pyMEF: a Python library for mixtures of exponential families
pyMEF is a Python framework allowing to manipulate, learn, simplify and compare mixtures of exponential families. It is designed to ease the use of various exponential families in mixture models.
See also jMEF for a Java implementation of the same kind of library and libmef for a faster C implementation.
What are exponential families?
An exponential family is a generic set of probability distributions that admit the following canonical distribution:
Exponential families are characterized by the log normalizer function F, and include the following well-known distributions: Gaussian (generic, isotropic Gaussian, diagonal Gaussian, rectified Gaussian or Wald distributions, lognormal), Poisson, Bernoulli, binomial, multinomial, Laplacian, Gamma (incl. chi-squared), Beta, exponential, Wishart, Dirichlet, Rayleigh, probability simplex, negative binomial distribution, Weibull, von Mises, Pareto distributions, skew logistic, etc.
Mixtures of exponential families provide a generic framework for handling Gaussian mixture models (GMMs also called MoGs for mixture of Gaussians), mixture of Poisson distributions, and Laplacian mixture models as well.
More pyMEF specific tutorials are available here:
Basic manipulation of mixture models
- Olivier Schwander, Frank Nielsen, Simplification de modèles de mélange issus d’estimateur par noyau, GRETSI 2011
- Olivier Schwander and Frank Nielsen, pyMEF - A framework for Exponential Families in Python, in Proceedings of the 2011 IEEE Workshop on Statistical Signal Processing
- Vincent Garcia, Frank Nielsen, and Richard Nock, Levels of details for Gaussian mixture models, in Proceedings of the Asian Conference on Computer Vision, Xi’an, China, September 2009
- Frank Nielsen and Vincent Garcia, Statistical exponential families: A digest with flash cards, arXiV, http://arxiv.org/abs/0911.4863, November 2009
- Frank Nielsen and Richard Nock, Sided and symmetrized Bregman centroids, in IEEE Transactions on Information Theory, 2009, 55, 2048-2059
- Frank Nielsen, Jean-Daniel Boissonnat and Richard Nock, On Bregman Voronoi diagrams, in ACM-SIAM Symposium on Data Mining, 2007, 746-755
- A. Banerjee, S. Merugu, I. Dhillon, and J. Ghosh, Clustering with Bregman divergences, in Journal of Machine Learning Research, 2005, 6, 234-245
“Intelligence is the faculty of manufacturing artificial objects, especially tools to make tools, and of indefinitely varying the manufacture.” Henri Bergson
GSI forge presents and lists packages and softwares usually opensource (Python and associated Github depositories, R and associated CRAN-R depositories) that can be useful in the statistical and informational analysis of data with a geometrical or topological approach.
Venus at the Forge of Vulcan, Le Nain Brothers, Musée Saint-Denis, Reims (Vulcan is the god of fire and god of metalworking and the forge, often depicted with a blacksmith’s hammer)
Cartan's father Joseph (1837-1917) was born in the village of Saint Victor de Morestel, which is 13 kilometers from Dolomieu. After he married Anne Cottaz (1841-1927) the family settled in Dolomieu, where Anne had lived. Joseph Cartan was the village blacksmith. Elie Cartan recalled that his childhood had passed under "blows of the anvil, which started every morning from dawn", and that "his mother, during those rare minutes when she was free from taking care of the children and the house, was working with a spinning-wheel".
Diffeomorphic Demons is an efficient algorithm for the diffeomorphic registration of N dimensional images. It is based on Thirion’s demons algorithm but works on a Lie group structure on diffeomorphic transformations. Typical 3D medical images can be registered in less than three minutes on a 2 x 2.8 GHz quad-core Intel Xeon Apple Mac pro computer. Diffeomorphic demons is now included in MedINRIA‘s image fusion module. The source code has been integrated into ITK since version 3.8. A command-line software can be found on the Insight Journal. Additionally, some standalone binaries and tutorials may be found on the Stark Lab website.
The Computational Geometry Algorithms Library
CGAL (Computational Geometry Algorithms Library) is a comprehensive library of geometric algorithms. The goal of CGAL is to advance the state of the art of geometric computing and to offer robust and efficient programs for research purpose and industrial applications. The initial development of CGAL is a joint effort of six groups in Europe partially funded by European Projects. The library consists of about 1,000,000 lines of C++ code with users all over the world. Since november 2003, CGAL is an Open Source Project. The spin-off Geometry Factory sells CGAL commercial licenses, support for CGAL and customized developments based on CGAL.
The library offers data structures and algorithms like triangulations, Voronoi diagrams, Boolean operations on polygons and polyhedra, point set processing, arrangements of curves, surface and volume mesh generation, geometry processing, alpha shapes, convex hull algorithms, shape analysis, AABB and KD trees...
Learn more about CGAL by browsing through the Package Overview.
GUDHI – Geometry Understanding in Higher Dimensions
The GUDHI library is a generic open source C++ library, with a Python interface, for Topological Data Analysis (TDA) and Higher Dimensional Geometry Understanding. The library offers state-of-the-art data structures and algorithms to construct simplicial complexes and compute persistent homology.
The library comes with data sets, demos, examples and test suites.
The INFOTOPO library is a generic open source suite of Python Programs (compatible with Python 3.4.x, on Linux, windows, or mac) for Information Topological Data Analysis. It is distrubuted freely under opensource GNU GPL V3 Licence and available on Github depository. The library offers state-of-the-art statistical high dimensional data structures analysis and algorithms to detect covarying patterns and clusters, multiscale data analysis.
INFOTOPO version 1.2
It computes all multivariate information functions: entropy, joint entropy between k random variables (Hk), mutual informations between k random variables (Ik), conditional entropies and mutual informations and provides their cohomological (and homotopy) visualisation in the form of information landscapes and information paths together with an approximation of the minimum information energy complex . It is applicable on any set of empirical data that is data with several trials-repetitions-essays (parameter m), and also allows to compute the undersampling regime, the degree k above which the sample size m is to small to provide good estimations of the information functions . The computational exploration is restricted to the simplicial sublattice of random variable (all the subsets of k=n random variables) and has hence a complexity in O(2^n). In this simplicial setting we can exhaustively estimate information functions on the simplicial information structure, that is joint-entropy Hk and mutual-informations Ik at all degrees k=<n and for every k-tuple, with a standard commercial personal computer (a laptop with processor Intel Core i7-4910MQ CPU @ 2.90GHz * up to k=n=21 in reasonable time (about 3 hours). The mathematical formalism can be found in [1,2,3,6], and its application as a neuroscience and data analysis method can be found in [1,4,5,6].
Baudot, Tapia, Goaillard, Topological Information Data Analysis: Poincare-Shannon Machine and Statistical Physic of Finite Heterogeneous Systems. PDF
 M. Tapia, P. Baudot, M. Dufour, C. Formisano-Tréziny, S. Temporal, M. Lasserre, J. Gabert, K. Kobayashi, JM. Goaillard . Information topology of gene expression profile in dopaminergic neurons PDF
 Baudot P., Bennequin D., The homological nature of entropy. Entropy, 2015, 17, 1-66; doi:10.3390. PDF
 Categories and Physics 2011. Classic and quantum Information topos.
 Information Topology: Statistical Physic of Complex Systems and Data Analysis -Topological and geometrical structures of information, CIRM LuminyFrance. 27-1 sept VIDEO-SLIDE
The INFOTOPO library is developed as part of the Channelomics project supported by the European Research Council, developped at UNIS Inserm 1072, and thanks previously to supports and hostings since 2007 of Max Planck Institute for Mathematic in the Sciences (MPI-MIS) and Complex System Instititute Paris-Ile-de-France (ISC-PIF) and Institut de Mathématiques de Jussieu - Paris Rive Gauche (IMJ-PRG)
A Matlab toolbox for optimization on manifolds
Optimization on manifolds is a powerful paradigm to address nonlinear optimization problems. With Manopt, it is easy to deal with various types of symmetries and constraints which arise naturally in applications, such as orthonormality and low rank.
Manifolds are mathematical sets with a smooth geometry, such as spheres. If you are facing a nonlinear (and possibly nonconvex) optimization problem with nice-looking constraints, symmetries or invariance properties, Manopt may just be the tool for you. Check out the manifolds library to find out!
Manopt comes with a large library of manifolds and ready-to-use Riemannian optimization algorithms. It is well documented and includes diagnostics tools to help you get started quickly. It provides flexibility in describing your cost function and incorporates an optional caching system for more efficiency.
It's open source
Check out the license and let us know how you use Manopt. Please cite this paper if you publish work using Manopt (bibtex).
Python package that performs computations on manifolds such as hyperspheres, hyperbolic spaces, spaces of symmetric positive definite matrices and Lie groups of transformations.
- Nina Miolane, Johan Mathe, Claire Donnat, Mikael Jorda, Xavier Pennec, geomstats: a Python Package for Riemannian Geometry in Machine Learning 2018 PDF
- Benjamin Hou , Nina Miolane, Bishesh Khanal, Matthew C.H. Lee , Amir Alansary,
Steven McDonagh, Joseph Hajnal, Daniel Rueckert, Ben Glocker, Bernhard Kainz, Deep Pose Estimation for Image-Based Registration PDF
TUMCREATE is a leading research institute set up by the Technical University of Munich, Germany in collaboration with the Singapore Government. TUMCREATE has received funding and support for the SPEEDCARGO project from the Civil Aviation Authority of Singapore (CAAS) & the National Research Foundation (NRF) to develop automation solutions that will transform the air freight logistics sector. The SPEEDCARGO solution is the world's first AI-powered robotic solution for automatic build-up and break down of aviation cargo pallets and will help Singapore lead the transformation of the logistics industry globally. The project is seeking technical experts with a passion for creating world class products, and a willingness to work in a fast paced, quality obsessed, multi-cultural global environment. On successful completion, the project will spin-off as a start-up with members of the project team having the option to join the start-up with benefits that include attractive ESOPs. Apply now if you are interested in working on cutting edge technologies, changing the world with your work and joining a dynamic start-up team.
For Open Position in Robotics, Vision, Mathematical Optimisation,
Mechatronics ... Please have a look ate the job posting under
For more info about the project please look at
For any another question please write to me at
suraj.nair [at] tum-create.edu.sg
This website is the home for the IHP semester “The Mathematics of Imaging” that will take place between January to April 2019.
You can also access the monthly seminar website from here.
Registration is free but mandatory to be able to participate to one (or to all!) of these conferences and to the pre-school.
All the conferences will take place at IHP.
- Variational methods and optimization in imaging, February 4th-8th 2019.
- Statistical Modeling for Shapes and Imaging, March 11th-15th 2019.
- Imaging and machine learning, April 1st-5th 2019.
January 7-11th 2019, CIRM pre-school for PhD students and postdocs.
Monthly seminar on imaging sciences
More information here.
High-school and general audience activity
- Jean-François Aujol (Bordeaux).
- Julie Delon (Paris 5)
- Agnès Desolneux (CNRS and ENS Cachan)
- Jalal Fadili (ENSICAEN)
- Bruno Galerne (Paris 5)
- Gabriel Peyré (CNRS and ENS)
- Coloma Ballester (Pompeu Fabra Univ., Spain)
- Andrea Bertozzi (UCLA, USA)
- Laure Blanc-Feraud (CNRS, Nice Sophia Antipolis Univ., France)
- Donald Geman (Johns Hopkins Univ., USA)
- Stephane Mallat (ENS Ulm, France)
- Simon Masnou (Univ. Lyon 1, France)
- Jean-Michel Morel (ENS Cachan, France)
- David Mumford (Brown University, Providence,USA)
- Mila Nikolova (CNRS, ENS Cachan, France)
- Joachim Weickert (Saarland Univ., Germany)
Institution: Faculty of Mathematics and Computer Science
Contact: Mr Prof. Dr. Dieter Hogrefe
Allocation: earliest possible
Publication: 28 May 2018
The Faculty of Mathematics and Computer Science at the Georg-August-University of Göttingen invites applications for the positions of four professorships at the Institute of Computer Science in the context of the establishment of a Campus Institute Data Science.
The newly established professors will participate in the development of the Campus Institute Data Science (CIDAS). CIDAS will be a joint institution of the members of the Göttingen Campus including the University, the University Medical Center, five Max-Planck-Institutes, the Leibniz Institute for Primate Research, the Academy of Sciences and the German Aerospace Center Göttingen. The members of CIDAS will pursue research at the convergence point of computer science, statistics, mathematics and application areas and will combine method development in the field of Data Science with cutting edge research in Göttingen’s research foci.
The future position holder shall have worked already successfully in research networks. We especially expect candidates to be willing to cooperate with colleagues in the Faculty of Mathematics and Computer Science as well as from other Faculties and Campus institutions. The candidates’ research interests should ideally align with one or several of Göttingen University’s research foci (https://www.uni-goettingen.de/en/505395.html).
The appointed professors will conduct their teaching in the study programmes of the Faculty, in particular in the Bachelor’s programmes ”Applied Data Science” and ”Mathematical Data Science” and in the Bachelor and Master programme ”Applied Computer Science”.
Conditions for appointment can be found in Section 25 of the Higher Education Law of Lower Saxony. As a Public Law Foundation, the University of Göttingen has the right to appoint professors. Further details can be obtained upon request.
Two positions of full professor (W3) for Data Science
The future position holder shall have expertise in computer science, especially related to the field of Data Science method development in one or several of the following areas: Computer Vision and Imaging, Data Fusion / Data Merging, Data Management, Big Data, Graph and Network Optimization and Modeling, Natural Language Processing. Candidates shall have a track record of high level research in these fields, have high potential for groundbreaking research in internationally visible research projects along with profound experience in mentoring and advising university students. Göttingen University is committed to research oriented teaching.
Full professor (W3) for Artificial Intelligence / Machine Learning
The future position holder shall have expertise in artificial intelligence, especially with a focus on machine learning or on foundations of machine learning. The University of Göttingen aims to establish machine learning as research area by recruitments in its fields of application as well as in theoretical foundations. Candidates shall have a track record of high level research in artificial intelligence, in particular machine learning, have high potential for groundbreaking research in internationally visible research projects along with profound experience in mentoring and advising university students. Göttingen University is committed to research oriented teaching.
Associate Professor (Tenure Track) (W2 t.t. W3) for Artificial Intelligence / Machine Learning
This professorship is funded by the German Federal and State governments through the programme “Bund-Länder-Programm zur Förderung des wissenschaftlichen Nachwuchses (Nachwuchspakt)”. The initial appointment will be limited to five years. The transition to the permanent position of Full Professor (W3) takes effect following a successful evaluation.
Preferably, the candidate has expertise in artificial intelligence, especially with a focus on machine learning or on foundations of machine learning. The University of Göttingen aims to establish machine learning as research area by recruitments in its fields of application as well as in theoretical foundations. The future position holder shall have a track record of high level research in artificial intelligence, in particular machine learning, have high potential for groundbreaking research in internationally visible research projects along with profound experience in mentoring and advising university students. Göttingen University is committed to research oriented teaching.
Applications of international scientists are highly encouraged. The University is an equal opportunity and family friendly employer. We specifically encourage women to apply. Disabled persons with appropriate qualifications and aptitude for the position will be given special consideration.
For further information please contact Prof. Dr. Dieter Hogrefe, Institute of Computer Science, hogrefe [at] informatik.uni-goettingen.de.
Applications including CV, publications list and details of teaching and research including funded research projects should be submitted by July 6th, 2018, online via the application portals stated above.
This international workshop is organized in the honor and in the presence of Dominique Jeulin, after he retired from École des Mines in 2016. The workshop aims to promote ideas and establish connections between researchers working in the wide field of mechanics and physics of heterogeneous media. The main themes of the workshop cover the theoretical and numerical modeling of microstructures and of their electrical, mechanical and transport properties, in solid mechanics and material science. Sessions will emphasize all main topics beloved by Dominique, ranging from probabilistic models, multiscale structures, microstructure evolution, image analysis for materials, homogenization, stochastic analysis, fracture and rupture processes, transport properties of nano-structures, localization in non-periodic media, percolation theory, fuel cell technology and the problem of triple percolation, electric & magnetic properties, optical properties, computational methods.
Preliminary programme (PDF)
François Willot and Samuel Forest (Mines ParisTech). To contact us, write to francois.willot [at] ensmp.fr.
Participants will have the opportunity to submit a paper for publication in a special issue of a Internationl Journal of Solids and Structures or Image Analysis and Stereology (submission due on November 1, 2018). Please let us know (francois.willot [at] ensmp.fr) if you plan to submit an article to this journal.
On this special occasion, a book at Presses de l'école des Mines will be published, with special articles that are of a scientific or “sentimental” nature. You are welcome to contribute to the book. The format is free, but a latex tample is available here.
The workshop will be held at the CNRS village La Vieille Perrotine, 140, route des Allards, 17310 Saint Pierre d'Oléron, in the Île d'Oléron, France (Google map link).
Oléron, the home of Pierre Loti, is a 30 km-wide island (nowadays linked to the coast of France by a bridge), made of small scattered villages, beaches and pinèdes (pine forests), famous for its palourdes (local clams). Further information is available here.
A shuttle (KEOLIS company) will link bring particpants from both the train station and airport of La Rochelle to Oléron on Sunday 17, late afternoon (train coming from Paris-Montparnasse, arriving at 16:16 and plane coming from Paris-Orly, arriving at 16:45). There will also be a shuttle for La Rochelle on June 22, after lunch.
To reach La Rochelle, you can take the train from e.g. Paris Gare Montparnasse. The two main Paris airports are Charles de Gaulle (CDG) and Orly (ORY). RER line B (from CDG) or Orlybus (from Orly) brings you to Denfert-Rochereau station, and metro line 6 from Denfert-Rochereau to Montparnasse train station in Paris.
At the moment (May 23), French rail workers continue to announce a strike for June 17 and June 22. If you have difficulties in finding how to come to Oléron or La Rochelle, or you plan to come by car and may be able to take participants with you, please let us know (francois.willot [at] ensmp.fr).
Participants to the workshop will be accommodated inside the CAES (simple CNRS village with single rooms). All meals will be provided inside the village. The 90 acres village includes two ponds, a forest area and offers a direct access to the seashore. It is located in the protected area of the marsh of Moëze, near Fort Royer and Fort Boyard oyster farms. See this link for more information.
A banquet will be held on thursday (June 21, 2018).
It is not possible to regiser anymore.
February 15, 2018: deadline for abstract submission (talk or poster).
March 13, 2018: notification of acceptance.
June 17-22, 2018: workshop.
I am hiring assistant of Bosch Center of Artificial Intelligence. We are now hiring research scientist and research engineer. I have registered the robotics worldwide mailing list but I see from the website that in order to send messages to all mailing list memebers, I should write email to you, right?
Below is the information that we would like to share, could you please help us to send this email in the mailing list?
Email Title: WELCOME TO BOSCH! We are looking for DL/ML/RL Research Scientists for diverse functions
Welcome! Join us to start shaping the future of real-world Artificial Intelligence!
The Bosch Center for Artificial Intelligence was founded in early 2017 to deploy cutting-edge AI technologies across Bosch products and services creating solutions that are "Invented for life." Currently over 100 employees are working for the Bosch Center for Artificial Intelligence (BCAI) in Renningen, Sunny Vale and Bangalore. The aim is strengthen Boschs position in the field of artificial intelligence. Bosch will invest around 300 million euros in the BCAI by 2021. At the Bosch Center for Artificial Intelligence, you will have the opportunity to work with large-scale data sets across domains, AI along maturity scales, multi-disciplinary and dynamic teams international teams.
We are looking for:
Deep Learning Researcher Generative Models
Deep Learning Researcher for Uncertainties in Behavior Models
Deep Learning Researcher Uncertainty
Reinforcement learning researcher for autonomous driving
Research Scientist in Probabilistic Modeling
Research Scientist for Reinforcement Learning
Reinforcement Learning under Constraints Researcher
Deep Learning Researcher Explainable AI
Research Engineer for Reinforcement Learning
Research Scientist Knowledge Representation/Logic Reasoning
Research Scientist Mathematical Optimization
Deep Learning Researcher for Autonomous Driving
Reinforcement Learning Research Engineer
Hierarchical Reinforcement Learning Researcher
Imitation Learning Researcher for Autonomous Driving
Machine Learning Research Engineer
Research Scientist Reinforcement Learning for Autonomous Systems & Robotics
Probabilistic graphical modeling researcher for autonomous driving
Research Scientist for Decision Making, Learning and Control for Robotic Sensor Networks
Learn more about BCAI:
Check out our worldwide job vacancies
Mit freundlichen Grüßen / Best regards
Please feel free to contact me if you have any further questions.
Thank you in advance.
Mit freundlichen Grüßen / Best regards
Two PhD positions in Deep Learning at the Idiap Research Institute, Switzerland, starting as early as possible but can be delayed if necessary.
** Please apply on-line at http://www.idiap.ch/~fleuret/apply **
The Machine Learning group at the Idiap Research Institute, affiliated with École Polytechnique Fédérale de Lausanne, seeks two PhD students in deep learning and computer vision to work on real-time semantic segmentation.
The candidates will be doctoral students at EPFL[4,5], working under the supervision of Dr. François Fleuret at the Idiap Research Institute.
Applicants must be self-sufficient programmers and have a strong background in mathematics. They should be interested in, and familiar with, applied probabilities, information theory, signal processing, optimization, algorithmic, and development in deep-learning frameworks.
The Idiap Research Institute is located in Valais, a scenic region in the south of Switzerland, surrounded by the highest mountains of Europe, and within close proximity to Lausanne and Geneva. The working language of Idiap is English.
IM-Simons postdoctoral Fellowship
The Institute of Mathematics of the Vietnam Academy of Science and Technology (Hanoi) invites applications for 4 positions of the IM-Simons postdoctoral Fellowship Program, 2018-2019. The initial appointment will be for one year, with possibility of extension up to a second year. Renewal for the second year will depend on a comprehensive review of the scientific activity of the fellow.
The targets of this program are:
· to attract foreign young researchers to work at the institute,
· to provide Vietnamese young researchers a bridge to a long-term scientific career.
Applications are invited from qualified researchers under 40 years of age who have submitted the doctoral thesis or have a PhD degree in mathematics no more than five years before the deadline of the application. Preferences will be given to research areas of the institute, see the website http://math.ac.vn/en/.
Salary and benefits
The salary will be 1.000 USD per month and a subsidy of 2.000 USD per year for accommodation. In addition, the fellows will receive a round flight ticket and a research grant up to 1.000 USD per year for attending conferences and small equipments.
How to Apply
Interested candidates should submit the following documents to im_simons [at] math.ac.vn:
· Curriculum Vitae;
· List of publications;
· A research proposal (no more than four pages);
· Two recommendation letters.
The submission deadline is July 15, 2018. Applications will be considered until the positions are filled. The expected starting date of the fellowship is September 1, 2018. Informal queries should be sent to the above email address.
The Universities of Pavia and Milano-Bicocca have a joint PhD program in Mathematics, with the contribution of INdAM:
The call for admissions 2018 is open (deadline June 7, 2018). A total of 13 fellowship are available in all domains of Mathematics, including Probability and Statistics. For more information, see the official pages:
(english version - details for Mathematics at page 26 of the Call)
(versione italiana - scheda per Matematica a pag. 30 del Bando)
Bernoulli Lecture - What is Probability?
27 March 2018 - CIB - EPFL - Switzerland
17:15 - 18:15
Room : BCH 2201
Mikhail Gromov, Université Paris-Saclay
The success of probability theory in mathematics and in theoretical physics is due not so much to its measure-theoretic foundation, but rather it is because it exploits and enhances the symmetries of the structures it applies to. We shall describe in this lecture two alternative approaches to the concept of probability, where
the first one is motivated by the ongoing revision of the set-theoretic language in mathematics, as it is being systematically superseded by the category-theoretic one;
the second approach is motivated by the needs of biology and linguistics, where the structures do not possess symmetries of the kind physical structures enjoy.