@inproceedings{coates2011analysis,
  abstract = {A great deal of research has focused on algorithms for learning features from unlabeled data. Indeed, much progress has been made on benchmark datasets like NORB and CIFAR-10 by employing increasingly complex unsupervised learning algorithms and deep models. In this paper, however, we show that several simple factors, such as the number of hidden nodes in the model, may be more important to achieving high performance than the learning algorithm or the depth of the model. Specifically, we will apply several off-the-shelf feature learning algorithms (sparse auto-encoders, sparse RBMs, K-means clustering, and Gaussian mixtures) to CIFAR-10, NORB, and STL datasets using only single-layer networks. We then present a detailed analysis of the effect of changes in the model setup: the receptive field size, number of hidden nodes (features), the step-size ("stride") between extracted features, and the effect of whitening. Our results show that large numbers of hidden nodes and dense feature extraction are critical to achieving high performance - so critical, in fact, that when these parameters are pushed to their limits, we achieve state-of-the-art performance on both CIFAR-10 and NORB using only a single layer of features. More surprisingly, our best performance is based on K-means clustering, which is extremely fast, has no hyper-parameters to tune beyond the model structure itself, and is very easy to implement. Despite the simplicity of our system, we achieve accuracy beyond all previously published results on the CIFAR-10 and NORB datasets (79.6% and 97.2% respectively).},
  author = {Coates, A. and Lee, H. and Ng, A.Y.},
  booktitle = {Proceedings of the Fourteenth International Conference on Artificial Intelligence and Statistics},
  editor = {Gordon, Geoffrey and Dunson, David and Dudík, Miroslav},
  interhash = {46cfb4b5b1c16c79a966512e07f67158},
  intrahash = {bcb2c1fd335ae57362cdf348ff727589},
  pages = {215--223},
  publisher = {JMLR W\&CP},
  series = {JMLR Workshop and Conference Proceedings},
  title = {An analysis of single-layer networks in unsupervised feature learning},
  url = {http://jmlr.csail.mit.edu/proceedings/papers/v15/coates11a.html},
  volume = 15,
  year = 2011
}

@inproceedings{coates2011detection,
  abstract = {Reading text from photographs is a challenging problem that has received a significant amount of attention. Two key components of most systems are (i) text detection from images and (ii) character recognition, and many recent methods have been proposed to design better feature representations and models for both. In this paper, we apply methods recently developed in machine learning -- specifically, large-scale algorithms for learning the features automatically from unlabeled data -- and show that they allow us to construct highly effective classifiers for both detection and recognition to be used in a high accuracy end-to-end system.},
  author = {Coates, A. and Carpenter, B. and Case, C. and Satheesh, S. and Suresh, B. and Wang, Tao and Wu, D.J. and Ng, A.Y.},
  booktitle = {International Conference on Document Analysis and Recognition (ICDAR)},
  doi = {10.1109/ICDAR.2011.95},
  interhash = {adb17817e5f95605a8066737ce0e8b7e},
  intrahash = {b550ca5ec5a8b61b64b17091f7b2eeab},
  issn = {1520-5363},
  month = sep,
  pages = {440--445},
  title = {Text Detection and Character Recognition in Scene Images with Unsupervised Feature Learning},
  url = {http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=6065350&tag=1},
  year = 2011
}

@incollection{leake2000casebased,
  abstract = {Case-based reasoning(CBR) is an artificial intelligence paradigm for reasoning and learning. Case-based reasoning solves new problems by retrieving stored records of prior problem-solving episodes (cases) and adapting their solutions to fit new circumstances. Each processing episode provides a new case that is stored for future reuse, making learning a natural side-effect of the reasoning process. Case-based reasoning is also studied within cognitive science as a model of human reasoning: studies show that people use recollections of prior problems to guide their reasoning in a wide range of tasks, such as programming, mathematical problem solving, diagnosis, decision making, and design.},
  acmid = {1074199},
  address = {Chichester, UK},
  author = {Leake, David B.},
  booktitle = {Encyclopedia of Computer Science},
  edition = {4th},
  editor = {Ralston, Anthony and Reilly, Edwin D. and Hemmendinger, David},
  interhash = {fa414e2f48be14bb94cbfbf2566e36af},
  intrahash = {b8526b7c03f1fc9bdd85863dfbf881a2},
  isbn = {0-470-86412-5},
  month = jun,
  numpages = {2},
  pages = {196--197},
  publisher = {John Wiley and Sons Ltd.},
  title = {Case-based reasoning},
  url = {http://dl.acm.org/citation.cfm?id=1074100.1074199},
  year = 2000
}

@article{raykar2010learning,
  abstract = {For many supervised learning tasks it may be infeasible (or very expensive) to obtain objective and reliable labels. Instead, we can collect subjective (possibly noisy) labels from multiple experts or annotators. In practice, there is a substantial amount of disagreement among the annotators, and hence it is of great practical interest to address conventional supervised learning problems in this scenario. In this paper we describe a probabilistic approach for supervised learning when we have multiple annotators providing (possibly noisy) labels but no absolute gold standard. The proposed algorithm evaluates the different experts and also gives an estimate of the actual hidden labels. Experimental results indicate that the proposed method is superior to the commonly used majority voting baseline.},
  acmid = {1859894},
  author = {Raykar, Vikas C. and Yu, Shipeng and Zhao, Linda H. and Valadez, Gerardo Hermosillo and Florin, Charles and Bogoni, Luca and Moy, Linda},
  interhash = {8113daf47997fddf48e4c6c79f2eba56},
  intrahash = {14220abe8babfab01c0cdd5ebd5e4b7c},
  issn = {1532-4435},
  issue_date = {3/1/2010},
  journal = {Journal of Machine Learning Research},
  month = aug,
  numpages = {26},
  pages = {1297--1322},
  publisher = {JMLR.org},
  title = {Learning From Crowds},
  url = {http://dl.acm.org/citation.cfm?id=1756006.1859894},
  volume = 11,
  year = 2010
}

@inproceedings{davis2006relationship,
  abstract = {Receiver Operator Characteristic (ROC) curves are commonly used to present results for binary decision problems in machine learning. However, when dealing with highly skewed datasets, Precision-Recall (PR) curves give a more informative picture of an algorithm's performance. We show that a deep connection exists between ROC space and PR space, such that a curve dominates in ROC space if and only if it dominates in PR space. A corollary is the notion of an achievable PR curve, which has properties much like the convex hull in ROC space; we show an efficient algorithm for computing this curve. Finally, we also note differences in the two types of curves are significant for algorithm design. For example, in PR space it is incorrect to linearly interpolate between points. Furthermore, algorithms that optimize the area under the ROC curve are not guaranteed to optimize the area under the PR curve.},
  address = {New York, NY, USA},
  author = {Davis, Jesse and Goadrich, Mark},
  booktitle = {ICML '06: Proceedings of the 23rd international conference on Machine learning},
  doi = {http://doi.acm.org/10.1145/1143844.1143874},
  interhash = {e4ea92aea3ff8bbb3eb04c64867505f2},
  intrahash = {4cc51d680241bab2326e28dfea42c9ea},
  isbn = {1-59593-383-2},
  location = {Pittsburgh, Pennsylvania},
  pages = {233--240},
  publisher = {ACM},
  title = {The relationship between Precision-Recall and ROC curves},
  url = {http://portal.acm.org/citation.cfm?id=1143874},
  year = 2006
}