Metamemory, defined as knowing about memory and mnemonic strategies, is an especially important form of metacognition. Scientific models are often prone to distancing the observer from the object or field of study whereas a metacognitive model in general tries to include the observer in the model. A metacognitive model differs from other scientific models in that the creator of the model is per definition also enclosed within it. There are generally two components of metacognition: (1) knowledge about cognition and (2) regulation of cognition. Metacognition can take many forms, such as reflecting on one's ways of thinking and knowing when and how to use particular strategies for problem-solving. The term comes from the root word meta, meaning "beyond", or "on top of". Metacognition is an awareness of one's thought processes and an understanding of the patterns behind them. Peters, J., Muelling, K., Altun, Y.: Relative entropy policy search.Thinking about thinking, higher-order thinking skills Geman, S., Geman, D.: Stochastic relaxation, Gibbs distributions, and the Bayesian restoration of images. George, E.I., McCulloch, R.E.: Variable selection via Gibbs sampling. Schaal, S., Peters, J., Nakanishi, J., Ijspeert, A.: Learning movement primitives. Kroemer, O., Sukhatme, G.S.: Learning relevant features for manipulation skills using meta-level priors. Rother, C., Kolmogorov, V., Blake, A.: GrabCut -interactive foreground extraction using iterated graph cuts. Metzen, J.H., Fabisch, A., Senger, L., de Gea Fernandez, J., Kirchner, E.A.: Towards learning of generic skills for robotic manipulation. JMLR 9, 2349–2376 (2008)ĭeisenroth, M.P., Englert, P., Peters, J., Fox, D.: Multi-task policy search for robotics. Krupka, E., Navot, A., Tishby, N.: Learning to select features using their properties. Lee, S., Chatalbashev, V., Vickrey, D., Koller, D.: Learning a meta-level prior for feature relevance from multiple related tasks. Piater, J., Jodogne, S., Detry, R., Kraft, D., Krueger, N., Kroemer, O., Peters, J.: Learning visual representations for perception-action systems. Levine, S., Finn, C., Darrell, T., Abbeel, P.: End-to-end training of deep visuomotor policies. Muelling, K., Kober, J., Kroemer, O., Peters, J.: Learning to select and generalize striking movements in robot table tennis. 32(11), 1238–1274 (2013)ĭa Silva, B.C., Baldassarre, G., Konidaris, G.D., Barto, A.: Learning parameterized motor skills on a humanoid robot. Kober, J., Bagnell, D., Peters, J.: Reinforcement learning in robotics: a survey. KeywordsĪrgall, B.D., Chernova, S., Veloso, M., Browning, B.: A survey of robot learning from demonstration. Our robot experiments show that using a meta-level prior results in better generalization performance and more efficient skill learning. The meta-level prior is transferred to new skills using meta features. This prior is computed using a meta-level prior, which is learned from previous skills. An informative prior over the features’ relevance can guide the robot’s feature selection process. Rather than relying on a human to select the relevant features, our work focuses on incorporating feature selection into the skill learning process. However, only a sparse set of the objects’ features will be relevant for generalizing the manipulation skill between different scenarios and objects. the position or length of an action-relevant part of an object. Manipulation skills need to adapt to the geometric features of the objects that they are manipulating, e.g.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |