Machine Learning is the study of methods for programming computers to learn. Computers are applied to a wide range of tasks, and for most of these it is relatively easy for programmers to design and implement the necessary software. However, there are many tasks for which this is difficult or impossible. These can be divided into four general categories. First, there are problems for which there exist no human experts. For example, in modern automated manufacturing facilities, there is a need to predict machine failures before they occur by analyzing sensor readings. Because the machines are new, there are no human experts who can be interviewed by a programmer to provide the knowledge necessary to build a computer system. A machine learning system can study recorded data and subsequent machine failures and learn prediction rules. Second, there are problems where human experts exist, but where they are unable to explain their expertise. This is the case in many perceptual tasks, such as speech recognition, hand-writing recognition, and natural language understanding. Virtually all humans exhibit expert-level abilities on these tasks, but none of them can describe the detailed steps that they follow as they perform them. Fortunately, humans can provide machines with examples of the inputs and correct outputs for these tasks, so machine learning algorithms can learn to map the inputs to the outputs. Third, there are problems where phenomena are changing rapidly. In finance, for example, people would like to predict the future behavior of the stock market, of consumer purchases, or of exchange rates. These behaviors change frequently, so that even if a programmer could construct a good predictive computer program, it would need to be rewritten frequently. A learning program can relieve the programmer of this burden by constantly modifying and tuning a set of learned prediction rules. Fourth, there are applications that need to be customized for each computer user separately. Consider, for example, a program to filter unwanted electronic mail messages. Different users will need different filters. It is unreasonable to expect each user to program his or her own rules, and it is infeasible to provide every user with a software engineer to keep the rules up-to-date. A machine learning system can learn which mail messages the user rejects and maintain the filtering rules automatically. Machine learning addresses many of the same research questions as the fields of statistics, data mining, and psychology, but with differences of emphasis. Statistics focuses on understanding the phenomena that have generated the data, often with the goal of testing different hypotheses about those phenomena. Data mining seeks to find patterns in the data that are understandable by people. Psychological studies of human learning aspire to understand the mechanisms underlying the various learning behaviors exhibited by people (concept learning, skill acquisition, strategy change, etc.).

Machine learning

/pdf/convex-analysisnoer-san-nojin-zhan-nituite-5dl2rbmoyr.pdf

Convex Analysisの二,三の進展について

“Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks”の学習報告

Statistical Analysis with Missing Data

An overview of the self-organizing map algorithm, on which the papers in this issue are based, is presented in this article.

The Self-Organizing Map

We study how different output layer types of a deep neural network learn and forget in continual learning settings. We describe the three factors affecting catastrophic forgetting in the output layer: (1) weights modifications, (2) interferences, and (3) projection drift. Our goal is to provide more insights into how different types of output layers can address (1) and (2). We also propose potential solutions and evaluate them on several benchmarks. We show that the best-performing output layer type depends on the data distribution drifts or the amount of data available. In particular, in some cases where a standard linear layer would fail, it is sufficient to change the parametrization and get significantly better performance while still training with SGD. Our results and analysis shed light on the dynamics of the output layer in continual learning scenarios and help select the best-suited output layer for a given scenario.

Continual Learning in Deep Networks: an Analysis of the Last Layer

ApproximationAlgorithmsforProbabilisticDeco dingIrinaRish,KalevKaskandRinaDechterDepartmentofInformationandComputerScienceUniversityofCalifornia,Irvinefirinar,kkask,dechterg@ics .uc i.eduAbstractIt was recently shown that the problem of deco ding messages transmitted throughanoisychannelcanb eformulatedasb eliefup datingtaskover aprobabilisticnet-work[13].Moreover,itasobservedthatiterativeapplicationofthe(lineartime)b eliefpropagation algorithm designedfor p olytrees [14] outp erformed state of the artdeco dingalgorithms, even though the corresp ondingnetworks may have many cycles.Much of recent research in co ding fo cuses on explainingthis phenomenon.This pap er demonstrates empiricallythat an approximation algorithmapprox-mpeforsolvingthemostprobableexplanation(MPE)problem,develop edwithinre-centlyprop osedmini-bucketeliminationframework[3],outp erformsiterativeb eliefpropagationonclassesofco dingnetworksthathaveb oundedinducdwidth.Ourexp erimentssuggestthatdeco dersbasedon ndinganMPEassignment(blo ck-wisedeco ding)canb esup eriortothecommonlyusedb eliefup datingdeco ders(bit-wisedeco ders).1Intro ductionInthispap erweevaluatethequalityofarecently prop osedapproximationscheme, calledmini-bucket, forprobabilistic deco ding.Recently, a class of parameterized approximation algorithms based on the bucket elimina-tion framework was prop osed and analyzed [3].The approximation scheme uses as a control-ling parameter a b ound on size of probabilistic functions created during variable elimination,allowingatrade-o b etweenaccuracyandeciency [7,5].Thealgorithmserepresentedand analyzed for several tasks, such as: nding the most probable explanation (mpe), ndingthe maximum a p osteriori hyp othesis (map), and b elief up dating.Encouraging results wereThisworkaspartiallysupp ortedbyNSFgrantIRI-9157636andAirForceOceofScienti cResearch grant, AFOSR 900136, Ro ckwell International and Amada of America.1

Approximation Algorithms for Probabilistic Decoding

Sparse models embed variable selection into model learning (e.g., by using l1-norm regularizer). In small-sample high-dimensional problems, this leads to improved generalization accuracy combined with interpretability, which is important in scientific applications such as biology. In this paper, we summarize our recent work on sparse models, including both sparse regression and sparse Gaussian Markov Random Fields (GMRF), in neuroimaging applications, such as functional MRI data analysis, where the central objective is to gain a better insight into brain functioning, besides just learning predictive models of mental states from imaging data.

http://ecmlpkdd2013.org/wp-content/uploads/2013/07/679.pdf

Functional MRI analysis with sparse models

We propose a novel machine-learning framework for dialogue modeling which uses representations based on hash functions. More specifically, each person's response is represented by a binary hashcode where each bit reflects presence or absence of a certain text pattern in the response. Hashcodes serve as compressed text representations, allowing for efficient similarity search. Moreover, hashcode of one person's response can be used as a feature vector for predicting the hashcode representing another person's response. The proposed hashing model of dialogue is obtained by maximizing a novel lower bound on the mutual information between the hashcodes of consecutive responses. We apply our approach in psychotherapy domain, evaluating its effectiveness on a real-life dataset consisting of therapy sessions with patients suffering from depression.

Dialogue Modeling Via Hash Functions: Applications to Psychotherapy

We consider an extension of the contextual bandit setting, motivated by several practical applications, where an unlabeled history of contexts can become available for pre-training before the online decision-making begins. We propose an approach for improving the performance of contextual bandit in such setting, via adaptive, dynamic representation learning, which combines offline pre-training on unlabeled history of contexts with online selection and modification of embedding functions. Our experiments on a variety of datasets and in different nonstationary environments demonstrate clear advantages of our approach over the standard contextual bandit.

Irina Rish

Papers

Continual Learning in Deep Networks: an Analysis of the Last Layer

Approximation Algorithms for Probabilistic Decoding

Functional MRI analysis with sparse models

Dialogue Modeling Via Hash Functions: Applications to Psychotherapy

Adaptive Representation Selection in Contextual Bandit with Unlabeled History.