-
Olivier Colliot7/8/24, 9:30 AM
Goal: Introduce the basics of ML and describe in details how to perform validation
- History and terminology
- Problem setup for ML basics (Model, loss, learning procedure, features)
- Generalization in ML (overfitting, underfitting and model selection)
- Validation (performance metrics, validation strategies, statistical analysis)
-
Guillaume Lemaitre7/8/24, 2:00 PM
Goal: Introduce the scikit-learn API, with a focus on practical insights on the model validation and selection.
Go to contribution page- Overview of a simple cross-validation scheme k-fold - Overview of metrics (Regression, Classification) - Model selection through SearchCV - Cross validation in complex settings (stratification, groups, non-iid data) -
Thomas Moreau (Inria)7/9/24, 9:30 AM
Goal: Introduce the different types of data, with a focus on time-series, and the different methodologies to apply on each type.
Go to contribution page- Overview of the different types of data: tabular data, time series, images, graph, signals. - Overview of the specific problems and jargon with time series and signals. - How to get back to a “classical” ML framework? - Practical illustrations... -
Romain Ménégaux7/9/24, 2:00 PM
Goal: Describe the main types of deep learning architectures, and apply them to a concrete example from life sciences.
Go to contribution page- Introduction: what is deep learning and why is everyone doing it? - Overview of the main types of deep learning architectures: MLP, convolutional, and transformers. When to use one or the other? - Overview of the different training and regularization... -
Filippo Vicentini (École Polytechnique - CPHT)7/10/24, 9:30 AM
-
Lucia Reining7/10/24, 10:15 AM
The understanding and prediction of properties of materials is a quantum many-body problem, and the observed phenomena often go well beyond the range that can be described with simple models. Recently, machine learning has emerged as a new tool that could potentially capture materials-specific or hidden universal features, and therefore help to analyse or design materials, and to improve...
Go to contribution page -
Marylou Gabrié (École Polytechnique)7/10/24, 11:30 AM
Deep generative models parametrize very flexible families of distributions able to fit complicated datasets of images or text. These models provide independent samples from complex high-distributions at negligible costs. On the other hand, sampling exactly a target distribution, such the Boltzmann distribution of a physical system, is typically challenging: either because of dimensionality,...
Go to contribution page -
Dr Danijela Markovic (CNRS Thales)7/10/24, 2:00 PM
Quantum computing aims to leverage the principles of quantum mechanics, such as superposition, to encode and process information in ways that classical computers cannot, potentially handling exponentially larger amounts of information. However, harnessing this computational advantage requires quantum algorithms capable of encoding data into superpositions and providing answers with minimal...
Go to contribution page -
Dr Bruno Loureiro (ENS Ulm)7/10/24, 2:45 PM
Measuring the uncertainty associated to a model's prediction is a central part of statistical practice. In the context of modern deep learning practice, several methods for quantifying the uncertainty of neural networks co-exist. Yet, theoretical guarantees for these methods are scarce in the theoretical literature. In this talk, I will discuss how some of them compare in a mathematically...
Go to contribution page -
Jorge FERNANDEZ-DE-COSSIO-DIAZ (ENS Paris)7/10/24, 4:00 PM
Riboswitches are structured allosteric RNA molecules that change conformation in response to a metabolite binding event, eventually triggering a regulatory response. Computational modelling of the structure of these molecules is complicated by a complex network of tertiary contacts, stabilized by the presence of their cognate metabolite. In this work, we focus on the aptamer domain of SAM-I...
Go to contribution page -
Matthieu Melennec7/10/24, 4:20 PM
Particle physics experiments like CMS (Compact Muon Solenoid) at the LHC and Super-Kamiokande let us probe the fundamental laws of physics by observing the interaction of high energy particles with various detectors. These particles leave their signatures in different sensors composing these detectors and a host of sophisticated algorithms are employed to reconstruct these particles by...
Go to contribution page -
Dr Rudy Morel (Flatiron Institute)7/10/24, 4:40 PM
Physicists routinely need probabilistic models for a number of tasks such as parameter inference or the generation of new realizations of a field. Establishing such models for highly non-Gaussian fields is a challenge, especially when the number of samples is limited. In this paper, we introduce scattering spectra models for stationary fields and we show that they provide accurate and robust...
Go to contribution page -
Riwal Plougonven7/11/24, 9:30 AMTalk (Long)
Climate models and Numerical Weather Prediction (NWP) Models describe the atmospheric circulation with a
Go to contribution page
limited resolution. There unavoidably remains processes that involve spatial scales shorter than the
grid scales, ie processes that are unresolved. Cloud processes, turbulence near the surface and internal
gravity waves propagating from lower to upper layers are among the main dynamical... -
Prof. Florence Tupin (Telecom Paris)7/11/24, 10:15 AM
In this talk I will first introduce the basics of radar imaging and present some applications
Go to contribution page
for climate science. I will then show how machine learning can make a key contribution
to improve radar data degraded by the speckle phenomenon and extract useful information.
I will focus on self-supervised methods allowing for exploiting a wide range of unlabeled data. -
Julien Le Sommer7/11/24, 11:30 AMTalk (Long)
-
Matthieu Blanke (Inria Paris, DI ENS)7/11/24, 12:15 PMTalk (Short)
Data assimilation is a central problem in many geophysical applications, such as weather forecasting. It aims to estimate the state of a potentially large system, such as the atmosphere, from sparse observations, supplemented by prior physical knowledge. The size of the systems involved and the complexity of the underlying physical equations make it a challenging task from a computational...
Go to contribution page -
Jerome Bobin7/11/24, 2:00 PMTalk (Long)
Inverse problems are ubiquitous in astrophysics, ranging from image reconstruction to unmixing or unsupervised com-
Go to contribution page
ponent separation, but they often share common challenges: i) how to deal with ill-posedness, which mandates the design of effective and physically relevant regularisation, ii) how to deal with the deluge of data coming from current and future experiments and iii) quantifying... -
Dr David Cornu (Observatoire de Paris)7/11/24, 2:45 PM
Large astronomical facilities generate an ever-increasing data volume, rapidly approaching the exascale, following the need for better resolution, better sensitivity, and larger wavelength coverage. Modern radio astronomy is strongly affected, especially regarding giant radio interferometers that produce large quantities of raw data. In particular, the forthcoming arrival of the SKA (Square...
Go to contribution page -
Damien Gratadour (Observatoire de Paris)7/11/24, 4:00 PM
The field of experimental astronomy is entering an exciting new era, with the emergence of extremely large telescopes, hosts to primary mirrors the size of several basketball courts. Among the many challenges associated with the construction and operations of such giant scientific infrastructures, the complexity of embedded computing facilities is notably heavy. In particular, the real-time...
Go to contribution page -
Dr Bertrand Thirion (Inria)7/12/24, 10:15 AM
Recent years have witnessed intense interactions between cognitive neuroscience and artificial intelligence, with the deep learning revolution driving new developments in neuroscience.
Go to contribution page
A first aspect concerns the processing of neuroscience data, which is often in the form of time courses. These data are often short and noisy, and suffer from poorly controlled confounding effects. AI-powered... -
Dr Christophe Pallier (EMR CNRS 9003 & INSERM-CEA Cognitive Neuroimaging Lab U992)7/12/24, 11:30 AM
Do representations proposed in linguistic theories, such as constituent trees, correspond to actual data structures constructed in real-time in the brain during language comprehension? And if so, what are the brain regions involved? This question was investigated in a series of functional magnetic resonance studies using various experimental paradigms, including repetition priming, syntactic...
Go to contribution page -
Fabian Gloeckle (Ecole des Ponts ParisTech)7/12/24, 12:15 PM
A machine readable and verifiable account of a large portion of human mathematics would change the way mathematicians can work, learn and collaborate. While impressive progress has been made in the mathematical standard libraries of proof assistants like Lean, Isabelle and Coq, the proportion of mathematical results formalized in such systems remains tiny overall. In the talk, I will argue...
Go to contribution page -
Dr Laura Cantini (Institut Pasteur)7/12/24, 2:00 PM
Single-cell data constitute a major breakthrough in life sciences. Their integration will enable us to investigate outstanding biological and medical questions thus far inaccessible. However, still few methods exist to integrate different single-cell modalities, corresponding to omics data (e.g. DNA methylation, proteome, chromatin accessibility), plus spatial positioning and images....
Go to contribution page -
Dr Judith Abecassis (Inria)7/12/24, 2:45 PM
The combination of artificial intelligence and the increasing digitization of the health sector opens up perspectives for using data for research and daily decision-making tools for patients and healthcare providers. However, the systematic deployment of these technologies requires better control of their performance, particularly in terms of generalization and explainability. These notions...
Go to contribution page -
Nona Naderi7/12/24, 4:00 PMTalk (Long)
-
Giulia Sambataro (École des Ponts ParisTech)Poster
In this work we adapt recent model reduction approaches to predict the solutions of
Go to contribution page
time-dependent parametrized problems describing crowd motion in the presence of ob-
stacles. The problem of interest is a discrete contact model, which is formulated as a
constrained least-squares optimization statement. The parametric variations in the prob-
lem (associated with the geometric configuration... -
Christoph Schönle (CMAP, Ecole Polytechnique)Poster
Generative models have started to integrate into the scientific computing toolkit. One notable instance of this integration is the utilization of normalizing flows (NF) in the development of sampling and variational inference algorithms. This work introduces a novel algorithm, GflowMC, which relies on a Metropolis-within-Gibbs framework within the latent space of NFs. This approach addresses...
Go to contribution page
Choose timezone
Your profile timezone: