Rika Antonova: Adaptive Kernels and Priors for Data-efficient Bayesian Optimization
Abstract
This talk will describe a framework for constructing simulation-based kernels for ultra data-efficient Bayesian optimization (BO). We consider challenging settings that allow only 10-20 hardware trials, with access to only mid/low-fidelity simulators. First, I will describe how we can construct an informed kernel by embedding the space of simulated trajectories into a lower-dimensional space of latent paths. Our sequential variational autoencoder handles large-scale learning from ample simulated data; its modular design ensures quick adaptation to close the sim-to-real gap. The approach does not require domain-specific knowledge. Hence, we are able to demonstrate on hardware that the same architecture works for different areas of robotics: locomotion and manipulation. For domains with severe sim-to-real mismatch, I will describe our variant of BO which ensures that discrepancies between simulation and reality do not hinder online adaption. Using task-oriented grasping as an example, I will demonstrate how this approach helps quick recovery in case of corrupted/degraded simulation. My longer-term research vision is to build priors from simulation without requiring a specific simulation scenario. So, I will conclude by providing the motivation for this direction, and will describe our initial work on ‘timescale fidelity’ priors. Such priors could help transfer-aware models and simulators to automatically adjust their timestep/frequency or planning horizon.
Bio
Currently, I am a PhD student at KTH (at the Robotics, Perception & Learning division in the group headed by Danica Kragic). My research is focused on constructing simulation-based kernels to improve data efficiency of Bayesian optimization (BO) for dynamic manipulation tasks, task-oriented grasping and locomotion. More generally, I am interested in reinforcement learning for robotics domains with a challenging sim-to-real gap. Previously, I was a research Masters student at the Robotics Institute at Carnegie Mellon University. At CMU I worked on BO for bipedal locomotion (with Akshara Rai and Chris Atkeson) and on improving data efficiency of BO for personalized education systems (with Emma Brunskill - my MS adviser). Earlier, I was a software engineer at Google. I started in the Personalized Search team in core ranking, then joined the Character Recognition (OCR) team and worked on the open-source OCR engine Tesseract, which was used for Google Books and StreetView.