Darrin P. Lewis, T. Jebara, William Stafford Noble
The power and popularity of kernel methods stem in part from their ability to handle diverse forms of structured inputs, including vectors, graphs and strings. Recently, several methods have been proposed for combining kernels from heterogeneous data sources. However, all of these methods produce stationary combinations; i.e., the relative weights of the various kernels do not vary among input examples. This article proposes a method for combining multiple kernels in a nonstationary fashion. The approach uses a large-margin latent-variable generative model within the maximum entropy discrimination (MED) framework. Latent parameter estimation is rendered tractable by variational bounds and an iterative optimization procedure. The classifier we use is a log-ratio of Gaussian mixtures, in which each component is implicitly mapped via a Mercer kernel function. We show that the support vector machine is a special case of this model. In this approach, discriminative parameter estimation is feasible via a fast sequential minimal optimization algorithm. Empirical results are presented on synthetic data, several benchmarks, and on a protein function annotation task.
{"title":"Nonstationary kernel combination","authors":"Darrin P. Lewis, T. Jebara, William Stafford Noble","doi":"10.1145/1143844.1143914","DOIUrl":"https://doi.org/10.1145/1143844.1143914","url":null,"abstract":"The power and popularity of kernel methods stem in part from their ability to handle diverse forms of structured inputs, including vectors, graphs and strings. Recently, several methods have been proposed for combining kernels from heterogeneous data sources. However, all of these methods produce stationary combinations; i.e., the relative weights of the various kernels do not vary among input examples. This article proposes a method for combining multiple kernels in a nonstationary fashion. The approach uses a large-margin latent-variable generative model within the maximum entropy discrimination (MED) framework. Latent parameter estimation is rendered tractable by variational bounds and an iterative optimization procedure. The classifier we use is a log-ratio of Gaussian mixtures, in which each component is implicitly mapped via a Mercer kernel function. We show that the support vector machine is a special case of this model. In this approach, discriminative parameter estimation is feasible via a fast sequential minimal optimization algorithm. Empirical results are presented on synthetic data, several benchmarks, and on a protein function annotation task.","PeriodicalId":124011,"journal":{"name":"Proceedings of the 23rd international conference on Machine learning","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134365092","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
A. Zheng, Michael I. Jordan, B. Liblit, M. Naik, A. Aiken
We describe a statistical approach to software debugging in the presence of multiple bugs. Due to sparse sampling issues and complex interaction between program predicates, many generic off-the-shelf algorithms fail to select useful bug predictors. Taking inspiration from bi-clustering algorithms, we propose an iterative collective voting scheme for the program runs and predicates. We demonstrate successful debugging results on several real world programs and a large debugging benchmark suite.
{"title":"Statistical debugging: simultaneous identification of multiple bugs","authors":"A. Zheng, Michael I. Jordan, B. Liblit, M. Naik, A. Aiken","doi":"10.1145/1143844.1143983","DOIUrl":"https://doi.org/10.1145/1143844.1143983","url":null,"abstract":"We describe a statistical approach to software debugging in the presence of multiple bugs. Due to sparse sampling issues and complex interaction between program predicates, many generic off-the-shelf algorithms fail to select useful bug predictors. Taking inspiration from bi-clustering algorithms, we propose an iterative collective voting scheme for the program runs and predicates. We demonstrate successful debugging results on several real world programs and a large debugging benchmark suite.","PeriodicalId":124011,"journal":{"name":"Proceedings of the 23rd international conference on Machine learning","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128344488","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Feature selection is often applied to high-dimensional data prior to classification learning. Using the same training dataset in both selection and learning can result in so-called feature subset selection bias. This bias putatively can exacerbate data over-fitting and negatively affect classification performance. However, in current practice separate datasets are seldom employed for selection and learning, because dividing the training data into two datasets for feature selection and classifier learning respectively reduces the amount of data that can be used in either task. This work attempts to address this dilemma. We formalize selection bias for classification learning, analyze its statistical properties, and study factors that affect selection bias, as well as how the bias impacts classification learning via various experiments. This research endeavors to provide illustration and explanation why the bias may not cause negative impact in classification as much as expected in regression.
{"title":"Feature subset selection bias for classification learning","authors":"Surendra K. Singhi, Huan Liu","doi":"10.1145/1143844.1143951","DOIUrl":"https://doi.org/10.1145/1143844.1143951","url":null,"abstract":"Feature selection is often applied to high-dimensional data prior to classification learning. Using the same training dataset in both selection and learning can result in so-called feature subset selection bias. This bias putatively can exacerbate data over-fitting and negatively affect classification performance. However, in current practice separate datasets are seldom employed for selection and learning, because dividing the training data into two datasets for feature selection and classifier learning respectively reduces the amount of data that can be used in either task. This work attempts to address this dilemma. We formalize selection bias for classification learning, analyze its statistical properties, and study factors that affect selection bias, as well as how the bias impacts classification learning via various experiments. This research endeavors to provide illustration and explanation why the bias may not cause negative impact in classification as much as expected in regression.","PeriodicalId":124011,"journal":{"name":"Proceedings of the 23rd international conference on Machine learning","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130285109","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Recently, a very appealing approach was proposed to compute the entire solution path for support vector classification (SVC) with very low extra computational cost. This approach was later extended to a support vector regression (SVR) model called ε-SVR. However, the method requires that the error parameter ε be set a priori, which is only possible if the desired accuracy of the approximation can be specified in advance. In this paper, we show that the solution path for ε-SVR is also piecewise linear with respect to ε. We further propose an efficient algorithm for exploring the two-dimensional solution space defined by the regularization and error parameters. As opposed to the algorithm for SVC, our proposed algorithm for ε-SVR initializes the number of support vectors to zero and then increases it gradually as the algorithm proceeds. As such, a good regression function possessing the sparseness property can be obtained after only a few iterations.
{"title":"Two-dimensional solution path for support vector regression","authors":"G. Wang, D. Yeung, F. Lochovsky","doi":"10.1145/1143844.1143969","DOIUrl":"https://doi.org/10.1145/1143844.1143969","url":null,"abstract":"Recently, a very appealing approach was proposed to compute the entire solution path for support vector classification (SVC) with very low extra computational cost. This approach was later extended to a support vector regression (SVR) model called ε-SVR. However, the method requires that the error parameter ε be set a priori, which is only possible if the desired accuracy of the approximation can be specified in advance. In this paper, we show that the solution path for ε-SVR is also piecewise linear with respect to ε. We further propose an efficient algorithm for exploring the two-dimensional solution space defined by the regularization and error parameters. As opposed to the algorithm for SVC, our proposed algorithm for ε-SVR initializes the number of support vectors to zero and then increases it gradually as the algorithm proceeds. As such, a good regression function possessing the sparseness property can be obtained after only a few iterations.","PeriodicalId":124011,"journal":{"name":"Proceedings of the 23rd international conference on Machine learning","volume":"87 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131450524","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
We introduce the use of learned shaping rewards in reinforcement learning tasks, where an agent uses prior experience on a sequence of tasks to learn a portable predictor that estimates intermediate rewards, resulting in accelerated learning in later tasks that are related but distinct. Such agents can be trained on a sequence of relatively easy tasks in order to develop a more informative measure of reward that can be transferred to improve performance on more difficult tasks without requiring a hand coded shaping function. We use a rod positioning task to show that this significantly improves performance even after a very brief training period.
{"title":"Autonomous shaping: knowledge transfer in reinforcement learning","authors":"G. Konidaris, A. Barto","doi":"10.1145/1143844.1143906","DOIUrl":"https://doi.org/10.1145/1143844.1143906","url":null,"abstract":"We introduce the use of learned shaping rewards in reinforcement learning tasks, where an agent uses prior experience on a sequence of tasks to learn a portable predictor that estimates intermediate rewards, resulting in accelerated learning in later tasks that are related but distinct. Such agents can be trained on a sequence of relatively easy tasks in order to develop a more informative measure of reward that can be transferred to improve performance on more difficult tasks without requiring a hand coded shaping function. We use a rod positioning task to show that this significantly improves performance even after a very brief training period.","PeriodicalId":124011,"journal":{"name":"Proceedings of the 23rd international conference on Machine learning","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129912713","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
We describe a family of embedding algorithms that are based on nonparametric estimates of mutual information (MI). Using Parzen window estimates of the distribution in the joint (input, embedding)-space, we derive a MI-based objective function for dimensionality reduction that can be optimized directly with respect to a set of latent data representatives. Various types of supervision signal can be introduced within the framework by replacing plain MI with several forms of conditional MI. Examples of the semi-(un)supervised algorithms that we obtain this way are a new model for manifold alignment, and a new type of embedding method that performs 'conditional dimensionality reduction'.
{"title":"Kernel information embeddings","authors":"R. Memisevic","doi":"10.1145/1143844.1143924","DOIUrl":"https://doi.org/10.1145/1143844.1143924","url":null,"abstract":"We describe a family of embedding algorithms that are based on nonparametric estimates of mutual information (MI). Using Parzen window estimates of the distribution in the joint (input, embedding)-space, we derive a MI-based objective function for dimensionality reduction that can be optimized directly with respect to a set of latent data representatives. Various types of supervision signal can be introduced within the framework by replacing plain MI with several forms of conditional MI. Examples of the semi-(un)supervised algorithms that we obtain this way are a new model for manifold alignment, and a new type of embedding method that performs 'conditional dimensionality reduction'.","PeriodicalId":124011,"journal":{"name":"Proceedings of the 23rd international conference on Machine learning","volume":"90 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123458007","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In Kernel Fisher discriminant analysis (KFDA), we carry out Fisher linear discriminant analysis in a high dimensional feature space defined implicitly by a kernel. The performance of KFDA depends on the choice of the kernel; in this paper, we consider the problem of finding the optimal kernel, over a given convex set of kernels. We show that this optimal kernel selection problem can be reformulated as a tractable convex optimization problem which interior-point methods can solve globally and efficiently. The kernel selection method is demonstrated with some UCI machine learning benchmark examples.
{"title":"Optimal kernel selection in Kernel Fisher discriminant analysis","authors":"Seung-Jean Kim, A. Magnani, Stephen P. Boyd","doi":"10.1145/1143844.1143903","DOIUrl":"https://doi.org/10.1145/1143844.1143903","url":null,"abstract":"In Kernel Fisher discriminant analysis (KFDA), we carry out Fisher linear discriminant analysis in a high dimensional feature space defined implicitly by a kernel. The performance of KFDA depends on the choice of the kernel; in this paper, we consider the problem of finding the optimal kernel, over a given convex set of kernels. We show that this optimal kernel selection problem can be reformulated as a tractable convex optimization problem which interior-point methods can solve globally and efficiently. The kernel selection method is demonstrated with some UCI machine learning benchmark examples.","PeriodicalId":124011,"journal":{"name":"Proceedings of the 23rd international conference on Machine learning","volume":"71 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128244721","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
We address the problem of efficiently learning Naive Bayes classifiers under class-conditional classification noise (CCCN). Naive Bayes classifiers rely on the hypothesis that the distributions associated to each class are product distributions. When data is subject to CCC-noise, these conditional distributions are themselves mixtures of product distributions. We give analytical formulas which makes it possible to identify them from data subject to CCCN. Then, we design a learning algorithm based on these formulas able to learn Naive Bayes classifiers under CCCN. We present results on artificial datasets and datasets extracted from the UCI repository database. These results show that CCCN can be efficiently and successfully handled.
{"title":"Efficient learning of Naive Bayes classifiers under class-conditional classification noise","authors":"François Denis, C. Magnan, L. Ralaivola","doi":"10.1145/1143844.1143878","DOIUrl":"https://doi.org/10.1145/1143844.1143878","url":null,"abstract":"We address the problem of efficiently learning Naive Bayes classifiers under class-conditional classification noise (CCCN). Naive Bayes classifiers rely on the hypothesis that the distributions associated to each class are product distributions. When data is subject to CCC-noise, these conditional distributions are themselves mixtures of product distributions. We give analytical formulas which makes it possible to identify them from data subject to CCCN. Then, we design a learning algorithm based on these formulas able to learn Naive Bayes classifiers under CCCN. We present results on artificial datasets and datasets extracted from the UCI repository database. These results show that CCCN can be efficiently and successfully handled.","PeriodicalId":124011,"journal":{"name":"Proceedings of the 23rd international conference on Machine learning","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128984295","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
When constructing a classifier from labeled data, it is important not to assign too much weight to any single input feature, in order to increase the robustness of the classifier. This is particularly important in domains with nonstationary feature distributions or with input sensor failures. A common approach to achieving such robustness is to introduce regularization which spreads the weight more evenly between the features. However, this strategy is very generic, and cannot induce robustness specifically tailored to the classification task at hand. In this work, we introduce a new algorithm for avoiding single feature over-weighting by analyzing robustness using a game theoretic formalization. We develop classifiers which are optimally resilient to deletion of features in a minimax sense, and show how to construct such classifiers using quadratic programming. We illustrate the applicability of our methods on spam filtering and handwritten digit recognition tasks, where feature deletion is indeed a realistic noise model.
{"title":"Nightmare at test time: robust learning by feature deletion","authors":"A. Globerson, S. Roweis","doi":"10.1145/1143844.1143889","DOIUrl":"https://doi.org/10.1145/1143844.1143889","url":null,"abstract":"When constructing a classifier from labeled data, it is important not to assign too much weight to any single input feature, in order to increase the robustness of the classifier. This is particularly important in domains with nonstationary feature distributions or with input sensor failures. A common approach to achieving such robustness is to introduce regularization which spreads the weight more evenly between the features. However, this strategy is very generic, and cannot induce robustness specifically tailored to the classification task at hand. In this work, we introduce a new algorithm for avoiding single feature over-weighting by analyzing robustness using a game theoretic formalization. We develop classifiers which are optimally resilient to deletion of features in a minimax sense, and show how to construct such classifiers using quadratic programming. We illustrate the applicability of our methods on spam filtering and handwritten digit recognition tasks, where feature deletion is indeed a realistic noise model.","PeriodicalId":124011,"journal":{"name":"Proceedings of the 23rd international conference on Machine learning","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128556446","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In the model-based policy search approach to reinforcement learning (RL), policies are found using a model (or "simulator") of the Markov decision process. However, for high-dimensional continuous-state tasks, it can be extremely difficult to build an accurate model, and thus often the algorithm returns a policy that works in simulation but not in real-life. The other extreme, model-free RL, tends to require infeasibly large numbers of real-life trials. In this paper, we present a hybrid algorithm that requires only an approximate model, and only a small number of real-life trials. The key idea is to successively "ground" the policy evaluations using real-life trials, but to rely on the approximate model to suggest local changes. Our theoretical results show that this algorithm achieves near-optimal performance in the real system, even when the model is only approximate. Empirical results also demonstrate that---when given only a crude model and a small number of real-life trials---our algorithm can obtain near-optimal performance in the real system.
{"title":"Using inaccurate models in reinforcement learning","authors":"P. Abbeel, M. Quigley, A. Ng","doi":"10.1145/1143844.1143845","DOIUrl":"https://doi.org/10.1145/1143844.1143845","url":null,"abstract":"In the model-based policy search approach to reinforcement learning (RL), policies are found using a model (or \"simulator\") of the Markov decision process. However, for high-dimensional continuous-state tasks, it can be extremely difficult to build an accurate model, and thus often the algorithm returns a policy that works in simulation but not in real-life. The other extreme, model-free RL, tends to require infeasibly large numbers of real-life trials. In this paper, we present a hybrid algorithm that requires only an approximate model, and only a small number of real-life trials. The key idea is to successively \"ground\" the policy evaluations using real-life trials, but to rely on the approximate model to suggest local changes. Our theoretical results show that this algorithm achieves near-optimal performance in the real system, even when the model is only approximate. Empirical results also demonstrate that---when given only a crude model and a small number of real-life trials---our algorithm can obtain near-optimal performance in the real system.","PeriodicalId":124011,"journal":{"name":"Proceedings of the 23rd international conference on Machine learning","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129335210","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：