Machine Learning

A non-monotonic method for large-scale nonnegative least squares

by Suvrit Sra

Brenning, A. & B. Lausen (2008): Estimating error rates in the classification of paired organs. Statistics in Medicine, 27(22): 4515-4531.

by Alexander Brenning

Clinical data from paired organs present a dependence structure that has to be considered when making statistical... more Clinical data from paired organs present a dependence structure that has to be considered when making statistical inference or evaluating classification rules with resampling-based techniques (bootstrap, cross-validation). We introduce a paired cross-validation approach for the estimation of misclassification error rates in the classification of data from paired organs. The dependence structure of the sample is honored by subject-level cross-validation. Theoretical considerations as well as a case–control study on glaucoma diagnosis and a simulation study show that the variance of the paired cross-validation estimator is considerably lower than in traditional cross-validation error estimation on one randomly selected eye per subject. The actual variance reduction is mainly controlled by the contribution of differential misclassification between both eyes to the overall error rate. By contrast, ‘ad hoc’ cross-validation ignoring the autocorrelation of paired organs leads to biased error estimates. Using the double-bagging technique, we also show that classification accuracy can be improved by using information from both eyes in training machine-learning classifiers. In glaucoma detection, the reduction in misclassification error rates by training data from both eyes is equivalent to an increase in the sample size by one-third to one-half, which is an important achievement in clinical studies.

Adler, W., A. Brenning, S. Potapov, M. Schmid & B. Lausen (2011): Ensemble classification of paired data. Computational Statistics & Data Analysis, 55(5): 1933-1941

by Alexander Brenning

In many medical applications, data are taken from paired organs or from repeated measurements of the same organ or... more In many medical applications, data are taken from paired organs or from repeated measurements of the same organ or subject. Subject based as opposed to observation based evaluation of these data results in increased efficiency of the estimation of the misclassification rate. A subject based approach for classification in the generation of bootstrap samples of bagging and bundling methods is analyzed. A simulation model is used to compare the performance of different strategies to create the bootstrap samples which are used to grow individual trees. The proposed approach is compared to linear discriminant analysis, logistic regression, random forests and gradient boosting. Finally, the simulation results are applied to glaucoma diagnosis using both eyes of glaucoma patients and healthy controls. It is demonstrated that the proposed subject based resampling reduces the misclassification rate.

EEG estimates of engagement and cognitive workload predict math problem solving outcomes

by Federico Cirett Galán

Multiple dynamic representations in the motor cortex during sensorimotor learning

by Diego Gutnisky

Tell me more? The effects of mental model soundness on personalizing an intelligent agent.

by Todd Kulesza

In the proceedings of CHI 2012. Honorable mention for Best Paper award.

What does a user need to know to productively work with an intelligent agent? Intelligent agents and recommender... more What does a user need to know to productively work with an intelligent agent? Intelligent agents and recommender systems are gaining widespread use, potentially creating a need for end users to understand how these systems operate in order to fix their agent's personalized behavior. This paper explores the effects of mental model soundness on such personalization by providing structural knowledge of a music recommender system in an empirical study. Our findings show that participants were able to quickly build sound mental models of the recommender system's reasoning, and that participants who most improved their mental models during the study were significantly more likely to make the recommender operate to their satisfaction. These results suggest that by helping end users understand a system's reasoning, intelligent agents may elicit more and better feedback, thus more closely aligning their output with each user's intentions.

Towards recognizing “cool”: Can end users help computer vision recognize subjective attributes of objects in images?

by Todd Kulesza

Published in the proceedings of IUI 2012.

Recent computer vision approaches are aimed at richer image interpretations that extend the standard recognition of... more Recent computer vision approaches are aimed at richer image interpretations that extend the standard recognition of objects in images (e.g., cars) to also recognize object attributes (e.g., cylindrical, has-stripes, wet). However, the more idiosyncratic and abstract the notion of an object attribute (e.g., "cool" car), the more challenging the task of attribute recognition. This paper considers whether end users can help vision algorithms recognize highly idiosyncratic attributes, referred to here as subjective attributes. We empirically investigated how end users recognized three subjective attributes of cars—"cool", "cute", and "classic". Our results suggest the feasibility of vision algorithms recognizing subjective attributes of objects, but an interactive approach beyond standard supervised learning from labeled training examples is needed.

Why-Oriented End-User Debugging of Naive Bayes Text Classification

by Todd Kulesza

Published in ACM Transactions on Interactive Intelligent Systems, Vol. 1, No. 1, October 2011.

Machine learning techniques are increasingly used in intelligent assistants, that is, software targeted at and... more Machine learning techniques are increasingly used in intelligent assistants, that is, software targeted at and continuously adapting to assist end users with email, shopping, and other tasks. Examples include desktop SPAM filters, recommender systems, and handwriting recognition. Fixing such intelligent assistants when they learn incorrect behavior, however, has received only limited attention. To directly support end-user “debugging” of assistant behaviors learned via statistical machine learning, we present a Why-oriented approach which allows users to ask questions about how the assistant made its predictions, provides answers to these “why” questions, and allows users to interactively change these answers to debug the assistant’s current and future predictions. To understand the strengths and weaknesses of this approach, we then conducted an exploratory study to investigate barriers that participants could encounter when debugging an intelligent assistant using our approach, and the information those participants requested to overcome these barriers. To help ensure the inclusiveness of our approach, we also explored how gender differences played a role in understanding barriers and information needs. We then used these results to consider opportunities for Why-oriented approaches to address user barriers and information needs.

Classification of dog barks: a machine learning approach

by Csaba Molnar

In this study we analyzed the possible contextspecific and individual-specific features of dog barks using a new... more In this study we analyzed the possible contextspecific and individual-specific features of dog barks using a new machine-learning algorithm. A pool containing more than 6,000 barks, which were recorded in six different communicative situations was used as the sound sample. The algorithm's task was to learn which acoustic features of the barks, which were recorded in different contexts and from different individuals, could be distinguished from another. The program conducted this task by analyzing barks emitted in previously identified contexts by identified dogs. After the best feature set had been obtained (with which the highest identification rate was achieved), the efficiency of the algorithm was tested in a classification task in which unknown barks were analyzed. The recognition rates we found were highly above chance level: the algorithm could categorize the barks according to their recorded situation with an efficiency of 43% and with an efficiency of 52% of the barking individuals. These findings suggest that dog barks have context-specific and individual-specific acoustic features. In our opinion, this machine learning method may provide an efficient tool for analyzing acoustic data in various behavioral studies.

Estimating motion with principal component regression strategies

by Vania V. Estrela

Audio based Context Awareness on a Pocket PC

by Filip Bonnevier

Evolving Regular Expressions for GeneChip Probe Performance Prediction

by Andrew Harrison

Creating regular expressions as mRNA motifs with GP to predict human exon splitting

by Andrew Harrison

Automated DNA Motif Discovery

by Andrew Harrison

Soft Cardinality + ML: Learning Adaptive Similarity Functions for Cross-lingual Textual Entailment (preprint)

by Sergio Jimenez

to present in SemEval-2012, *SEM, First Joint Conference on Lexical and Computational Semantics, co-located with NAACL-HTL, Montreal, Canada

This paper presents a novel approach for building adaptive similarity functions based on cardinality using machine... more This paper presents a novel approach for building adaptive similarity functions based on cardinality using machine learning. Unlike current approaches that build feature sets using similarity scores, we have developed these feature sets with the cardinalities of the commonalities and differences between pairs of objects being compared. This approach allows the machine-learning algorithm to obtain an asymmetric similarity function suitable for directional judgments. Besides using the classic set cardinality, we used soft cardinality to allow flexibility in the comparison between words. Our approach used only the information from the surface of the text, a stop-word remover and a stemmer to address the cross-lingual textual entailment task 8 at SEMEVAL 2012. We have the third best result among the 29 systems submitted by 10 teams. Additionally, this paper presents better results compared with the best official score.

Concurrent Activity Recognition for Clinical Work

by Afsaneh Doryab

In 2012 IEEE World Congress in Computational Intelligence

We present an approach to learning to recognize concurrent activities based on multiple data streams. One example is... more We present an approach to learning to recognize concurrent activities based on multiple data streams. One example is recognition of concurrent activities in hospital operating rooms based on multiple wearable and embedded sensors. This problem differs from standard time series classification in that there is no natural single target dimension, as multiple activities are performed at the same time. Hence, most existing approaches fail. The key innovations that allow us to tackle this problem is (1) learning to recognize base activities from raw sensor data, (2) creating artificial joint activities from base activities using frequent pattern mining and (3) handling temporal dependency using virtual evidence boosting.

Prediction as a candidate for learning deep hierarchical models of data

by Rasmus Berg Palm

Recent findings [HOT06] have made possible the learning of deep layered hier- archical representations of data... more Recent findings [HOT06] have made possible the learning of deep layered hier- archical representations of data mimicking the brains working. It is hoped that this paradigm will unlock some of the power of the brain and lead to advances towards true AI.
In this thesis I implement and evaluate state-of-the-art deep learning models and using these as building blocks I investigate the hypothesis that predicting the time-to-time sensory input is a good learning objective. I introduce the Predictive Encoder (PE) and show that a simple non-regularized learning rule, minimizing prediction error on natural video patches leads to receptive fields similar to those found in Macaque monkey visual area V1. I scale this model to video of natural scenes by introducing the Convolutional Predictive Encoder (CPE) and show similar results. Both models can be used in deep architectures as a deep learning module.

Jensen-Bregman LogDet Divergence for Efficient Similarity Computations on Positive Defnite Tensors

by Anoop Cherian

Covariance matrices provide an easy platform for fusing multiple features compactly and as a result have found immense... more Covariance matrices provide an easy platform for fusing multiple features compactly and as a result have found immense success in several computer vision applications, including activity recognition, visual surveillance, and diffusion tensor
imaging. An important task in all of these applications is to compute the distance
between covariance matrices using a (dis)similarity function, for which the natural
choice is the Riemannian metric corresponding to the manifold inhabited by these
matrices. As this Riemannian manifold is not flat, the dissimilarities should take
into account the curvature of the manifold. As a result such distance computations
tend to slow down, especially when the matrix dimensions are large or gradients
are required. Further, suitability of the metric to enable efficient nearest neighbor
retrieval is an important requirement in the contemporary times of big data analytics. To alleviate these difficulties, this paper proposes a novel dissimilarity measure
for covariances, the Jensen-Bregman LogDet Divergence (JBLD). This divergence
enjoys several desirable theoretical properties, at the same time is computationally less demanding (compared to standard measures). To address the problem
of efficient nearest neighbor retrieval on large covariance datasets, we propose a
metric tree framework using kmeans clustering on JBLD. We demonstrate the superior performance of JBLD on covariance datasets from several computer vision
applications.

Every Moment Is a Learning Time”: Conversation with Michel Alhadeff-Jones

by Michel Alhadeff-Jones

Hodeck, M. (2008). “Every Moment Is a Learning Time”: Conversation with Michel Alhadeff-Jones, Teacher Writers for a Public Voice - “Inter–View” Bulletin, 3, n°3-4, 3-7.

In December 2008, Maria Hodeck, editor of the Teachers Writer for a Public Voice “Inter-view Bulletin” (Teachers... more In December 2008, Maria Hodeck, editor of the Teachers Writer for a Public Voice “Inter-view Bulletin” (Teachers College, Columbia University), invited me to have an electronic conversation with her about the new course I am proposing this Spring 2009 on “Time and Learning“.

This text is the the transcript of our conversation. Through autobiographical elements, this conversation introduces several themes such as: time and alienation; the multiplicity of temporalities; time and teaching; informal learning; paces, movements, and the rhythms of learning; time and agency; rhythms and health.

http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0035860

by jessica schrouff

published in PLoS One, 04/26/2012

Predicting a particular cognitive state from a specific pattern of fMRI voxel values is still a methodological... more Predicting a particular cognitive state from a specific pattern of fMRI voxel values is still a methodological challenge. Decoding brain activity is usually performed in highly controlled experimental paradigms characterized by a series of distinct states induced by a temporally constrained experimental design. In more realistic conditions, the number, sequence and duration of mental states are unpredictably generated by the individual, resulting in complex and imbalanced fMRI data sets. This study tests the classification of brain activity, acquired on 16 volunteers using fMRI, during mental imagery, a condition in which the number and duration of mental events were not externally imposed but self-generated. To deal with these issues, two classification techniques were considered (Support Vector Machines, SVM, and Gaussian Processes, GP), as well as different feature extraction methods (General Linear Model, GLM and SVM). These techniques were combined in order to identify the procedures leading to the highest accuracy measures. Our results showed that 12 data sets out of 16 could be significantly modeled by either SVM or GP. Model accuracies tended to be related to the degree of imbalance between classes and to task performance of the volunteers. We also conclude that the GP technique tends to be more robust than SVM to model unbalanced data sets.