Publication
Efficient local linearity regularization to overcome catastrophic overfitting
Related publications (44)
Generalization error in high-dimensional perceptrons: Approaching Bayes error with convex optimization
Florent Gérard Krzakala, Lenka Zdeborová
We consider a commonly studied supervised classification of a synthetic dataset whose labels are generated by feeding a one-layer non-linear neural network with random iid inputs. We study the generalization performances of standard classifiers in the high ...
Curran Associates, Inc.2020Learning Multi-Label Aerial Image Classification Under Label Noise: A Regularization Approach Using Word Embeddings
Training deep neural networks requires well-annotated datasets. However, real world datasets are often noisy, especially in a multi-label scenario, i.e. where each data point can be attributed to more than one class. To this end, we propose a regularizatio ...
IEEE2020Regularization via Mass Transportation
Daniel Kuhn, Soroosh Shafieezadeh Abadeh, Peyman Mohajerin Esfahani
The goal of regression and classification methods in supervised learning is to minimize the empirical risk, that is, the expectation of some loss function quantifying the prediction error under the empirical distribution. When facing scarce training data, ...
2019CoCoA: A General Framework for Communication-Efficient Distributed Optimization
The scale of modern datasets necessitates the development of efficient distributed optimization methods for machine learning. We present a general-purpose framework for distributed computing environments, CoCoA, that has an efficient communication scheme a ...
MICROTOME PUBL2018Peptide-Cation Systems: Conformational Search, Benchmark Evaluation, and Force Field Parameter Adjustment Using Regularized Linear Regression
Metal cations often play an important role in shaping the three-dimensional structure of peptides. As an example, the model system AcPheAla5LysH+ is investigated in order to fully understand the forces that stabilize its helical structure. In particular, t ...
EPFL2018Robust and adaptive approaches for Relative Geologic Time Estimation
For a geoscientist, the Relative Geologic Time (RGT) is an important tool to perform chronostratigraphic analysis. However, automatically estimate an RGT image from a seismic image can be a challenging task where we have to respect seismic features, the de ...
ELSEVIER SCIENCE BV2018Sparsistency of $\ell_1$-Regularized $M$-Estimators
Volkan Cevher, Jonathan Mark Scarlett, Yen-Huan Li
We consider the model selection consistency or sparsistency of a broad set of -regularized -estimators for linear and non-linear statistical models in a unified fashion. For this purpose, we propose the local structured smoothness condition (LSS ...
2015Enhancing statistical performance of data-driven controller tuning via L2-regularization
Noniterative data-driven techniques are design methods that allow optimal feedback control laws to be derived from input-output (I/O) data only, without the need of a model of the process. A drawback of these methods is that, in their standard formulation, ...
Pergamon-Elsevier Science Ltd2014On L2-Regularization for Virtual Reference Feedback Tuning
The Virtual Reference Feedback Tuning (VRFT) approach is a design method that allow optimal feedback control laws to be derived from input-output (I/O) data only, without need of a model of the process. A drawback of this methods is that, in its standard f ...
2013The dilaton Wess-Zumino action in higher dimensions
We present a general formula for the Wess-Zumino action associated with the Weyl anomaly, given in a curved background for any even number of dimensions. The result is obtained by considering a finite Weyl transformation of counterterms in dimensional regu ...
Springer2013