Publications related to Probability distribution fitting

Exploiting the Signal-Leak Bias in Diffusion Models

Sabine Süsstrunk, Radhakrishna Achanta, Mahmut Sami Arpa, Martin Nicolas Everaert, Athanasios Fitsios

There is a bias in the inference pipeline of most diffusion models. This bias arises from a signal leak whose distribution deviates from the noise distribution, creating a discrepancy between training and inference processes. We demonstrate that this signa ...

2024

On the use of Cramér-Rao Lower Bound for least-variance circuit parameters identification of Li-ion cells

Mario Paolone, Vladimir Sovljanski

Electrochemical Impedance Spectroscopy (EIS) and Equivalent Circuit Models (ECMs) are widely used to characterize the impedance and estimate parameters of electrochemical systems such as batteries. We use a generic ECM with ten parameters grouped to model ...

2024

On distributional autoregression and iterated transportation

Victor Panaretos, Laya Ghodrati

We consider the problem of defining and fitting models of autoregressive time series of probability distributions on a compact interval of Double-struck capital R. An order-1 autoregressive model in this context is to be understood as a Markov chain, where ...

Wiley2024

Federated Learning under Covariate Shifts with Generalization Guarantees

Volkan Cevher, Grigorios Chrysos, Fanghui Liu, Thomas Michaelsen Pethick

This paper addresses intra-client and inter-client covariate shifts in federated learning (FL) with a focus on the overall generalization performance. To handle covariate shifts, we formulate a new global model training paradigm and propose Federated Impor ...

2023

Personalized Privacy-Preserving Distributed Learning on Heterogeneous Data

Michael Christoph Gastpar, Aditya Pradeep

One major challenge in distributed learning is to efficiently learn for each client when the data across clients is heterogeneous or non iid (not independent or identically distributed). This provides a significant challenge as the data of the other client ...

2023

Bringing casting to microfabrication: processing and deformation of microcastings

Luciano Borasi

This work extends the range of pathways for the production of metallic microcomponents by downscaling metal casting. This is accomplished by using either of two different molding techniques, namely femtosecond laser micromachining or lithographic silicon m ...

EPFL2023

Towards a multiscale point cloud structural similarity metric

Touradj Ebrahimi, Davi Nachtigall Lazzarotto

Point clouds are effective data structures for the rep- resentation of three-dimensional media and hence adopted in a wide range of practical applications. In many cases, the portrayed data is expected to be visualized by humans. After acquisition, point c ...

IEEE2023

Hierarchical processing underpins competition in tactile perceptual bistability

Andrea Ferrario

Ambiguous sensory information can lead to spontaneous alternations between perceptual states, recently shown to extend to tactile perception. The authors recently proposed a simplified form of tactile rivalry which evokes two competing percepts for a fixed ...

SPRINGER2023

Scalable and Privacy-Preserving Federated Principal Component Analysis

Jean-Pierre Hubaux, Juan Ramón Troncoso-Pastoriza, Apostolos Pyrgelis, David Jules Froelicher, Joao André Gomes de Sá e Sousa, Jean-Philippe Léonard Bossuat

Principal component analysis (PCA) is an essential algorithm for dimensionality reduction in many data science domains. We address the problem of performing a federated PCA on private data distributed among multiple data providers while ensuring data confi ...

IEEE COMPUTER SOC2023

Near-Minimax Optimal Estimation With Shallow ReLU Neural Networks

Rahul Parhi

We study the problem of estimating an unknown function from noisy data using shallow ReLU neural networks. The estimators we study minimize the sum of squared data-fitting errors plus a regularization term proportional to the squared Euclidean norm of the ...

IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC2023

Benign Overfitting in Deep Neural Networks under Lazy Training

Volkan Cevher, Grigorios Chrysos, Fanghui Liu, Zhenyu Zhu

This paper focuses on over-parameterized deep neural networks (DNNs) with ReLU activation functions and proves that when the data distribution is well-separated, DNNs can achieve Bayesoptimal test error for classification while obtaining (nearly) zero-trai ...

2023

System Support for Robust Distributed Learning

Arsany Hany Abdelmessih Guirguis

Machine learning (ML) applications are ubiquitous. They run in different environments such as datacenters, the cloud, and even on edge devices. Despite where they run, distributing ML training seems the only way to attain scalable, high-quality learning. B ...

EPFL2022

Score Matching Enables Causal Discovery of Nonlinear Additive Noise Models

Volkan Cevher, Paul Thierry Yves Rolland

This paper demonstrates how to recover causal graphs from the score of the data distribution in non-linear additive (Gaussian) noise models. Using score matching algorithms as a building block, we show how to design a new generation of scalable causal disc ...

JMLR-JOURNAL MACHINE LEARNING RESEARCH2022

Dalton: Learned Partitioning for Distributed Data Streams

Anastasia Ailamaki, Eleni Zapridou, Ioannis Mytilinis

To sustain the input rate of high-throughput streams, modern stream processing systems rely on parallel execution. However, skewed data yield imbalanced load assignments and create stragglers that hinder scalability. Deciding on a static partitioning for a ...

2022

Failure and success of the spectral bias prediction for Laplace Kernel Ridge Regression: the case of low-dimensional data

Matthieu Wyart, Umberto Maria Tomasini, Antonio Sclocchi

Recently, several theories including the replica method made predictions for the generalization error of Kernel Ridge Regression. In some regimes, they predict that the method has a 'spectral bias': decomposing the true function f* on the eigenbasis of the ...

JMLR-JOURNAL MACHINE LEARNING RESEARCH2022

Investigation and Validation of Unstructured Mesh Methodologies for Modeling Experimental Reactors

Andreas Pautz, Mathieu Hursin, Carlo Fiorina, Tom Mager

This paper summarizes a methodology developed at Ecole Polytechnique Federale de Lausanne for the neutronic modeling of the CROCUS experimental reactor and proposes solutions to the challenges one may face while modeling a research reactor with a complex g ...

MDPI2022

Data-based model maintenance in the era of industry 4.0: A methodology

Dimitrios Kyritsis, Antoine Pélissier, Foivos Psarommatis Giannakopoulos

Despite the high number of investments for data-based models in the expansion of Industry 4.0, too little effort has been made to ensure the maintenance of those models. In a data-streaming environment, data-based models are subject to concept drifts. A co ...

ELSEVIER SCI LTD2022

Considerations on the partial safety factor format for reinforced concrete structures accounting for multiple failure modes

Aurelio Muttoni, Miguel Fernández Ruiz, Qianhui Yu

The increasing usage of nonlinear analyses for the design of reinforced concrete structures and the necessity of codes of practice to provide a consistent safety format for them is one of the challenges that new generations of codes of practice are facing. ...

ELSEVIER SCI LTD2022

On the Use of Benford’s Law to Assess the Quality of the Data Provided by Lightning Locating Systems

Farhad Rachidi-Haeri, Marcos Rubinstein, Amirhossein Mostajabi, Wolfgang Schulz

Lightning causes significant damage and casualties globally by directly striking humans and livestock, by igniting forest fires, and by inducing electrical surges in electronic infrastructure, airplanes, rockets, etc. Monitoring the evolution of thundersto ...

2022

Deep Domain Adaptation in Earth Observation

Devis Tuia, Benjamin Alexander Kellenberger, Nicolas Courty

When applied to new datasets, acquired at different time moments, with different sensors or under different acquisition conditions, deep learning models might fail spectacularly. This is because they have learned from the data distribution observed during ...

Wiley2021