You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

zotero_export.bib 20KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301
  1. @article{barto_neuronlike_1983,
  2. abstract = {It is shown how a system consisting of two neuronlike adaptive elements can solve a difficult learning control problem. The task is to balance a pole that is hinged to a movable cart by applying forces to the cart's base. It is argued that the learning problems faced by adaptive elements that are components of adaptive networks are at least as difficult as this version of the pole-balancing problem. The learning system consists of a single associative search element (ASE) and a single adaptive critic element (ACE). In the course of learning to balance the pole, the ASE constructs associations between input and output by searching under the influence of reinforcement feedback, and the ACE constructs a more informative evaluation function than reinforcement feedback alone can provide. The differences between this approach and other attempts to solve problems using neurolike elements are discussed, as is the relation of this work to classical and instrumental conditioning in animal learning studies and its possible implications for research in the neurosciences.},
  3. author = {Barto, A. G. and Sutton, R. S. and Anderson, C. W.},
  4. date = {1983-09},
  5. doi = {10.1109/TSMC.1983.6313077},
  6. issn = {0018-9472},
  7. journaltitle = {IEEE Transactions on Systems, Man, and Cybernetics},
  8. keywords = {adaptive control,adaptive critic element,Adaptive systems,animal learning studies,associative search element,Biological neural networks,learning control problem,learning systems,movable cart,neural nets,neuronlike adaptive elements,Neurons,Pattern recognition,Problem-solving,Supervised learning,Training},
  9. number = {5},
  10. pages = {834-846},
  11. title = {Neuronlike Adaptive Elements That Can Solve Difficult Learning Control Problems},
  12. volume = {SMC-13}
  13. }
  14. @report{bishop_mixture_1994,
  15. author = {Bishop, Christopher M.},
  16. date = {1994},
  17. title = {Mixture Density Networks}
  18. }
  19. @article{bodin_latent_2017,
  20. abstract = {We introduce Latent Gaussian Process Regression which is a latent variable extension allowing modelling of non-stationary processes using stationary GP priors. The approach is built on extending the input space of a regression problem with a latent variable that is used to modulate the covariance function over the input space. We show how our approach can be used to model non-stationary processes but also how multi-modal or non-functional processes can be described where the input signal cannot fully disambiguate the output. We exemplify the approach on a set of synthetic data and provide results on real data from geostatistics.},
  21. archivePrefix = {arXiv},
  22. author = {Bodin, Erik and Campbell, Neill D. F. and Ek, Carl Henrik},
  23. date = {2017-07-18},
  24. eprint = {1707.05534},
  25. eprinttype = {arxiv},
  26. keywords = {Computer Science - Learning,Statistics - Machine Learning},
  27. primaryClass = {cs, stat},
  28. title = {Latent {{Gaussian Process Regression}}},
  29. url = {http://arxiv.org/abs/1707.05534},
  30. urldate = {2017-08-29}
  31. }
  32. @article{brockman_openai_2016,
  33. abstract = {OpenAI Gym is a toolkit for reinforcement learning research. It includes a growing collection of benchmark problems that expose a common interface, and a website where people can share their results and compare the performance of algorithms. This whitepaper discusses the components of OpenAI Gym and the design decisions that went into the software.},
  34. archivePrefix = {arXiv},
  35. author = {Brockman, Greg and Cheung, Vicki and Pettersson, Ludwig and Schneider, Jonas and Schulman, John and Tang, Jie and Zaremba, Wojciech},
  36. date = {2016-06-05},
  37. eprint = {1606.01540},
  38. eprinttype = {arxiv},
  39. keywords = {Computer Science - Artificial Intelligence,Computer Science - Machine Learning},
  40. primaryClass = {cs},
  41. title = {{{OpenAI Gym}}},
  42. url = {http://arxiv.org/abs/1606.01540},
  43. urldate = {2018-09-24}
  44. }
  45. @article{choi_choicenet_2018,
  46. abstract = {In this paper, we focus on the supervised learning problem with corrupted training data. We assume that the training dataset is generated from a mixture of a target distribution and other unknown distributions. We estimate the quality of each data by revealing the correlation between the generated distribution and the target distribution. To this end, we present a novel framework referred to here as ChoiceNet that can robustly infer the target distribution in the presence of inconsistent data. We demonstrate that the proposed framework is applicable to both classification and regression tasks. ChoiceNet is evaluated in comprehensive experiments, where we show that it constantly outperforms existing baseline methods in the handling of noisy data. Particularly, ChoiceNet is successfully applied to autonomous driving tasks where it learns a safe driving policy from a dataset with mixed qualities. In the classification task, we apply the proposed method to the MNIST and CIFAR-10 datasets and it shows superior performances in terms of robustness to noisy labels.},
  47. archivePrefix = {arXiv},
  48. author = {Choi, Sungjoon and Hong, Sanghoon and Lim, Sungbin},
  49. date = {2018-05-16},
  50. eprint = {1805.06431},
  51. eprinttype = {arxiv},
  52. keywords = {Computer Science - Machine Learning,Statistics - Machine Learning},
  53. primaryClass = {cs, stat},
  54. shorttitle = {{{ChoiceNet}}},
  55. title = {{{ChoiceNet}}: {{Robust Learning}} by {{Revealing Output Correlations}}},
  56. url = {http://arxiv.org/abs/1805.06431},
  57. urldate = {2018-08-23}
  58. }
  59. @inproceedings{choi_robust_2016,
  60. author = {Choi, Sungjoon and Lee, Kyungjae and Oh, Songhwai},
  61. booktitle = {Robotics and {{Automation}} ({{ICRA}}), 2016 {{IEEE International Conference}} On},
  62. date = {2016},
  63. pages = {470--475},
  64. publisher = {{IEEE}},
  65. title = {Robust Learning from Demonstration Using Leveraged {{Gaussian}} Processes and Sparse-Constrained Optimization}
  66. }
  67. @inproceedings{damianou_deep_2013,
  68. abstract = {In this paper we introduce deep Gaussian process (GP) models. Deep GPs are a deep belief network based on Gaussian process mappings. The data is modeled as the output of a multivariate GP. The inpu...},
  69. author = {Damianou, Andreas and Lawrence, Neil},
  70. booktitle = {Artificial {{Intelligence}} and {{Statistics}}},
  71. date = {2013-04-29},
  72. eventtitle = {Artificial {{Intelligence}} and {{Statistics}}},
  73. keywords = {60G15; 58E30,Computer Science - Learning,G.1.2,G.3,I.2.6,Mathematics - Probability,Statistics - Machine Learning},
  74. langid = {english},
  75. pages = {207-215},
  76. title = {Deep {{Gaussian Processes}}},
  77. url = {http://proceedings.mlr.press/v31/damianou13a.html},
  78. urldate = {2018-10-02}
  79. }
  80. @inproceedings{depeweg_decomposition_2018,
  81. author = {Depeweg, Stefan and Hernandez-Lobato, Jose-Miguel and Doshi-Velez, Finale and Udluft, Steffen},
  82. booktitle = {International {{Conference}} on {{Machine Learning}}},
  83. date = {2018},
  84. pages = {1192--1201},
  85. title = {Decomposition of {{Uncertainty}} in {{Bayesian Deep Learning}} for {{Efficient}} and {{Risk}}-Sensitive {{Learning}}}
  86. }
  87. @article{depeweg_learning_2016,
  88. abstract = {We present an algorithm for model-based reinforcement learning that combines Bayesian neural networks (BNNs) with random roll-outs and stochastic optimization for policy learning. The BNNs are trained by minimizing alpha-divergences, allowing us to capture complicated statistical patterns in the transition dynamics, e.g. multi-modality and heteroskedasticity, which are usually missed by other common modeling approaches. We illustrate the performance of our method by solving a challenging benchmark where model-based approaches usually fail and by obtaining promising results in a real-world scenario for controlling a gas turbine.},
  89. archivePrefix = {arXiv},
  90. author = {Depeweg, Stefan and Hernández-Lobato, José Miguel and Doshi-Velez, Finale and Udluft, Steffen},
  91. date = {2016-05-23},
  92. eprint = {1605.07127},
  93. eprinttype = {arxiv},
  94. keywords = {Computer Science - Learning,Statistics - Machine Learning},
  95. primaryClass = {cs, stat},
  96. title = {Learning and {{Policy Search}} in {{Stochastic Dynamical Systems}} with {{Bayesian Neural Networks}}},
  97. url = {http://arxiv.org/abs/1605.07127},
  98. urldate = {2016-06-06}
  99. }
  100. @inproceedings{hein_benchmark_2017,
  101. abstract = {In the research area of reinforcement learning (RL), frequently novel and promising methods are developed and introduced to the RL community. However, although many researchers are keen to apply their methods on real-world problems, implementing such methods in real industry environments often is a frustrating and tedious process. Generally, academic research groups have only limited access to real industrial data and applications. For this reason, new methods are usually developed, evaluated and compared by using artificial software benchmarks. On one hand, these benchmarks are designed to provide interpretable RL training scenarios and detailed insight into the learning process of the method on hand. On the other hand, they usually do not share much similarity with industrial real-world applications. For this reason we used our industry experience to design a benchmark which bridges the gap between freely available, documented, and motivated artificial benchmarks and properties of real industrial problems. The resulting industrial benchmark (IB) has been made publicly available to the RL community by publishing its Java and Python code, including an OpenAI Gym wrapper, on Github. In this paper we motivate and describe in detail the IB's dynamics and identify prototypic experimental settings that capture common situations in real-world industry control problems.},
  102. author = {Hein, D. and Depeweg, S. and Tokic, M. and Udluft, S. and Hentschel, A. and Runkler, T. A. and Sterzing, V.},
  103. booktitle = {2017 {{IEEE Symposium Series}} on {{Computational Intelligence}} ({{SSCI}})},
  104. date = {2017-11},
  105. doi = {10.1109/SSCI.2017.8280935},
  106. eventtitle = {2017 {{IEEE Symposium Series}} on {{Computational Intelligence}} ({{SSCI}})},
  107. keywords = {academic research groups,artificial software benchmarks,Automobiles,benchmark environment,Benchmark testing,frustrating process,Games,Helicopters,industrial control,Industrial control,industrial data,Industries,industry environments,industry experience,interpretable RL training scenarios,learning (artificial intelligence),learning process,motivated artificial benchmarks,public domain software,real-world applications,real-world industry control problems,reinforcement learning,research area,RL community,tedious process,Wind turbines},
  108. pages = {1-8},
  109. title = {A Benchmark Environment Motivated by Industrial Control Problems}
  110. }
  111. @inproceedings{hensman_gaussian_2013,
  112. author = {Hensman, James and Fusi, Nicolo and Lawrence, Neil D.},
  113. booktitle = {Uncertainty in {{Artificial Intelligence}}},
  114. date = {2013},
  115. keywords = {Computer Science - Learning,Statistics - Machine Learning},
  116. pages = {282},
  117. publisher = {{Citeseer}},
  118. title = {Gaussian {{Processes}} for {{Big Data}}}
  119. }
  120. @article{hensman_scalable_2015,
  121. abstract = {Gaussian process classification is a popular method with a number of appealing properties. We show how to scale the model within a variational inducing point framework, outperforming the state of the art on benchmark datasets. Importantly, the variational formulation can be exploited to allow classification in problems with millions of data points, as we demonstrate in experiments. Copyright 2015 by the authors.},
  122. author = {Hensman, James and Matthews, Alexander G. de G. and Ghahramani, Zoubin},
  123. date = {2015},
  124. journaltitle = {Journal of Machine Learning Research},
  125. keywords = {Statistics - Machine Learning},
  126. pages = {351-360},
  127. title = {Scalable Variational {{Gaussian}} Process Classification},
  128. volume = {38}
  129. }
  130. @article{hodge_survey_2004,
  131. author = {Hodge, Victoria and Austin, Jim},
  132. date = {2004},
  133. journaltitle = {Artificial intelligence review},
  134. number = {2},
  135. pages = {85--126},
  136. title = {A Survey of Outlier Detection Methodologies},
  137. volume = {22}
  138. }
  139. @article{jacobs_adaptive_1991,
  140. author = {Jacobs, Robert A. and Jordan, Michael I. and Nowlan, Steven J. and Hinton, Geoffrey E.},
  141. date = {1991},
  142. journaltitle = {Neural computation},
  143. number = {1},
  144. pages = {79--87},
  145. title = {Adaptive Mixtures of Local Experts},
  146. volume = {3}
  147. }
  148. @incollection{kaiser_bayesian_2018,
  149. author = {Kaiser, Markus and Otte, Clemens and Runkler, Thomas and Ek, Carl Henrik},
  150. booktitle = {Advances in {{Neural Information Processing Systems}} 31},
  151. date = {2018},
  152. editor = {Bengio, S. and Wallach, H. and Larochelle, H. and Grauman, K. and Cesa-Bianchi, N. and Garnett, R.},
  153. keywords = {Computer Science - Learning,Computer Science - Machine Learning,Statistics - Machine Learning},
  154. pages = {6995--7004},
  155. publisher = {{Curran Associates, Inc.}},
  156. title = {Bayesian {{Alignments}} of {{Warped Multi}}-{{Output Gaussian Processes}}},
  157. url = {http://papers.nips.cc/paper/7931-bayesian-alignments-of-warped-multi-output-gaussian-processes.pdf},
  158. urldate = {2019-01-23}
  159. }
  160. @inproceedings{kingma_variational_2015,
  161. author = {Kingma, Diederik P and Salimans, Tim and Welling, Max},
  162. booktitle = {Advances in {{Neural Information Processing Systems}} 28},
  163. date = {2015},
  164. editor = {Cortes, C. and Lawrence, N. D. and Lee, D. D. and Sugiyama, M. and Garnett, R.},
  165. pages = {2575--2583},
  166. publisher = {{Curran Associates, Inc.}},
  167. title = {Variational {{Dropout}} and the {{Local Reparameterization Trick}}},
  168. url = {http://papers.nips.cc/paper/5666-variational-dropout-and-the-local-reparameterization-trick.pdf},
  169. urldate = {2018-09-12}
  170. }
  171. @article{lazaro-gredilla_overlapping_2012,
  172. author = {Lázaro-Gredilla, Miguel and Van Vaerenbergh, Steven and Lawrence, Neil D.},
  173. date = {2012},
  174. journaltitle = {Pattern Recognition},
  175. keywords = {Computer Science - Artificial Intelligence,Computer Science - Machine Learning,Statistics - Machine Learning},
  176. number = {4},
  177. pages = {1386--1395},
  178. title = {Overlapping Mixtures of {{Gaussian}} Processes for the Data Association Problem},
  179. volume = {45}
  180. }
  181. @article{maddison_concrete_2016,
  182. abstract = {The reparameterization trick enables optimizing large scale stochastic computation graphs via gradient descent. The essence of the trick is to refactor each stochastic node into a differentiable function of its parameters and a random variable with fixed distribution. After refactoring, the gradients of the loss propagated by the chain rule through the graph are low variance unbiased estimators of the gradients of the expected loss. While many continuous random variables have such reparameterizations, discrete random variables lack useful reparameterizations due to the discontinuous nature of discrete states. In this work we introduce Concrete random variables---continuous relaxations of discrete random variables. The Concrete distribution is a new family of distributions with closed form densities and a simple reparameterization. Whenever a discrete stochastic node of a computation graph can be refactored into a one-hot bit representation that is treated continuously, Concrete stochastic nodes can be used with automatic differentiation to produce low-variance biased gradients of objectives (including objectives that depend on the log-probability of latent stochastic nodes) on the corresponding discrete graph. We demonstrate the effectiveness of Concrete relaxations on density estimation and structured prediction tasks using neural networks.},
  183. archivePrefix = {arXiv},
  184. author = {Maddison, Chris J. and Mnih, Andriy and Teh, Yee Whye},
  185. date = {2016-11-02},
  186. eprint = {1611.00712},
  187. eprinttype = {arxiv},
  188. keywords = {Computer Science - Machine Learning,Statistics - Machine Learning},
  189. primaryClass = {cs, stat},
  190. shorttitle = {The {{Concrete Distribution}}},
  191. title = {The {{Concrete Distribution}}: {{A Continuous Relaxation}} of {{Discrete Random Variables}}},
  192. url = {http://arxiv.org/abs/1611.00712},
  193. urldate = {2018-09-12}
  194. }
  195. @article{matthews_gpflow_2017,
  196. author = {Matthews, Alexander G. de G. and van der Wilk, Mark and Nickson, Tom and Fujii, Keisuke and Boukouvalas, Alexis and León-Villagrá, Pablo and Ghahramani, Zoubin and Hensman, James},
  197. date = {2017},
  198. journaltitle = {Journal of Machine Learning Research},
  199. number = {40},
  200. options = {useprefix=true},
  201. pages = {1--6},
  202. shorttitle = {{{GPflow}}},
  203. title = {{{GPflow}}: {{A Gaussian}} Process Library Using {{TensorFlow}}},
  204. url = {http://www.jmlr.org/papers/volume18/16-537/16-537.pdf},
  205. urldate = {2017-09-27},
  206. volume = {18}
  207. }
  208. @inproceedings{rasmussen_infinite_2002,
  209. author = {Rasmussen, Carl E. and Ghahramani, Zoubin},
  210. booktitle = {Advances in {{Neural Information Processing Systems}} 14},
  211. date = {2002},
  212. editor = {Dietterich, T. G. and Becker, S. and Ghahramani, Z.},
  213. pages = {881--888},
  214. publisher = {{MIT Press}},
  215. title = {Infinite {{Mixtures}} of {{Gaussian Process Experts}}},
  216. url = {http://papers.nips.cc/paper/2055-infinite-mixtures-of-gaussian-process-experts.pdf},
  217. urldate = {2018-08-23}
  218. }
  219. @article{rezende_stochastic_2014,
  220. author = {Rezende, Danilo Jimenez and Mohamed, Shakir and Wierstra, Daan},
  221. date = {2014-01-16},
  222. langid = {english},
  223. title = {Stochastic {{Backpropagation}} and {{Approximate Inference}} in {{Deep Generative Models}}},
  224. url = {https://arxiv.org/abs/1401.4082},
  225. urldate = {2018-09-12}
  226. }
  227. @book{rousseeuw_robust_2005,
  228. author = {Rousseeuw, Peter J. and Leroy, Annick M.},
  229. date = {2005},
  230. publisher = {{John wiley \& sons}},
  231. title = {Robust Regression and Outlier Detection},
  232. volume = {589}
  233. }
  234. @inproceedings{salimbeni_doubly_2017,
  235. author = {Salimbeni, Hugh and Deisenroth, Marc},
  236. booktitle = {Advances in {{Neural Information Processing Systems}} 30},
  237. date = {2017},
  238. editor = {Guyon, I. and Luxburg, U. V. and Bengio, S. and Wallach, H. and Fergus, R. and Vishwanathan, S. and Garnett, R.},
  239. keywords = {Statistics - Machine Learning},
  240. pages = {4588--4599},
  241. publisher = {{Curran Associates, Inc.}},
  242. title = {Doubly {{Stochastic Variational Inference}} for {{Deep Gaussian Processes}}},
  243. url = {http://papers.nips.cc/paper/7045-doubly-stochastic-variational-inference-for-deep-gaussian-processes.pdf},
  244. urldate = {2018-10-02}
  245. }
  246. @article{tensorflow2015-whitepaper,
  247. author = {Abadi, Martín and Agarwal, Ashish and Barham, Paul and Brevdo, Eugene and Chen, Zhifeng and Citro, Craig and Corrado, Greg S. and Davis, Andy and Dean, Jeffrey and Devin, Matthieu and Ghemawat, Sanjay and Goodfellow, Ian and Harp, Andrew and Irving, Geoffrey and Isard, Michael and Jia, Yangqing and Jozefowicz, Rafal and Kaiser, Lukasz and Kudlur, Manjunath and Levenberg, Josh and Mané, Dandelion and Monga, Rajat and Moore, Sherry and Murray, Derek and Olah, Chris and Schuster, Mike and Shlens, Jonathon and Steiner, Benoit and Sutskever, Ilya and Talwar, Kunal and Tucker, Paul and Vanhoucke, Vincent and Vasudevan, Vijay and Viégas, Fernanda and Vinyals, Oriol and Warden, Pete and Wattenberg, Martin and Wicke, Martin and Yu, Yuan and Zheng, Xiaoqiang},
  248. date = {2015},
  249. note = {Software available from tensorflow.org},
  250. title = {{{TensorFlow}}: {{Large}}-{{Scale Machine Learning}} on {{Heterogeneous Systems}}},
  251. url = {https://www.tensorflow.org/}
  252. }
  253. @inproceedings{titsias_variational_2009,
  254. author = {Titsias, Michalis K.},
  255. booktitle = {{{AISTATS}}},
  256. date = {2009},
  257. pages = {567--574},
  258. title = {Variational {{Learning}} of {{Inducing Variables}} in {{Sparse Gaussian Processes}}.},
  259. url = {http://www.jmlr.org/proceedings/papers/v5/titsias09a/titsias09a.pdf},
  260. urldate = {2017-04-06},
  261. volume = {5}
  262. }
  263. @inproceedings{tresp_mixtures_2001,
  264. author = {Tresp, Volker},
  265. booktitle = {Advances in {{Neural Information Processing Systems}} 13},
  266. date = {2001},
  267. editor = {Leen, T. K. and Dietterich, T. G. and Tresp, V.},
  268. pages = {654--660},
  269. publisher = {{MIT Press}},
  270. title = {Mixtures of {{Gaussian Processes}}},
  271. url = {http://papers.nips.cc/paper/1900-mixtures-of-gaussian-processes.pdf},
  272. urldate = {2018-09-26}
  273. }