Neural networks often learn simple explanations that fit the majority of the data while memorizing exceptions that deviate from these explanations. This behavior leads to poor generalization when the learned explanations rely on spurious correlations. In this work, we formalize $\textit{the interplay between memorization and generalization}$, showing that spurious correlations would particularly lead to poor generalization when are combined with memorization. Memorization can reduce training loss to zero, leaving no incentive to learn robust, generalizable patterns. To address this, we propose $\textit{memorization-aware training}$ (MAT), which uses held-out predictions as a signal of memorization to shift a model's logits. MAT encourages learning robust patterns invariant across distributions, improving generalization under distribution shifts.

Describing skills in natural language has the potential to provide an accessible way to inject human knowledge about decision-making into an AI system. We present MaestroMotif, a method for AI-assisted skill design, which yields high-performing and adaptable agents. MaestroMotif leverages the capabilities of Large Language Models (LLMs) to effectively create and reuse skills. It first uses an LLM's feedback to automatically design rewards corresponding to each skill, starting from their natural language description. Then, it employs an LLM's code generation abilities, together with reinforcement learning, for training the skills and combining them to implement complex behaviors specified in language. We evaluate MaestroMotif using a suite of complex tasks in the NetHack Learning Environment (NLE), demonstrating that it surpasses existing approaches in both performance and usability.

Compositional generalization is a crucial step towards developing data-efficient intelligent machines that generalize in human-like ways. In this work, we tackle a challenging form of distribution shift, termed compositional shift, where some attribute combinations are completely absent at training but present in the test distribution. This shift tests the model’s ability to generalize compositionally to novel attribute combinations in discriminative tasks. We model the data with flexible additive energy distributions, where each energy term represents an attribute, and derive a simple alternative to empirical risk minimization termed compositional risk minimization (CRM). We first train an additive energy classifier to predict the multiple attributes and then adjust this classifier to tackle compositional shifts. We provide an extensive theoretical analysis of CRM, where we show that our proposal extrapolates to special affine hulls of seen attribute combinations. Empirical evaluations on benchmark datasets confirms the improved robustness of CRM compared to other methods from the literature designed to tackle various forms of subpopulation shifts.

Exploring rich environments and evaluating one's actions without prior knowledge is immensely challenging. In this paper, we propose Motif, a general method to interface such prior knowledge from a Large Language Model (LLM) with an agent. Motif is based on the idea of grounding LLMs for decision-making without requiring them to interact with the environment: it elicits preferences from an LLM over pairs of captions to construct an intrinsic reward, which is then used to train agents with reinforcement learning. We evaluate Motif's performance and behavior on the challenging, open-ended and procedurally-generated NetHack game. Surprisingly, by only learning to maximize its intrinsic reward, Motif achieves a higher game score than an algorithm directly trained to maximize the score itself. When combining Motif's intrinsic reward with the environment reward, our method significantly outperforms existing approaches and makes progress on tasks where no advancements have ever been made without demonstrations. Finally, we show that Motif mostly generates intuitive human-aligned behaviors which can be steered easily through prompt modifications, while scaling well with the LLM size and the amount of information given in the prompt.

Masked Image Modeling (MIM) is a promising self-supervised learning approach that enables learning from unlabeled images. Despite its recent success, learning good representations through MIM remains challenging because it requires predicting the right semantic content in accurate locations. For example, given an incomplete picture of a dog, we can guess that there is a tail, but we cannot determine its exact location. In this work, we propose to incorporate location uncertainty to MIM by using stochastic positional embeddings (StoP). Specifically, we condition the model on stochastic masked token positions drawn from a gaussian distribution. We show that using StoP reduces overfitting to location features and guides the model toward learning features that are more robust to location uncertainties. Quantitatively, using StoP improves downstream MIM performance on a variety of downstream tasks. For example, linear probing on ImageNet using ViT-B is improved by $+1.7\%$, and by $2.5\%$ for ViT-H using 1% of the data.

Environment annotations are essential for the success of many out-of-distribution (OOD) generalization methods. Unfortunately, these are costly to obtain and often limited by human annotators’ biases. To achieve robust generalization, it is essential to develop algorithms for automatic environment discovery within datasets. Current proposals, which divide examples based on their training error, suffer from one fundamental problem. These methods introduce hyper-parameters and early-stopping criteria, which require a validation set with human-annotated environments, the very information subject to discovery. In this paper, we propose Cross-Risk Minimization (XRM) to address this issue. XRM trains twin networks, each learning from one random half of the training data, while imitating confident held-out mistakes made by its sibling. XRM provides a recipe for hyper-parameter tuning, does not require early-stopping, and can discover environments for all training and validation data. Algorithms built on top of XRM environments achieve oracle worst-group-accuracy, addressing a long-standing challenge in OOD generalization.

This paper demonstrates an approach for learning highly semantic image representations without relying on hand-crafted data-augmentations. We introduce the Image-based Joint-Embedding Predictive Architecture (I-JEPA), a non-generative approach for self-supervised learning from images. The idea behind I-JEPA is simple: from a single context block, predict the representations of various target blocks in the same image. A core design choice to guide I-JEPA towards producing semantic representations is the masking strategy; specifically, it is crucial to (a) sample target blocks with sufficiently large scale (semantic), and to (b) use a sufficiently informative (spatially distributed) context block. Empirically, when combined with Vision Transformers, we find I-JEPA to be highly scalable. For instance, we train a ViT-Huge/14 on ImageNet using 16 A100 GPUs in under 72 hours to achieve strong downstream performance across a wide range of tasks, from linear classification to object counting and depth prediction.

Synthetic image datasets offer unmatched advantages for designing and evaluating deep neural networks: they make it possible to (i) render as many data samples as needed, (ii) precisely control each scene and yield granular ground truth labels (and captions), (iii) precisely control distribution shifts between training and testing to isolate variables of interest for sound experimentation.Despite such promise, the use of synthetic image data is still limited -- and often played down -- mainly due to their lack of realism. Most works therefore rely on datasets of real images, which have often been scraped from public images on the internet, and may have issues with regards to privacy, bias, and copyright, while offering little control over how objects precisely appear.In this work, we present a path to democratize the use of photorealistic synthetic data: we develop a new generation of interactive environments for representation learning research, that offer both controllability and realism. We use the Unreal Engine, a powerful game engine well known in the entertainment industry, to produce PUG (Photorealistic Unreal Graphics) environments and datasets for representation learning. Using PUG for evaluation and fine-tuning, we demonstrate the potential of PUG to both enable more rigorous evaluations and to improve model training.

Self-supervised learning (SSL) algorithms can produce useful image representations by learning to associate different parts of natural images with one another. However, when taken to the extreme, SSL models can unintendedly memorize specific parts in individual training samples rather than learning semantically meaningful associations. In this work, we perform a systematic study of the unintended memorization of image-specific information in SSL models -- which we refer to as déjà vu memorization. Concretely, we show that given the trained model and a crop of a training image containing only the background (e.g., water, sky, grass), it is possible to infer the foreground object with high accuracy or even visually reconstruct it. Furthermore, we show that déjà vu memorization is common to different SSL algorithms, is exacerbated by certain design choices, and cannot be detected by conventional techniques for evaluating representation quality. Our study of déjà vu memorization reveals previously unknown privacy risks in SSL models, as well as suggests potential practical mitigation strategies.

Abstract Computational approaches to the study of language emergence can help us understand how natural languages are shaped by cognitive and sociocultural factors. Previous work focused on tasks where agents refer to a single entity. In contrast, we study how agents predicate, that is, how they express that some relation holds between several entities. We introduce a setup where agents talk about a variable number of entities that can be partially observed by the listener. In the presence of a least-effort pressure, they tend to discuss only entities that are not observed by the listener. Thus we can obtain artificial phrases that denote a single entity, as well as artificial sentences that denote several entities. In natural languages, if we ignore the verb, phrases are usually concatenated, either in a specific order or by adding case markers to form sentences. Our setup allows us to quantify how much this holds in emergent languages using a metric we call concatenability. We also measure transitivity, which quantifies the importance of word order. We demonstrate the usefulness of this new setup and metrics for studying factors that influence argument structure. We compare agents having access to input representations structured into pre-segmented objects with properties, versus unstructured representations. Our results indicate that the awareness of object structure yields a more natural sentence organization.

We approach the problem of implicit regularization in deep learning from a geometrical viewpoint. We highlight a regularization effect induced by a dynamical alignment ofthe neural tangent features introduced by Jacot et al. (2018), along a small number of task-relevant directions. This can be interpreted as a combined mechanism of feature selection and compression. By extrapolating a new analysis of Rademacher complexity bounds for linear models, we motivate and study a heuristic complexity measure that captures this phenomenon, in terms of sequences of tangent kernel classes along optimization paths. The code for our experiments is available as https://github.com/tfjgeorge/ntk_alignment.

Recent advances in variational inference enable the modelling of highly structured joint distributions, but are limited in their capacity to scale to the high-dimensional setting of stochastic neural networks. This limitation motivates a need for scalable parameterizations of the noise generation process, in a manner that adequately captures the dependencies among the various parameters. In this work, we address this need and present the Kronecker Flow, a generalization of the Kronecker product to invertible mappings designed for stochastic neural networks. We apply our method to variational Bayesian neural networks on predictive tasks, PAC-Bayes generalization bound estimation, and approximate Thompson sampling in contextual bandits. In all setups, our methods prove to be competitive with existing methods and betterthan the baselines.

Autoregressive language models are powerful and relatively easy to train.However, these models are usually trained without explicit conditioning labels and do not offer easy ways to control global aspects such as sentiment or topic during generation.Bowman et al. (2016) adapted the Variational Autoencoder (VAE) for natural language with the sequenceto-sequence architecture and claimed that the latent vector was able to capture such global features in an unsupervised manner.We question this claim.We measure which words benefit most from the latent information by decomposing the reconstruction loss per position in the sentence.Using this method, we find that VAEs are prone to memorizing the first words and the sentence length, producing local features of limited usefulness.To alleviate this, we investigate alternative architectures based on bag-of-words assumptions and language model pretraining.These variants learn latent variables that are more global, i.e., more predictive of topic or sentiment labels.Moreover, using reconstructions, we observe that they decrease memorization: the first word and the sentence length are not recovered as accurately than with the baselines, consequently yielding more diverse reconstructions.

The success of adversarial formulations in machine learning has brought renewed motivation for smooth games. In this work, we focus on the class of stochastic Hamiltonian methods and provide the first convergence guarantees for certain classes of stochastic smooth games. We propose a novel unbiased estimator for the stochastic Hamiltonian gradient descent (SHGD) and highlight its benefits. Using tools from the optimization literature we show that SHGD converges linearly to the neighbourhood of a stationary point. To guarantee convergence to the exact solution, we analyze SHGD with a decreasing step-size and we also present the first stochastic variance reduced Hamiltonian method. Our results provide the first global non-asymptotic last-iterate convergence guarantees for the class of stochastic unconstrained bilinear games and for the more general class of stochastic games that satisfy a “sufficiently bilinear" condition, notably including some non-convex non-concave problems. We supplement our analysis with experiments on stochastic bilinear and sufficiently bilinear games, where our theory is shown to be tight, and on simple adversarial machine learning formulations.

Stochastic variance-reduced gradient (SVRG) is an optimization method originally designed for tackling machine learning problems with a finite sum structure. SVRG was later shown to work for policy evaluation, a problem in reinforcement learning in which one aims to estimate the value function of a given policy. SVRG makes use of gradient estimates at two scales. At the slower scale, SVRG computes a full gradient over the whole dataset, which could lead to prohibitive computation costs. In this work, we show that two variants of SVRG for policy evaluation could significantly diminish the number of gradient calculations while preserving a linear convergence speed. More importantly, our theoretical result implies that one does not need to use the entire dataset in every epoch of SVRG when it is applied to policy evaluation with linear function approximation. Our experiments demonstrate large computational savings provided by the proposed methods.

The existence of adversarial examples capable of fooling trained neural network classifiers calls for a much better understanding of possible attacks to guide the development of safeguards against them. This includes attack methods in the challenging {\em non-interactive blackbox} setting, where adversarial attacks are generated without any access, including queries, to the target model. Prior attacks in this setting have relied mainly on algorithmic innovations derived from empirical observations (e.g., that momentum helps), lacking principled transferability guarantees. In this work, we provide a theoretical foundation for crafting transferable adversarial examples to entire hypothesis classes. We introduce \textit{Adversarial Example Games} (AEG), a framework that models the crafting of adversarial examples as a min-max game between a generator of attacks and a classifier. AEG provides a new way to design adversarial examples by adversarially training a generator and a classifier from a given hypothesis class (e.g., architecture). We prove that this game has an equilibrium, and that the optimal generator is able to craft adversarial examples that can attack any classifier from the corresponding hypothesis class. We demonstrate the efficacy of AEG on the MNIST and CIFAR-10 datasets, outperforming prior state-of-the-art approaches with an average relative improvement of $29.9\%$ and $47.2\%$ against undefended and robust models (Table \ref{table:q2} \& \ref{table:q3}) respectively.

Trust Region Policy Optimization (TRPO) and Proximal Policy Optimization (PPO) are among the most successful policy gradient approaches in deep reinforcement learning (RL). While these methods achieve state-of-the-art performance across a wide range of challenging tasks, there is room for improvement in the stabilization of the policy learning and how the off-policy data are used. In this paper we revisit the theoretical foundations of these algorithms and propose a new algorithm which stabilizes the policy improvement through a proximity term that constrains the discounted state-action visitation distribution induced by consecutive policies to be close to one another. This proximity term, expressed in terms of the divergence between the visitation distributions, is learned in an off-policy and adversarial manner. We empirically show that our proposed method can have a beneficial effect on stability and improve final performance in benchmark high-dimensional control tasks.

The goal of MRI reconstruction is to restore a high fidelity image from partially observed measurements. This partial view naturally induces reconstruction uncertainty that can only be reduced by acquiring additional measurements. In this paper, we present a novel method for MRI reconstruction that, at inference time, dynamically selects the measurements to take and iteratively refines the prediction in order to best reduce the reconstruction error and, thus, its uncertainty. We validate our method on a large scale knee MRI dataset, as well as on ImageNet. Results show that (1) our system successfully outperforms active acquisition baselines; (2) our uncertainty estimates correlate with error maps; and (3) our ResNet-based architecture surpasses standard pixel-to-pixel models in the task of MRI reconstruction. The proposed method not only shows high-quality reconstructions but also paves the road towards more applicable solutions for accelerating MRI.

The apparent contradiction in the title is a wordplay on the different meanings attributed to the word reproducible across different scientific fields. What we imply is that unreproducible findings can be built upon reproducible methods. Without denying the importance of facilitating the reproduction of methods, we deem important to reassert that reproduction of findings is a fundamental step of the scientific inquiry. We argue that the commendable quest towards easy deterministic reproducibility of methods and numerical results should not have us forget the even more important necessity of ensuring the reproducibility of empirical findings and conclusions by properly accounting for essential sources of variations. We provide experiments to exemplify the brittleness of current common practice in the evaluation of models in the field of deep learning, showing that even if the results could be reproduced, a slightly different experiment would not support the findings. We hope to help clarify the distinction between exploratory and empirical research in the field of deep learning and believe more energy should be devoted to proper empirical research in our community. This work is an attempt to promote the use of more rigorous and diversified methodologies. It is not an attempt to impose a new methodology and it is not a critique on the nature of exploratory research.

We present two techniques to improve landmark localization in images from partially annotated datasets. Our primary goal is to leverage the common situation where precise landmark locations are only provided for a small data subset, but where class labels for classification or regression tasks related to the landmarks are more abundantly available. First, we propose the framework of sequential multitasking and explore it here through an architecture for landmark localization where training with class labels acts as an auxiliary signal to guide the landmark localization on unlabeled data. A key aspect of our approach is that errors can be backpropagated through a complete landmark localization model. Second, we propose and explore an unsupervised learning technique for landmark localization based on having a model predict equivariant landmarks with respect to transformations applied to the image. We show that these techniques, improve landmark prediction considerably and can learn effective detectors even when only a small fraction of the dataset has landmark labels. We present results on two toy datasets and four real datasets, with hands and faces, and report new state-of-the-art on two datasets in the wild, e.g. with only 5% of labeled images we outperform previous state-of-the-art trained on the AFLW dataset.

Monolingual dictionaries are widespread and semantically rich resources.This paper presents a simple model that learns to compute word embeddings by processing dictionary definitions and trying to reconstruct them.It exploits the inherent recursivity of dictionaries by encouraging consistency between the representations it uses as inputs and the representations it produces as outputs.The resulting embeddings are shown to capture semantic similarity better than regular distributional methods and other dictionary-based methods.In addition, the method shows strong performance when trained exclusively on dictionary data and generalizes in one shot.

Off-policy learning is key to scaling up reinforcement learning as it allows to learn about a target policy from the experience generated by a different behavior policy. Unfortunately, it has been challenging to combine off-policy learning with function approximation and multi-step bootstrapping in a way that leads to both stable and efficient algorithms. In this work, we show that the Tree Backup and Retrace algorithms are unstable with linear function approximation, both in theory and in practice with specific examples. Based on our analysis, we then derive stable and efficient gradient-based algorithms using a quadratic convex-concave saddle-point formulation. By exploiting the problem structure proper to these algorithms, we are able to provide convergence guarantees and finite-sample bounds. The applicability of our new analysis also goes beyond Tree Backup and Retrace and allows us to provide new convergence rates for the GTD and GTD2 algorithms without having recourse to projections or Polyak averaging.

Optimization algorithms that leverage gradient covariance information, such as variants of natural gradient descent (Amari, 1998), offer the prospect of yielding more effective descent directions. For models with many parameters, the covari- ance matrix they are based on becomes gigantic, making them inapplicable in their original form. This has motivated research into both simple diagonal approxima- tions and more sophisticated factored approximations such as KFAC (Heskes, 2000; Martens & Grosse, 2015; Grosse & Martens, 2016). In the present work we draw inspiration from both to propose a novel approximation that is provably better than KFAC and amendable to cheap partial updates. It consists in tracking a diagonal variance, not in parameter coordinates, but in a Kronecker-factored eigenbasis, in which the diagonal approximation is likely to be more effective. Experiments show improvements over KFAC in optimization speed for several deep network architectures.

Deep neural networks with alternating convolutional, max-pooling and decimation layers are widely used in state of the art architectures for computer vision. Max-pooling purposefully discards precise spatial information in order to create features that are more robust, and typically organized as lower resolution spatial feature maps. On some tasks, such as whole-image classification, max-pooling derived features are well suited, however, for tasks requiring precise localization, such as pixel level prediction and segmentation, max-pooling destroys exactly the information required to perform well. Precise localization may be preserved by shallow convnets without pooling but at the expense of robustness. Can we have our max-pooled multilayered cake and eat it too? Several papers have proposed summation and concatenation based methods for combining upsampled coarse, abstract features with finer features to produce robust pixel level predictions. Here we introduce another model - dubbed Recombinator Networks - where coarse features inform finer features early in their formation such that finer features can make use of several layers of computation in deciding how to use coarse features. The model is trained once, end-to-end and performs better than summation-based architectures, reducing the error from the previous state of the art on two facial keypoint datasets, AFW and AFLW, by 30% and beating the current state-of-the-art on 300W without using extra data. We improve performance even further by adding a denoising prediction model based on a novel convnet formulation.

An important class of problems involves training deep neural networks with sparse prediction targets of very high dimension D. These occur naturally in e.g. neural language models or the learning of word-embeddings, often posed as predicting the probability of next words among a vocabulary of size D (e.g. 200,000). Computing the equally large, but typically non-sparse D-dimensional output vector from a last hidden layer of reasonable dimension d (e.g. 500) incurs a prohibitive $O(Dd)$ computational cost for each example, as does updating the $D \times d$ output weight matrix and computing the gradient needed for backpropagation to previous layers. While efficient handling of large sparse network inputs is trivial, this case of large sparse targets is not, and has thus so far been sidestepped with approximate alternatives such as hierarchical softmax or sampling-based approximations during training. In this work we develop an original algorithmic approach that, for a family of loss functions that includes squared error and spherical softmax, can compute the exact loss, gradient update for the output weights, and gradient for backpropagation, all in $O(d^2)$ per example instead of $O(Dd)$, remarkably without ever computing the D-dimensional output. The proposed algorithm yields a speedup of $\frac{D}{4d}$, i.e. two orders of magnitude for typical sizes, for that critical part of the computations that often dominates the training time in this kind of network architecture.

We investigate the problem of transforming an input sequence into a high-dimensional output sequence in order to transcribe polyphonic audio music into symbolic notation. We introduce a probabilistic model based on a recurrent neural network that is able to learn realistic output distributions given the input and we devise an efficient algorithm to search for the global mode of that distribution. The resulting method produces musically plausible transcriptions even under high levels of noise and drastically outperforms previous state-of- the-art approaches on five datasets of synthesized sounds and real recordings, approximately halving the test error rate.

In this paper we present the techniques used for the University of Montréal's team submissions to the 2013 Emotion Recognition in the Wild Challenge. The challenge is to classify the emotions expressed by the primary human subject in short video clips extracted from feature length movies. This involves the analysis of video clips of acted scenes lasting approximately one-two seconds, including the audio track which may contain human voices as well as background music. Our approach combines multiple deep neural networks for different data modalities, including: (1) a deep convolutional neural network for the analysis of facial expressions within video frames; (2) a deep belief net to capture audio information; (3) a deep autoencoder to model the spatio-temporal information produced by the human actions depicted within the entire scene; and (4) a shallow network architecture focused on extracted features of the mouth of the primary human subject in the scene. We discuss each of these techniques, their performance characteristics and different strategies to aggregate their predictions. Our best single model was a convolutional neural network trained to predict emotions from static frames using two large data sets, the Toronto Face Database and our own set of faces images harvested from Google image search, followed by a per frame aggregation strategy that used the challenge training data. This yielded a test set accuracy of 35.58%. Using our best strategy for aggregating our top performing models into a single predictor we were able to produce an accuracy of 41.03% on the challenge test set. These compare favorably to the challenge baseline test set accuracy of 27.56%.

Recent work has shown how denoising and contractive autoencoders implicitly capture the structure of the data generating density, in the case where the corruption noise is Gaussian, the reconstruction error is the squared error, and the data is continuous-valued. This has led to various proposals for sampling from this implicitly learned density function, using Langevin and Metropolis-Hastings MCMC. However, it remained unclear how to connect the training procedure of regularized auto-encoders to the implicit estimation of the underlying data generating distribution when the data are discrete, or using other forms of corruption process and reconstruction errors. Another issue is the mathematical justification which is only valid in the limit of small corruption noise. We propose here a different attack on the problem, which deals with all these issues: arbitrary (but noisy enough) corruption, arbitrary reconstruction loss (seen as a log-likelihood), handling both discrete and continuous-valued variables, and removing the bias due to non-infinitesimal corruption noise (or non-infinitesimal contractive penalty).

The success of machine learning algorithms generally depends on data representation, and we hypothesize that this is because different representations can entangle and hide more or less the different explanatory factors of variation behind the data. Although specific domain knowledge can be used to help design representations, learning with generic priors can also be used, and the quest for AI is motivating the design of more powerful representation-learning algorithms implementing such priors. This paper reviews recent work in the area of unsupervised feature learning and deep learning, covering advances in probabilistic models, autoencoders, manifold learning, and deep networks. This motivates longer term unanswered questions about the appropriate objectives for learning good representations, for computing representations (i.e., inference), and the geometrical connections between representation learning, density estimation, and manifold learning.

We combine three important ideas present in previous work for building classi- fiers: the semi-supervised hypothesis (the input distribution contains information about the classifier), the unsupervised manifold hypothesis (data density concen- trates near low-dimensional manifolds), and the manifold hypothesis for classifi- cation (different classes correspond to disjoint manifolds separated by low den- sity). We exploit a novel algorithm for capturing manifold structure (high-order contractive auto-encoders) and we show how it builds a topological atlas of charts, each chart being characterized by the principal singular vectors of the Jacobian of a representation mapping. This representation learning algorithm can be stacked to yield a deep architecture, and we combine it with a domain knowledge-free version of the TangentProp algorithm to encourage the classifier to be insensitive to local directions changes along the manifold. Record-breaking classification results are obtained.

Two recently proposed learning algorithms, herding and fast persistent contrastive divergence (FPCD), share the following interesting characteristic: they exploit changes in the model parameters while sampling in order to escape modes and mix better during the sampling process that is part of the learning algorithm. We justify such approaches as ways to escape modes while keeping approximately the same asymptotic distribution of the Markov chain. In that spirit, we extend FPCD using an idea borrowed from Herding in order to obtain a pure sampling algorithm, which we call the rates-FPCD sampler. Interestingly, this sampler can improve the model as we collect more samples, since it optimizes a lower bound on the log likelihood of the training data. We provide empirical evidence that this new algorithm displays substantially better and more robust mixing than Gibbs sampling.

Denoising autoencoders have been previously shown to be competitive alternatives to restricted Boltzmann machines for unsupervised pretraining of each layer of a deep architecture. We show that a simple denoising autoencoder training criterion is equivalent to matching the score (with respect to the data) of a specific energy-based model to that of a nonparametric Parzen density estimator of the data. This yields several useful insights. It defines a proper probabilistic model for the denoising autoencoder technique, which makes it in principle possible to sample from them or rank examples by their energy. It suggests a different way to apply score matching that is related to learning to denoise and does not require computing second derivatives. It justifies the use of tied weights between the encoder and decoder and suggests ways to extend the success of denoising autoencoders to a larger family of energy-based models.

Previous work has shown that the difficulties in learning deep generative or discriminative models can be overcome by an initial unsupervised learning step that maps inputs to useful intermediate representations. We introduce and motivate a new training principle for unsupervised learning of a representation based on the idea of making the learned representations robust to partial corruption of the input pattern. This approach can be used to train autoencoders, and these denoising autoencoders can be stacked to initialize deep architectures. The algorithm can be motivated from a manifold learning and information theoretic perspective or from a generative model perspective. Comparative experiments clearly show the surprising advantage of corrupting the input of autoencoders on a pattern classification benchmark suite. 1.

To escape from the curse of dimensionality, we claim that one can learn non-local functions, in the sense that the value and shape of the learned function at x must be inferred using examples that may be far from x. With this objective, we present a non-local non-parametric density esti- mator. It builds upon previously proposed Gaussian mixture models with regularized covariance matrices to take into account the local shape of the manifold. It also builds upon recent work on non-local estimators of the tangent plane of a manifold, which are able to generalize in places with little training data, unlike traditional, local, non-parametric models.

Convexity has recently received a lot of attention in the machine learning community, and the lack of convexity has been seen as a major disadvantage of many learning algorithms, such as multi-layer artificial neural networks. We show that training multi-layer neural networks in which the number of hidden units is learned can be viewed as a convex optimization problem. This problem involves an infinite number of variables, but can be solved by incrementally inserting a hidden unit at a time, each time finding a linear classifier that minimizes a weighted sum of errors.

In this letter, we show a direct relation between spectral embedding methods and kernel principal components analysis and how both are special cases of a more general learning problem: learning the principal eigenfunctions of an operator defined from a kernel and the unknown data-generating density. Whereas spectral embedding methods provided only coordinates for the training points, the analysis justifies a simple extension to out-of-sample examples (the Nyström formula) for multidimensional scaling (MDS), spectral clustering, Laplacian eigenmaps, locally linear embedding (LLE), and Isomap. The analysis provides, for all such spectral embedding methods, the definition of a loss function, whose empirical average is minimized by the traditional algorithms. The asymptotic expected value of that loss defines a generalization performance and clarifies what these algorithms are trying to learn. Experiments with LLE, Isomap, spectral clustering, and MDS show that this out-of-sample embedding formula generalizes well, with a level of error comparable to the effect of small perturbations of the training set on the embedding.

Several unsupervised learning algorithms based on an eigendecompo- sition provide either an embedding or a clustering only for given train- ing points, with no straightforward extension for out-of-sample examples short of recomputing eigenvectors. This paper provides a unified frame- work for extending Local Linear Embedding (LLE), Isomap, Laplacian Eigenmaps, Multi-Dimensional Scaling (for dimensionality reduction) as well as for Spectral Clustering. This framework is based on seeing these algorithms as learning eigenfunctions of a data-dependent kernel. Numerical experiments show that the generalizations performed have a level of error comparable to the variability of the embedding algorithms due to the choice of training data.

The similarity between objects is a fundamental element of many learn- ing algorithms. Most non-parametric methods take this similarity to be fixed, but much recent work has shown the advantages of learning it, in particular to exploit the local invariances in the data or to capture the possibly non-linear manifold on which most of the data lies. We propose a new non-parametric kernel density estimation method which captures the local structure of an underlying manifold through the leading eigen- vectors of regularized local covariance matrices. Experiments in density estimation show significant improvements with respect to Parzen density estimators. The density estimators can also be used within Bayes classi- fiers, yielding classification rates similar to SVMs and much superior to the Parzen classifier.

Estimating insurance premia from data is a difficult regression problem for several reasons: the large number of variables, many of which are .discrete, and the very peculiar shape of the noise distri(cid:173) bution, asymmetric with fat tails, with a large majority zeros and a few unreliable and very large values. We compare several machine learning methods for estimating insurance premia, and test them on a large data base of car insurance policies. We find that func(cid:173) tion approximation methods that do not optimize a squared loss, like Support Vector Machines regression, do not work well in this context. Compared methods include decision trees and generalized linear models. The best results are obtained with a mixture of experts, which better identifies the least and most risky contracts, and allows to reduce the median premium by charging more to the most risky customers.

Guided by an initial idea of building a complex (non linear) decision surface with maximal local margin in input space, we give a possible geometrical intuition as to why K-Nearest Neighbor (KNN) algorithms often perform more poorly than SVMs on classification tasks. We then propose modified K-Nearest Neighbor algorithms to overcome the per- ceived problem. The approach is similar in spirit to Tangent Distance, but with invariances inferred from the local neighborhood rather than prior knowledge. Experimental results on real world classification tasks sug- gest that the modified KNN algorithms often give a dramatic improve- ment over standard KNN and perform as well or better than SVMs.

In the Support Vector Machines (SVM) framework, the positive-definite kernel can be seen as representing a fixed similarity measure between two patterns, and a discriminant function is obtained by taking a linear combination of the kernels computed at training examples called support vectors. We investigate learning architectures in which the kernel functions can be replaced by more general similarity measures that can have arbitrary internal parameters. The training criterion used in SVMs is not appropriate for this purpose so we adopt the simple criterion that is generally used when training neural networks for classification tasks. Several experiments are performed which show that such Neural Support Vector Networks perform similarly to SVMs while requiring significantly fewer support vectors, even when the similarity measure has no internal parameters.

A goal of statistical language modeling is to learn the joint probability function of sequences of words. This is intrinsically difficult because of the curse of dimensionality: we propose to fight it with its own weapons. In the proposed approach one learns simultaneously (1) a distributed rep(cid:173) resentation for each word (i.e. a similarity between words) along with (2) the probability function for word sequences, expressed with these repre(cid:173) sentations. Generalization is obtained because a sequence of words that has never been seen before gets high probability if it is made of words that are similar to words forming an already seen sentence. We report on experiments using neural networks for the probability function, showing on two text corpora that the proposed approach very significantly im(cid:173) proves on a state-of-the-art trigram model.

We present a new image compression technique called "DjVu" that is specifically geared towards the compression of scanned documents in color at high resolution. With DjVu, a magazine page in color at 300 dpi typically occupies between 40 KB and 80 KB, approximately 5 to 10 times better than JPEG for a similar level of readability. Using a combination of hidden Markov model techniques and MDL-driven heuristics, DjVu first classifies each pixel in the image as either foreground (text, drawings) or background (pictures, photos, paper texture). The pixel categories form a bitonal image which is compressed using a pattern matching technique that takes advantage of the similarities between character shapes. A progressive, wavelet-based compression technique, combined with a masking algorithm, is then used to compress the foreground and background images at lower resolutions while minimizing the number of bits spent on the pixels that are not visible in the foreground and background planes. Encoders, decoders, and real-time, memory efficient plug-ins for various web browsers are available for all the major platforms.