such complex scenarios, it is inevitable that autonomous robots are equipped with sophisticated learning capabilities which enable it to learn from human teachers as well as from self-improvement. In this talk, I will present our work on information-geometric policy search methods for learning complex motor skills. Our algorithms use information-geometric insights to exploit curvature and path information in order to perform efficient local search at the level of single elemental motions, also called movement primitives. Simultaneously to local search, the algorithms search on a global level by selecting between distinct solutions, allowing us to represent a versatile solution space with high quality solutions. Our algorithms can be used to efficiently learn motor skills, generalize these motions to different situations, learn reactive skills that can react to perturbations and select and learn when to switch between these motions. I will also briefly show how to extend our algorithms to learn from preference-based feedback instead of a numeric reward signal, enabling a human expert to guide the learning agent without the need for manual reward tuning. While I will use dynamic motor games, such as table tennis, as motivation throughout my talk, I will also shortly present how to apply similar methods for robot grasping and manipulation tasks.