On Inductive Biases for Machine Learning in Data Constrained Settings

Abstract

Learning with limited data is one of the biggest problems of machine learning. Current approaches to this issue consist in learning general representations from huge amounts of data before fine-tuning the model on a small dataset of interest. While such technique, coined transfer learning, is very effective in domains such as computer vision or natural langage processing, it does not yet solve common problems of deep learning such as model interpretability or the overall need for data. This thesis explores a different answer to the problem of learning expressive models in data constrained settings: instead of relying on big datasets to learn neural networks, we will replace some modules by known functions reflecting the structure of the data. Very often, these functions will be drawn from the rich literature of kernel methods. Indeed, many kernels can reflect the underlying structure of the data, thus sparing learning parameters to some extent. Our approach falls under the hood of "inductive biases", which can be defined as hypothesis on the data at hand restricting the space of models to explore during learning. We demonstrate the effectiveness of this approach in the context of sequences, such as sentences in natural language or protein sequences, and graphs, such as molecules. We also highlight the relationship between our work and recent advances in deep learning. Additionally, we study convex machine learning models. Here, rather than proposing new models, we wonder which proportion of the samples in a dataset is really needed to learn a "good" model. More precisely, we study the problem of safe sample screening, i.e, executing simple tests to discard uninformative samples from a dataset even before fitting a machine learning model, without affecting the optimal model. Such techniques can be used to prune datasets or mine for rare samples.

arXiv commentPhD thesis defended on January 19th, 2022

References in corpus (34)

• Scikit-learn: Machine Learning in Python
• An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
• Auto-Encoding Variational Bayes
• Neural Machine Translation by Jointly Learning to Align and Translate
• Intriguing properties of neural networks
• Generative Adversarial Networks
• Geometric deep learning: going beyond Euclidean data
• HuggingFace's Transformers: State-of-the-art Natural Language Processing
• Longformer: The Long-Document Transformer
• Scaling Laws for Neural Language Models
• Evaluating Large Language Models Trained on Code
• Linformer: Self-Attention with Linear Complexity
• CamemBERT: a Tasty French Language Model
• Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gauges
• Graph Transformer Networks
• Fast Computation of Wasserstein Barycenters
• A Generalization of Transformer Networks to Graphs
• Efficient Transformers: A Survey
• Diffusion Improves Graph Learning
• Graph-Bert: Only Attention is Needed for Learning Graph Representations
• Lasso Screening Rules via Dual Polytope Projection
• Graph Neural Tangent Kernel: Fusing Graph Neural Networks with Graph Kernels
• Safe Feature Elimination for the LASSO and Sparse Supervised Learning Problems
• White-box vs Black-box: Bayes Optimal Strategies for Membership Inference
• Measuring the Algorithmic Efficiency of Neural Networks
• A Safe Screening Rule for Sparse Logistic Regression
• GraphiT: Encoding Graph Structure in Transformers
• Multilevel Clustering via Wasserstein Means
• A Kernel Perspective for Regularizing Deep Neural Networks
• Convolutional Kernel Networks for Graph-Structured Data
• On the Definiteness of Earth Mover's Distance and Its Relation to Set Intersection
• Recurrent Kernel Networks
• Safe Sample Screening for Support Vector Machines
• Screening Data Points in Empirical Risk Minimization via Ellipsoidal Regions and Safe Loss Functions