@inproceedings{mavridis2022CDC,
  title={Sparse Gaussian Process Regression using Progressively Growing Learning Representations},
  author={Mavridis, Christos N and Kontoudis, George P and Baras, John S},
  booktitle={IEEE Conference on Decision and Control (CDC)},
  abstract = {We present a new sparse Gaussian process regression model whose covariance function is parameterized by the locations of a progressively growing set of pseudo-inputs generated by an online deterministic annealing optimization algorithm. A series of entropy-regularized optimization problems is solved sequentially, introducing a bifurcation phenomenon, according to which, pseudo-inputs are gradually generated. This results in an active learning approach, which, in contrast to most existing works, can modify already selected pseudoinputs and is trained using a recursive gradient-free stochastic approximation algorithm. Finally, the proposed algorithm is able to incorporate prior knowledge in the form of a probability density, according to which new observations are sampled. Experimental results showcase the effcacy and potential advantages of the proposed methodology.},
  keywords={Gaussian processes, active learning,gradient-free optimization},
  pages = {1454--1459},
  year={2022},
  keywords={Gaussian processes, active learning, gradient-free optimization},
  url_pdf = {CDC22_Mavridis_SparseGaussianProcessRegressionProgressivelyGrowingDatasets.pdf},
  doi = {10.1109/CDC51059.2022.9992933}
}

{"_id":"vZfMmiowhvT5nfGKp","bibbaseid":"mavridis-kontoudis-baras-sparsegaussianprocessregressionusingprogressivelygrowinglearningrepresentations-2022","author_short":["Mavridis, C. N","Kontoudis, G. P","Baras, J. S"],"bibdata":{"bibtype":"inproceedings","type":"inproceedings","title":"Sparse Gaussian Process Regression using Progressively Growing Learning Representations","author":[{"propositions":[],"lastnames":["Mavridis"],"firstnames":["Christos","N"],"suffixes":[]},{"propositions":[],"lastnames":["Kontoudis"],"firstnames":["George","P"],"suffixes":[]},{"propositions":[],"lastnames":["Baras"],"firstnames":["John","S"],"suffixes":[]}],"booktitle":"IEEE Conference on Decision and Control (CDC)","abstract":"We present a new sparse Gaussian process regression model whose covariance function is parameterized by the locations of a progressively growing set of pseudo-inputs generated by an online deterministic annealing optimization algorithm. A series of entropy-regularized optimization problems is solved sequentially, introducing a bifurcation phenomenon, according to which, pseudo-inputs are gradually generated. This results in an active learning approach, which, in contrast to most existing works, can modify already selected pseudoinputs and is trained using a recursive gradient-free stochastic approximation algorithm. Finally, the proposed algorithm is able to incorporate prior knowledge in the form of a probability density, according to which new observations are sampled. Experimental results showcase the effcacy and potential advantages of the proposed methodology.","keywords":"Gaussian processes, active learning, gradient-free optimization","pages":"1454–1459","year":"2022","url_pdf":"CDC22_Mavridis_SparseGaussianProcessRegressionProgressivelyGrowingDatasets.pdf","doi":"10.1109/CDC51059.2022.9992933","bibtex":"@inproceedings{mavridis2022CDC,\n title={Sparse Gaussian Process Regression using Progressively Growing Learning Representations},\n author={Mavridis, Christos N and Kontoudis, George P and Baras, John S},\n booktitle={IEEE Conference on Decision and Control (CDC)},\n abstract = {We present a new sparse Gaussian process regression model whose covariance function is parameterized by the locations of a progressively growing set of pseudo-inputs generated by an online deterministic annealing optimization algorithm. A series of entropy-regularized optimization problems is solved sequentially, introducing a bifurcation phenomenon, according to which, pseudo-inputs are gradually generated. This results in an active learning approach, which, in contrast to most existing works, can modify already selected pseudoinputs and is trained using a recursive gradient-free stochastic approximation algorithm. Finally, the proposed algorithm is able to incorporate prior knowledge in the form of a probability density, according to which new observations are sampled. Experimental results showcase the effcacy and potential advantages of the proposed methodology.},\n keywords={Gaussian processes, active learning,gradient-free optimization},\n pages = {1454--1459},\n year={2022},\n keywords={Gaussian processes, active learning, gradient-free optimization},\n url_pdf = {CDC22_Mavridis_SparseGaussianProcessRegressionProgressivelyGrowingDatasets.pdf},\n doi = {10.1109/CDC51059.2022.9992933}\n}\n\n","author_short":["Mavridis, C. N","Kontoudis, G. P","Baras, J. S"],"key":"mavridis2022CDC","id":"mavridis2022CDC","bibbaseid":"mavridis-kontoudis-baras-sparsegaussianprocessregressionusingprogressivelygrowinglearningrepresentations-2022","role":"author","urls":{" pdf":"http://www.georgekontoudis.com/publications/CDC22_Mavridis_SparseGaussianProcessRegressionProgressivelyGrowingDatasets.pdf"},"keyword":["Gaussian processes","active learning","gradient-free optimization"],"metadata":{"authorlinks":{}},"downloads":1,"html":""},"bibtype":"inproceedings","biburl":"http://www.georgekontoudis.com/publications/Kontoudis_mypubs.bib","dataSources":["ENfCwQrwsiuEJtsqK","ya2CyA73rpZseyrZ8","2252seNhipfTmjEBQ"],"keywords":["gaussian processes","active learning","gradient-free optimization"],"search_terms":["sparse","gaussian","process","regression","using","progressively","growing","learning","representations","mavridis","kontoudis","baras"],"title":"Sparse Gaussian Process Regression using Progressively Growing Learning Representations","year":2022,"downloads":1}