Search

In many large-scale classification problems, classes are organized in a known hierarchy, typically represented as a tree expressing the inclusion of classes in superclasses. We introduce a loss for this type of supervised hierarchical classification. It utilizes the knowledge of the hierarchy to assign each example not only to a class but also to all encompassing superclasses. Applicable to any feedforward architecture with a softmax output layer, this loss is a proper scoring rule, in that its expectation is minimized by the true posterior class probabilities. This property allows us to simultaneously pursue consistent classification objectives between superclasses and fine-grained classes, and eliminates the need for a performance trade-off between different granularities. We conduct an experimental study on three reference benchmarks, in which we vary the size of the training sets to cover a diverse set of learning scenarios. Our approach does not entail any significant additional computational cost compared with the loss of cross-entropy. It improves accuracy and reduces the number of coarse errors, with predicted labels that are distant from ground-truth labels in the tree.

The field of industrial defect detection using machine learning and deep learning is a subject of active research. Datasets, also called benchmarks, are used to compare and assess research results. There is a number of datasets in industrial visual inspection, of varying quality. Thus, it is a difficult task to determine which dataset to use. Generally speaking, datasets which include a testing set, with precise labeling and made in real-world conditions should be preferred. We propose a study of existing benchmarks to compare and expose their characteristics and their use-cases. A study of industrial metrics requirements, as well as testing procedures, will be presented and applied to the studied benchmarks. We discuss our findings by examining the current state of benchmarks for industrial visual inspection, and by exposing guidelines on the usage of benchmarks.

Missing data can be informative. Ignoring this information can lead to misleading conclusions when the data model does not allow information to be extracted from the missing data. We propose a co-clustering model, based on the Latent Block Model, that aims to take advantage of this nonignorable nonresponses, also known as Missing Not At Random data (MNAR). A variational expectation-maximization algorithm is derived to perform inference and a model selection criterion is presented. We assess the proposed approach on a simulation study, before using our model on the voting records from the lower house of the French Parliament, where our analysis brings out relevant groups of MPs and texts, together with a sensible interpretation of the behavior of non-voters.

Camera-based end-to-end driving neural networks bring the promise of a low-cost system that maps camera images to driving control commands. These networks are appealing because they replace laborious hand engineered building blocks but their black-box nature makes them difficult to delve in case of failure. Recent works have shown the importance of using an explicit intermediate representation that has the benefits of increasing both the interpretability and the accuracy of networks' decisions. Nonetheless, these camera-based networks reason in camera view where scale is not homogeneous and hence not directly suitable for motion forecasting. In this paper, we introduce a novel monocular camera-only holistic end-to-end trajectory planning network with a Bird-Eye-View (BEV) intermediate representation that comes in the form of binary Occupancy Grid Maps (OGMs). To ease the prediction of OGMs in BEV from camera images, we introduce a novel scheme where the OGMs are first predicted as semantic masks in camera view and then warped in BEV using the homography between the two planes. The key element allowing this transformation to be applied to 3D objects such as vehicles, consists in predicting solely their footprint in camera-view, hence respecting the flat world hypothesis implied by the homography.

Learning generic representations with deep networks requires massive training samples and significant computer resources. To learn a new specific task, an important issue is to transfer the generic teacher's representation to a student network. In this paper, we propose to use a metric between representations that is based on a functional view of neurons. We use optimal transport to quantify the match between two representations, yielding a distance that embeds some invariances inherent to the representation of deep networks. This distance defines a regularizer promoting the similarity of the student's representation with that of the teacher. Our approach can be used in any learning context where representation transfer is applicable. We experiment here on two standard settings: inductive transfer learning, where the teacher's representation is transferred to a student network of same architecture for a new related task, and knowledge distillation, where the teacher's representation is transferred to a student of simpler architecture for the same task (model compression). Our approach also lends itself to solving new learning problems; we demonstrate this by showing how to directly transfer the teacher's representation to a simpler architecture student for a new related task.

In inductive transfer learning, fine-tuning pre-trained convolutional networks substantially outperforms training from scratch. When using fine-tuning, the underlying assumption is that the pre-trained model extracts generic features, which are at least partially relevant for solving the target task, but would be difficult to extract from the limited amount of data available on the target task. However, besides the initialization with the pre-trained model and the early stopping, there is no mechanism in fine-tuning for retaining the features learned on the source task. In this paper, we investigate several regularization schemes that explicitly promote the similarity of the final solution with the initial model. We show the benefit of having an explicit inductive bias towards the initial model, and we eventually recommend a simple $L^2$ penalty with the pre-trained model being a reference as the baseline of penalty for transfer learning tasks., Comment: Accepted at ICML 2018

This paper tackles the problem of endogenous link prediction for Knowledge Base completion. Knowledge Bases can be represented as directed graphs whose nodes correspond to entities and edges to relationships. Previous attempts either consist of powerful systems with high capacity to model complex connectivity patterns, which unfortunately usually end up overfitting on rare relationships, or in approaches that trade capacity for simplicity in order to fairly model all relationships, frequent or not. In this paper, we propose Tatec a happy medium obtained by complementing a high-capacity model with a simpler one, both pre-trained separately and then combined. We present several variants of this model with different kinds of regularization and combination strategies and show that this approach outperforms existing methods on different types of relationships by achieving state-of-the-art results on four benchmarks of the literature., Comment: 26 pages

Numerous variable selection methods rely on a two-stage procedure, where a sparsity-inducing penalty is used in the first stage to predict the support, which is then conveyed to the second stage for estimation or inference purposes. In this framework, the first stage screens variables to find a set of possibly relevant variables and the second stage operates on this set of candidate variables, to improve estimation accuracy or to assess the uncertainty associated to the selection of variables. We advocate that more information can be conveyed from the first stage to the second one: we use the magnitude of the coefficients estimated in the first stage to define an adaptive penalty that is applied at the second stage. We give two examples of procedures that can benefit from the proposed transfer of information, in estimation and inference problems respectively. Extensive simulations demonstrate that this transfer is particularly efficient when each stage operates on distinct subsamples. This separation plays a crucial role for the computation of calibrated p-values, allowing to control the False Discovery Rate. In this setup, the proposed transfer results in sensitivity gains ranging from 50% to 100% compared to state-of-the-art.

Non-linear performance measures are widely used for the evaluation of learning algorithms. For example, $F$-measure is a commonly used performance measure for classification problems in machine learning and information retrieval community. We study the theoretical properties of a subset of non-linear performance measures called pseudo-linear performance measures which includes $F$-measure, \emph{Jaccard Index}, among many others. We establish that many notions of $F$-measures and \emph{Jaccard Index} are pseudo-linear functions of the per-class false negatives and false positives for binary, multiclass and multilabel classification. Based on this observation, we present a general reduction of such performance measure optimization problem to cost-sensitive classification problem with unknown costs. We then propose an algorithm with provable guarantees to obtain an approximately optimal classifier for the $F$-measure by solving a series of cost-sensitive classification problems. The strength of our analysis is to be valid on any dataset and any class of classifiers, extending the existing theoretical results on pseudo-linear measures, which are asymptotic in nature. We also establish the multi-objective nature of the $F$-score maximization problem by linking the algorithm with the weighted-sum approach used in multi-objective optimization. We present numerical experiments to illustrate the relative importance of cost asymmetry and thresholding when learning linear classifiers on various $F$-measure optimization tasks., Comment: Extended Version of the NIPS 2014 Paper

This paper proposes a new interpretation of sparse penalties such as the elastic-net and the group-lasso. Beyond providing a new viewpoint on these penalization schemes, our approach results in a unified optimization strategy. Our experiments demonstrate that this strategy, implemented on the elastic-net, is computationally extremely efficient for small to medium size problems. Our accompanying software solves problems very accurately, at machine precision, in the time required to get a rough estimate with competing state-of-the-art algorithms. We illustrate on real and artificial datasets that this accuracy is required to for the correctness of the support of the solution, which is an important element for the interpretability of sparsity-inducing penalties.

We present a novel approach to the formulation and the resolution of sparse Linear Discriminant Analysis (LDA). Our proposal, is based on penalized Optimal Scoring. It has an exact equivalence with penalized LDA, contrary to the multi-class approaches based on the regression of class indicator that have been proposed so far. Sparsity is obtained thanks to a group-Lasso penalty that selects the same features in all discriminant directions. Our experiments demonstrate that this approach generates extremely parsimonious models without compromising prediction performances. Besides prediction, the resulting sparse discriminant directions are also amenable to low-dimensional representations of data. Our algorithm is highly efficient for medium to large number of variables, and is thus particularly well suited to the analysis of gene expression data., Comment: Appears in Proceedings of the 29th International Conference on Machine Learning (ICML 2012)

We consider the problems of estimation and selection of parameters endowed with a known group structure, when the groups are assumed to be sign-coherent, that is, gathering either nonnegative, nonpositive or null parameters. To tackle this problem, we propose the cooperative-Lasso penalty. We derive the optimality conditions defining the cooperative-Lasso estimate for generalized linear models, and propose an efficient active set algorithm suited to high-dimensional problems. We study the asymptotic consistency of the estimator in the linear regression setup and derive its irrepresentable conditions, which are milder than the ones of the group-Lasso regarding the matching of groups with the sparsity pattern of the true parameters. We also address the problem of model selection in linear regression by deriving an approximation of the degrees of freedom of the cooperative-Lasso estimator. Simulations comparing the proposed estimator to the group and sparse group-Lasso comply with our theoretical results, showing consistent improvements in support recovery for sign-coherent groups. We finally propose two examples illustrating the wide applicability of the cooperative-Lasso: first to the processing of ordinal variables, where the penalty acts as a monotonicity prior; second to the processing of genomic data, where the set of differentially expressed probes is enriched by incorporating all the probes of the microarray that are related to the corresponding genes., Comment: Published in at http://dx.doi.org/10.1214/11-AOAS520 the Annals of Applied Statistics (http://www.imstat.org/aoas/) by the Institute of Mathematical Statistics (http://www.imstat.org)

Gaussian Graphical Models provide a convenient framework for representing dependencies between variables. Recently, this tool has received a high interest for the discovery of biological networks. The literature focuses on the case where a single network is inferred from a set of measurements, but, as wetlab data is typically scarce, several assays, where the experimental conditions affect interactions, are usually merged to infer a single network. In this paper, we propose two approaches for estimating multiple related graphs, by rendering the closeness assumption into an empirical prior or group penalties. We provide quantitative results demonstrating the benefits of the proposed approaches. The methods presented in this paper are embeded in the R package 'simone' from version 1.0-0 and later.