{ Transfert technologique }
Un grand nombre d’avancées technologiques en apprentissage statistique sont de nature théorique. Epoch-intelligence se nourrit de ces développements pour produire de nouvelles techniques.
{ Des solutions faites maison }
Epoch-intelligence conçoit, élabore et produit ses propres solutions aux problèmes modernes de l’apprentissage statistique. Elles seront progressivement utilisées dans les expertises proposées aux entreprises ainsi que dans les formations.
Formations
Epoch-intelligence propose une offre de formations exigeante couvrant un large spectre de la science des données.
Nos formations vont du niveau débutant en R jusqu’aux dernières avancées en machine learning pour ingénieurs et chercheurs expérimentés.
en savoir plus
{ Un socle théorique solide }
La philosophie de nos formations est de transmettre une compréhension profonde des mécanismes utilisés dans les méthodes statistiques modernes afin de rendre les stagiaires autonomes dans leurs futurs projets.
{ Des applications adaptées }
Que vous soyez un industriel, un laboratoire ou une compagnie d’assurance, les applications sont adaptées au domaine spécifique de votre entreprise.
{ Venez avec vos données }
Une séance finale de 2h est prévue pour chaque formation afin que les stagiaires puissent appliquer les méthodes étudiées sur leurs propres données.
{ Une mise en pratique systématique }
Toutes les notions introduites sont mises en application via des travaux pratiques sur machine. Les stagiaires sont à la fois guidés et mis en autonomie afin de s’approprier pleinement les méthodes.
{ Une offre sur-mesure }
À votre demande, il est possible de créer une offre de formations entièrement personnalisée que nous élaborerons conjointement.
{ Nous nous rendons chez vous }
Que ce soit dans vos locaux ou ailleurs, les formateurs se déplacent pour vous former dans votre ville. Nous favorisons toujours le présentiel et le contact humain.
Machine Learning & Random Forests
21 heures 1200€ HT pour R{ Objectifs }
- Master the fundamental concepts of learning theory
- Understand and identify different types of learning
- Know how to measure the predictive performance of a model
- Understand the construction of a CART tree
- Know how to optimize a CART tree
- Understand the tools of ensemble methods
- Know how to build and optimize a random forest
- Know how to compare methods and select the most efficient one
- Understand the concept of variable importance
- Perform variable selection based on a specific objective
- Scientific Master’s degree or equivalent
- Advanced knowledge in mathematics and statistics
- Highly proficient in R programming
- Entirely in-person and in-company training
- Our trainers are all passionate PhD holders
- Alternation between theoretical presentations and practical work on real data
- Practice on numerous datasets
- Provision of a booklet and digital course materials with bibliographic references
- Course materials will be distributed at the end of the training
- Corrections for all exercises will be distributed at the end of the training
- Epoch-Intelligence is a Qualiopi-certified training organization.
- The quality certification has been awarded for the following action category: TRAINING ACTIONS.
- Our certificate can be viewed here.
Requirements
Teaching and Technical Resources
Qualiopi Certification
{ Programme }
- Learning theory
- Supervised vs unsupervised learning
- KNN algorithm
- Bayes predictor
- CART trees
- Node impurity and splitting criterion
- Pruning and overfitting
- Handling missing data
- Ensemble methods
- Bagging
- Random Forests (RF-RI)
- OOB prediction and OOB error
- Variable importance
- Classical variable selection
- VSURF method
- Boosting
- Request the full brochure at formations@epoch-intelligence.fr or via the contact form.
- Attendance sheet at the beginning of each session
- A positioning questionnaire will be conducted to verify prerequisites
- Assessment of concept understanding through solving open-ended problems
- Among the 28 participants, 100% reported being very satisfied with the training, 100% satisfied, and 97% ready to recommend it to a colleague.
- We encourage anyone with a disability to notify Epoch-Intelligence as early as possible at psh@epoch-intelligence.fr
- This training is accessible within 4 weeks from the initial client contact.
Monitoring and Evaluation of Learning
Satisfaction Rate
Accessibility
Access Time
Advanced Random Forests
21 heures 1200€ HT pour R{ Objectifs }
- Master the concepts of statistical learning
- Understand modern issues in statistical learning
- Master classical random forests
- Know how to measure the importance of explanatory variables
- Master unsupervised and semi-supervised random forests
- Estimate the conditional distribution in a learning problem using distributional random forests
- Understand the concept of weighting functions and combine them through resampling
- Build a causality model
- Handle missing data
- Detect outliers in a dataset
- Perform survival analysis
- Model longitudinal data
- Scientific Master’s degree or equivalent
- Advanced knowledge in mathematics and statistics
- Proficiency in R programming
- Entirely in-person and in-company training
- Our trainers are all passionate PhD holders
- Alternation between theoretical presentations and practical work on real data
- Provision of a booklet and digital course materials with bibliographic references
- Exercises will be corrected by the trainer in front of all participants at the end of each practical session
- Epoch-Intelligence is a Qualiopi-certified training organization.
- The quality certification has been awarded for the following action category: TRAINING ACTIONS.
- Our certificate can be viewed here.
Requirements
Teaching and Technical Resources
Qualiopi Certification
{ Programme }
- Recap on random forests
- Distributional random forests
- Learning the conditional distribution
- Splitting criteria and weighting function
- Quantile forests and prediction intervals
- Causality models
- Biostatistical models
- Repeated and longitudinal data
- Mixed models and random forests
- Survival random forests
- Extensions of random forests
- Handling missing data
- Random forests for high-performance computing
- Fréchet forests
- Outlier detection
- Request the full brochure at formations@epoch-intelligence.fr or via the contact form.
- Attendance sheet at the beginning of each session
- A positioning questionnaire will be conducted to verify prerequisites
- Assessment of concept understanding through solving open-ended problems
- Among the 8 participants, 75% reported being very satisfied with the training and felt the objectives had been achieved, 87.5% satisfied, and 63% ready to recommend it to a colleague.
- We encourage anyone with a disability to notify Epoch-Intelligence as early as possible at psh@epoch-intelligence.fr
- This training is accessible within 4 weeks from the initial client contact.
Monitoring and Evaluation of Learning
Satisfaction Rate
Accessibility
Access Time
Modern Artificial Intelligence
28 heures 1600€ HT pour R{ Objectifs }
- Understand the positioning of Artificial Intelligence (AI)
- Identify major types of algorithms and their origins
- Master optimization through Genetic Algorithms
- Know how to formulate a fitness function to address a problem
- Build a neural network
- Understand the concept of regularization in a network
- Be capable of optimizing a deep neural network using genetic algorithms
- Understand the concept of convolution
- Know how to manipulate images
- Build an image classifier
- Create an associative memory
- Know how to build a Kohonen self-organizing map
- Scientific Master’s degree or equivalent
- Advanced knowledge in mathematics and statistics
- Highly proficient in R programming
- Entirely in-person and in-company training
- Our trainers are all passionate PhD holders
- Alternation between theoretical presentations and practical work on real data
- Practice on numerous datasets
- Provision of a booklet and digital course materials with bibliographic references
- Course materials will be distributed at the end of the training
- Corrections for all exercises will be distributed at the end of the training
- Epoch-Intelligence is a Qualiopi-certified training organization.
- The quality certification has been awarded for the following action category: TRAINING ACTIONS.
- Our certificate can be viewed here.
Requirements
Teaching and Technical Resources
Qualiopi Certification
{ Programme }
- Concept of an intelligent agent
- Major challenges in AI
- Genetic algorithms
- Fitness function
- Selection, crossover, mutation, and replacement operators
- Neural networks
- Multi-layer perceptron
- Backpropagation algorithm
- Regularization and optimization
- Computer vision
- Convolution
- Convolutional neural networks
- Unsupervised networks
- Hopfield network and associative memory
- Hebb’s rule
- Dense associative memories
- Restricted Boltzmann machines
- Kohonen self-organizing networks
- Request the full brochure at formations@epoch-intelligence.fr or via the contact form.
- Attendance sheet at the beginning of each session
- A positioning questionnaire will be conducted to verify prerequisites
- Assessment of concept understanding through solving open-ended problems
- Among the 8 participants, 100% reported being very satisfied with the training and felt the objectives had been achieved, 100% satisfied, and 75% ready to recommend it to a colleague.
- We encourage anyone with a disability to notify Epoch-Intelligence as early as possible at psh@epoch-intelligence.fr
- This training is accessible within 4 weeks from the initial client contact.
Monitoring and Evaluation of Learning
Satisfaction Rate
Accessibility
Access Time
Basics of Programming with R
21 heures 990€ HT pour R{ Objectifs }
- Master the RStudio environment
- Know how to search for, install, and use packages
- Know how to use the help system
- Develop your own functions
- Use basic algorithmic tools
- Write a report with RMarkdown
- Import and manipulate a dataset
- Produce simple statistical descriptors
- Create a basic package
- Scientific bachelor’s degree or equivalent
- Basic knowledge of statistics
- Comfortable with computer tools
- Entirely in-person and in-company training
- Our trainers are all passionate PhD holders
- Alternation between theoretical presentations and practical work on real data
- Provision of numerous datasets
- Course materials will be distributed at the end of the training
- Corrections for all exercises will be distributed at the end of the training
- Epoch-Intelligence is a Qualiopi-certified training organization.
- The quality certification has been awarded for the following action category: TRAINING ACTIONS.
- Our certificate can be viewed here.
Requirements
Teaching and Technical Resources
Qualiopi Certification
{ Programme }
- Discovering the console
- Introduction to best practices
- Presentation of variable types
- Manipulation of algebraic operators
- Manipulation of vectors and matrices
- Managing loops and conditions
- Importing tables in .csv, .txt, and .tsv formats
- Managing data frames
- Exploring dataset characteristics
- Transforming a dataset
- Managing packages
- Producing simple graphs
- Manipulating and composing functions
- Solving and modeling simple mathematical problems
- Producing descriptive indicators on real-world data (medicine, ecology, marketing, etc.)
- Writing your first package
- Request the full brochure at formations@epoch-intelligence.fr or via the contact form.
- Attendance sheet at the beginning of each session
- A positioning questionnaire will be conducted to verify prerequisites
- Assessment of concept understanding through solving open-ended data analysis problems
- Of the 25 participants, 80% said they were very satisfied with the training and 92% considered that the objectives had been achieved, 100% satisfied, and 89% ready to recommend it to a colleague.
- We encourage anyone with a disability to notify Epoch-Intelligence as early as possible at psh@epoch-intelligence.fr
- This training is accessible within 4 weeks from the initial client contact.
Monitoring and Evaluation of Learning
Satisfaction Rate
Accessibility
Access Time
Basics in Statistics
21 heures 990€ HT pour R{ Objectifs }
- Understand the different types of variables
- Be able to create appropriate graphical representations
- Be capable of producing simple statistical descriptors
- Be able to identify a statistical problem
- Have basic knowledge of probability theory
- Be able to perform simple and multiple linear regression
- Know a wide range of statistical tests
- Be able to perform and interpret the results of a test
- Be able to perform variable selection
- Be capable of comparing models
- Be able to perform logistic regression
- Scientific bachelor’s degree or equivalent
- Basic knowledge of mathematics
- Basic knowledge of R
- Entirely in-person and in-company training
- Our trainers are all passionate PhD holders
- Alternation between theoretical presentations and practical work on real data
- Practice on numerous datasets
- Provision of a booklet and a digital course material with bibliographic references
- Course materials will be distributed at the end of the training
- Corrections for all exercises will be distributed at the end of the training
- Epoch-Intelligence is a Qualiopi-certified training organization.
- The quality certification has been awarded for the following action category: TRAINING ACTIONS.
- Our certificate can be viewed here.
Requirements
Teaching and Technical Resources
Qualiopi Certification
{ Programme }
- What is Statistics
- Probability Theory
- Random Variable and Distribution
- Expectation, Variance, and Standard Deviation
- Common Continuous Distributions
- Independence
- Univariate Descriptive Statistics
- Qualitative Variables and Continuous Quantitative Variables
- Measures of Central Tendency and Dispersion
- Bivariate Descriptive Statistics
- Frequency Table
- Cramér’s V
- Covariance and Pearson Correlation
- Test Theory
- Test Statistic, Rejection Region, and p-value
- Mean Testing
- Simple and Multivariate Linear Regression
- Least Squares Estimator
- Prediction and Model Quality
- Variable Selection
- Logistic Regression
- Logistic Function and Sigmoid Function
- Risk and Odds Ratio
- Extensions of the Linear Model
- Basis Functions, Polynomial Regression, and Cubic Splines
- Request the full brochure at formations@epoch-intelligence.fr or through the contact form.
- Attendance sheet at the beginning of each session
- A positioning questionnaire will be conducted to verify prerequisites
- Assessment of concept understanding through the resolution of open-ended problems
- Among the 16 participants, 77% reported being very satisfied with the training, 92% satisfied and felt the objectives had been achieved, and 89% ready to recommend it to a colleague.
- We encourage anyone with a disability to notify Epoch-Intelligence as early as possible at psh@epoch-intelligence.fr
- This training is accessible within 4 weeks from the initial client contact.
Monitoring and Evaluation of Learning
Satisfaction Rate
Accessibility
Access Time
Datamining
21 heures 1100€ HT pour R{ Objectifs }
- Understand the different methods of dimension reduction and how to apply them to R
- Select a dimension reduction method according to the problem
- Learn how to extract information from data
- Learn how to interpret results and graphical output in R
- Understand the difference between supervised and unsupervised classification
- Highlight groups of individuals and hierarchical relationships
- Learn how to combine dimension reduction and clustering
- Master’s degree in science or equivalent
- Advanced knowledge of mathematics and statistics
- Comfortable with R programming
- Entirely face-to-face and in-company training
- Our trainers are all passionate PhDs
- Alternating theoretical and practical presentations on real data
- Numerous data sets made available
- Course materials will be distributed at the end of the course
- A correction of all exercises will be distributed at the end of the course.
- Epoch-intelligence is a Qualiopi-certified training organization.
- Quality certification has been awarded for the following category of action : TRAINING ACTIONS.
- Our certificate can be viewed here.
Requirements
Teaching and technical resources
Qualiopi certification
{ Programme }
- Principal component analysis (PCA)
- Correspondence Analysis (CA)
- Multiple Correspondence Analysis (MCA)
- Multiple Factor Analysis (MFA)
- Presentation of classification problem
- Hierarchical clustering
- Expectation-Maximization algorithm for Gaussian mixtures
- Classification based on a metric approach (k-means, k-medoids,…)
- Principle of inertia and Huygens’ theorem
- Application on real data (biology, ecology, economics, . . . )
- Ask for the full leaflet at formations@epoch-intelligence.fr or via the contact form.
- Attendance sheet at the beginning of each session
- A positioning questionnaire will be carried out to check prerequisites
- Assessment of concept assimilation by solving open-ended problems
- We invite anyone with a disability to contact Epoch-intelligence at psh@epoch-intelligence.fr as soon as possible.
- This training course is available within 4 weeks of contact by the client.
Follow-up and evaluation
Accessibility
Availability
Advanced Statistics for High-Dimensional Data
28 heures 1500€ HT pour R{ Objectifs }
- Understand the challenge of high-dimensional data
- Identify approaches that fail in high dimensions
- Gain in-depth mastery of the linear model
- Build a regression model in high-dimensional settings
- Adapt a model to the nature of the data
- Perform model selection
- Construct and optimize a synthetic representation space
- Perform multiple test corrections
- Scientific Master’s degree or equivalent
- Advanced knowledge in mathematics and statistics
- Proficiency in R programming
- Entirely in-person and in-company training
- Our trainers are all passionate PhD holders
- Alternation between theoretical presentations and practical work on real data
- Provision of numerous datasets
- Course materials will be distributed at the end of the training
- Corrections for all exercises will be distributed at the end of the training
- Epoch-Intelligence is a Qualiopi-certified training organization.
- The quality certification has been awarded for the following action category: TRAINING ACTIONS.
- Our certificate can be viewed here.
Requirements
Teaching and Technical Resources
Qualiopi Certification
{ Programme }
- The paradigm of high-dimensional data
- Prediction challenges
- Generalization error and quadratic risk
- Sparse linear model
- Model selection
- The principle of penalization
- Ridge and Lasso regression
- Penalized logistic regression
- Risk and odds ratio
- Linear SVMs
- Kernel SVMs
- PLS regression and sparse PLS
- Low-rank matrix
- Singular value decomposition
- Robust PCA
- Graphical models
- Multiple test corrections
- Request the full brochure at formations@epoch-intelligence.fr or via the contact form.
- Attendance sheet at the beginning of each session
- A positioning questionnaire will be conducted to verify prerequisites
- Assessment of concept understanding through solving open-ended problems
- Among the 15 participants, 100% reported being very satisfied with the training, 100% satisfied, and 89% ready to recommend it to a colleague.
- We encourage anyone with a disability to notify Epoch-Intelligence as early as possible at psh@epoch-intelligence.fr
- This training is accessible within 4 weeks from the initial client contact.
Monitoring and Evaluation of Learning
Satisfaction Rate
Accessibility
Access Time
{ Valoriser vos données }
Nous vous aidons à extraire, visualiser et exploiter les informations présentes dans vos données.
{ Veille technologique }
La recherche scientifique a son propre langage et ses propres codes. Economisez du temps : epoch-intelligence se charge d’effectuer la recherche bibliographique en amont du lancement de vos projets.
{ Évaluation des algorithmes }
Epoch-intelligence propose d’analyser et d’optimiser vos algorithmes afin de vous faire gagner en performance.
{ Prédire pour mieux décider }
Nous vous aidons dans la conception de modèles mathématiques prédictifs adaptés à vos objectifs. Ces modèles constitueront un outil d’aide à la décision opérationnelle dans votre activité.
Qui sommes-nous ?
Marine, PhD
Fondatrice
Intérêts scientifiques :
{ Intérêts scientifiques }
- Théorie des tests
- Analyse différentielle
- Génomique
Théorie des tests, Analyse différentielle, Génomique
Louis, PhD
Fondateur
Intérêts scientifiques :
{ Intérêts scientifiques }
- Théorie de l'apprentissage statistique
- Données complexes
- Grande dimension
Théorie de l’apprentissage statistique, Données complexes, Grande dimension
Contactez-nous !