Principal component regression

In statistics, principal component regression (PCR) is a regression analysis technique that is based on principal component analysis (PCA). More specifically, PCR is used for estimating the unknown regression coefficients in a standard linear regression model. In PCR, instead of regressing the dependent variable on the explanatory variables directly, the principal components of the explanatory variables are used as regressors. One typically uses only a subset of all the principal components for regression, making PCR a kind of regularized procedure and also a type of shrinkage estimator. Often the principal components with higher variances (the ones based on eigenvectors corresponding to the higher eigenvalues of the sample variance-covariance matrix of the explanatory variables) are selected as regressors. However, for the purpose of predicting the outcome, the principal components with low variances may also be important, in some cases even more important. One major use of PCR l

Theory of Deep Learning: Neural Tangent Kernel and Beyond

Arthur Ulysse Jacot-Guillarmod

In the recent years, Deep Neural Networks (DNNs) have managed to succeed at tasks that previously appeared impossible, such as human-level object recognition, text synthesis, translation, playing games and many more. In spite of these major achievements, our understanding of these models, in particular of what happens during their training, remains very limited. This PhD started with the introduction of the Neural Tangent Kernel (NTK) to describe the evolution of the function represented by the network during training. In the infinite-width limit, i.e. when the number of neurons in the layers of the network grows to infinity, the NTK converges to a deterministic and time-independent limit, leading to a simple yet complete description of the dynamics of infinitely-wide DNNs. This allowed one to give the first general proof of convergence of DNNs to a global minimum, and yielded the first description of the limiting spectrum of the Hessian of the loss surface of DNNs throughout training.More importantly, the NTK plays a crucial role in describing the generalization abilities of DNNs, i.e. the performance of the trained network on unseen data. The NTK analysis uncovered a direct link between the function learned by infinitely wide DNNs and Kernel Ridge Regression predictors, whose generalization properties are studied in this thesis using tools of random matrix theory. Our analysis of KRR reveals the importance of the eigendecomposition of the NTK, which is affected by a number of architectural choices. In very deep networks, an ordered regime and a chaotic regime appear, determined by the choice of non-linearity and the balance between the weights and bias parameters; these two phases are characterized by different speeds of decay of the eigenvalues of the NTK, leading to a tradeoff between convergence speed and generalization. In practical contexts such as Generative Adversarial Networks or Topology Optimization, the network architecture can be chosen to guarantee certain properties of the NTK and its spectrum.These results give an almost complete description DNNs in this infinite-width limit. It is then natural to wonder how it extends to finite-width networks used in practice. In the so-called NTK regime, the discrepancy between finite- and infinite-widths DNNs is mainly a result of the variance w.r.t. to the sampling of the parameters, as shown empirically and mathematically relying on the similarity between DNNs and random feature models.In contrast to the NTK regime, where the NTK remains constant during training, there exist so-called active regimes, where the evolution of the NTK is significant, which appear in a number of settings. One such regime appears in Deep Linear Networks with a very small initialization, where the training dynamics approach a sequence of saddle-points, representing linear maps of increasing rank, leading to a low-rank bias which is absent in the NTK regime.

EPFL2022

Locality defeats the curse of dimensionality in convolutional teacher-student scenarios

Francesco Cagnetta, Alessandro Favero, Matthieu Wyart

Convolutional neural networks perform a local and translationally-invariant treatment of the data: quantifying which of these two aspects is central to their success remains a challenge. We study this problem within a teacher-student framework for kernel regression, using `convolutional' kernels inspired by the neural tangent kernel of simple convolutional architectures of given filter size. Using heuristic methods from physics, we find in the ridgeless case that locality is key in determining the learning curve exponent

\beta

(that relates the test error

\epsilon_t\sim P^{-\beta}

to the size of the training set

P

), whereas translational invariance is not. In particular, if the filter size of the teacher

t

is smaller than that of the student

s

\beta

is a function of

s

only and does not depend on the input dimension. We confirm our predictions on

\beta

empirically. We conclude by proving, under a natural universality assumption, that performing kernel regression with a ridge that decreases with the size of the training set leads to similar learning curve exponents to those we obtain in the ridgeless case.

2021

Structure-property maps with Kernel principal covariates regression

Michele Ceriotti, Guillaume André Jean Fraux, Benjamin Aaron Helfrecht

Data analyses based on linear methods constitute the simplest, most robust, and transparent approaches to the automatic processing of large amounts of data for building supervised or unsupervised machine learning models. Principal covariates regression (PCovR) is an underappreciated method that interpolates between principal component analysis and linear regression and can be used conveniently to reveal structure-property relations in terms of simple-to-interpret, low-dimensional maps. Here we provide a pedagogic overview of these data analysis schemes, including the use of the kernel trick to introduce an element of non-linearity while maintaining most of the convenience and the simplicity of linear approaches. We then introduce a kernelized version of PCovR and a sparsified extension, and demonstrate the performance of this approach in revealing and predicting structure-property relations in chemistry and materials science, showing a variety of examples including elemental carbon, porous silicate frameworks, organic molecules, amino acid conformers, and molecular materials.

2020