• Geo-Sci-Info

    Academic Fellowship positions

    The University of Nottingham is a research intensive university ranked in the top 10 in the UK and in the top 30 of Europe. It has a strong tradition in Cognition, Perception, Mathematical Biology, and structural and functional Imaging. To support further development across these areas, and stimulate new research opportunities, the University has recently made a major investment in Computational Neuroscience.

    As part of this investment we are inviting applications for three independent fellowships (3 years). The positions are ideal for researchers with a strong vision that are planning to develop their own research group in a supportive, interdisciplinary environment with outstanding research facilities.

    We are interested in candidates working across a broad spectrum of research topics areas in Computational Neuroscience, including functional neuroimaging, neural networks, and models of cognition. Research methodologies could include numerical simulation, machine learning, theoretical neuroscience, and innovative data analysis. The Fellows will be based in the School of Psychology and/or Mathematical Sciences.

    For the online application form see
    https://www.nottingham.ac.uk/jobs/currentvacancies/ref/SCI219418

    Informal inquiries:
    Prof Mark van Rossum, mark.vanrossum [at] nottingham.ac.uk, or
    Prof Mark Humphries, lpzmdh [at] exmail.nottingham.ac.uk.

    posted in Jobs offers - Call for projects read more
  • Geo-Sci-Info

    Capture du 2018-07-25 06-24-56.png

    OFFICIAL WEBPAGE

    DOWNLOAD PROGRAM

    Conférence-débat de l'Académie des sciences, de 14h30 à 16h45, dans la Grande salle des séances de l’Institut de France - Inscription obligatoire avant le 9 octobre 2018

    Organisateur
    Alain Berthoz, professeur honoraire au Collège de France, membre de l'Académie des sciences et de l’Académie des technologies.

    Intervenants

    • Neil Burgess, professeur et directeur de l’Institute of Cognitive Neuroscience, University College London, membre de la Royal Society et de l’Academy of Medical Sciences
    • Laure Rondi-Reig, directrice de recherche au CNRS à l’Institut de Biologie Paris–Seine, Université Pierre-et-Marie-Curie
    • Alain Berthoz, professeur honoraire au Collège de France, membre de l'Académie des sciences et de l’Académie des technologies
    • Daniel Bennequin, professeur émérite au département de mathématiques de l’université Paris Diderot.

    Résumé
    Le thème général de cette séance sera d’exposer des découvertes et théories récentes concernant le traitement cérébral de l’espace et plus particulièrement la flexibilité des stratégies cognitives et la diversité des réseaux de traitement des divers espaces d’action.
    Neil Burgess décrira les divers types de neurones de la formation hippocampique qui codent la place, la direction, les bords, leur organisation en grilles et les relations avec les objets. Il décrira un modèle mathématique de transformation entre les divers référentiels spatiaux (égocentré, allocentré etc.) et son utilisation pour prédire des symptômes de déficits de la mémoire spatiale dans des pathologies post-traumatiques.
    Laure Rondi-Reig décrira les réseaux neuronaux impliqués dans des activités de navigation spatiale, l’apprentissage et les troubles de la mémoire, dans le cas du vieillissement par exemple. Des données expérimentales sur les mécanismes neurophysiologiques et la diversité des réseaux corticaux et sous corticaux de la cognition spatiale et des modèles inspirés des neurosciences computationnelles seront présentés.
    Alain Berthoz proposera que le cerveau traite avec des réseaux différents et des géométries différentes pour les divers espaces (corps, préhension, locomoteur, environnemental). Il décrira des données récentes sur les bases neurales des changements de perspective, le développement chez l’enfant des fonctions visuo-spatiales et l’intervention de la manipulation des référentiels spatiaux dans l’empathie et son implication en pathologie psychiatrique.
    Daniel Bennequin proposera que pour guider l’adaptation des actions et des perceptions, les cerveaux des animaux et de l’homme mettent en place une variété de géométries, Euclidiennes et non Euclidiennes, et de dynamiques. L’exposé présentera un travail sur une nouvelle sorte de géométrie : un topos d’espaces au-dessus d’une catégorie (site) représentant la préparation et l’exécution d’une classe de mouvements; par exemple la préhension, la locomotion, la navigation, l’imagination.

    posted in Brain and Spaces - conference - Academie des Sciences read more
  • Geo-Sci-Info

    Capture du 2018-07-16 10-08-30.png

    OFFICIAL WEBSITE

    This ICML workshop seeks to broaden the role of geometric models and thinking in machine-learning research.

    The interplay between machine learning and geometry is an active field of research drawing the attention of researchers from many fields as it offers not only beautiful mathematical and statistical theory but also substantial impact on important real-world problems in machine learning. Examples of this interplay include, but are not limited to:

    • geometric insights into deep learning (helping us to better understand and improve these models);
    • statistical models that respect and exploit the constraints of geometric data ("statistics on manifolds");
    • (Riemannian) manifold learning;
    • viewing probability distributions as points on a nonlinear manifold ("information geometry");
    • optimization over nonlinear manifolds;
    • Wasserstein spaces and their applications;
    • understanding the effect of encoding group invariances (e.g. rotation invariance) on the intrinsic geometry of the data space, which becomes a quotient space.

    This ICML workshop is co-located with ICML, IJCAI and AAMAS in Stockholm.

    The Bosch center for AI generously sponsors a prize for the best abstract, which will be presented at the workshop.

    Confirmed Speakers

    Venue

    The workshop takes place at the ICML conference venue, room A4.

    Organizers

    The conference is jointly organized by:

    For matters regarding the conference, you can contact the organizers at gimli.meeting[at]gmail.com.

    We are grateful for funding from the Bosch Center for Artificial Intelligence, the Villum Fonden Young Investigator program and the European Research Council (ERC) through a starting grant.

    posted in Geometry in Machine Learning - GiMLi workshop - ICML conference read more
  • Geo-Sci-Info

    Postdoc and PhD student positions in machine learning, Finnish Center for Artificial Intelligence

    Finnish Center for Artificial Intelligence (FCAI; http://fcai.fi) is looking for exceptional doctoral students and postdoctoral researchers to tackle complex and exciting problems in the field of machine learning. Come and join us to create the Real AI that is data-efficient, trustworthy and understandable!
    FCAI brings together the world-class expertise of Aalto University and the University of Helsinki in AI research, strengthened further with an extensive set of companies and public sector partners, creating an attractive, world-class ICT hub in Helsinki metropolitan area.
    Our research agenda is spearheaded by 5 research programs with multiple research groups involved in each. We are currently hiring doctoral students and postdoctoral researchers in the following FCAI research programs and the detailed projects listed below.

    Research programs:

    1. Agile probabilistic AI. Keywords: Probabilistic programming; Robust and automated Bayesian machine learning
    2. Simulator-based inference: Approximate Bayesian Computation ABC; likelihood-free inference; Generative adversarial networks (GAN); applications in many fields including medicine, materials design, visualization, business, …
    3. Next generation data-efficient deep learning; including deep reinforcement learning
    4. Privacy-preserving and secure AI: Privacy-preserving machine learning; differential privacy; adversarial machine learning
    5. Interactive AI: Interactive machine learning; probabilistic inference of cognitive models from data; probabilistic programming for behavioral sciences

    Specific projects:

    1. Constraint-Based Optimization and Machine Learning
    2. Probabilistic Machine Learning
    3. Probabilistic machine learning for personalized medicine
    4. Probabilistic modeling and machine learning for bioinformatics
    5. Non-parametric probabilistic machine learning
    6. Bioinformatics and computational biology
    7. Computational HCI
    8. Privacy-preserving federated machine learning
    9. Probabilistic user modelling in interactive human-in-the-loop machine learning

    Deadline for doctoral students: August 12, 2018. Postdoc applications received by August 12, 2018 will receive full treatment.

    More details here: http://www.fcai.fi/fcainews/2018/7/2/postdoc-and-doctoral-student-positions-in-machine-learning

    posted in Jobs offers - Call for projects read more
  • Geo-Sci-Info

    1] Research Associate /Senior Research Associate positions, University of Bristol, UK
    [2] Postdoc position in Computational Neuroscience at the CRM, Barcelona, Spain
    [3] 1 research associate and 2 funded PhD positions on the evolution of neural learning and plasticity, Loughborough University, UK
    [4] Two PhD Student Positions in Computer Vision for Learnings Systems at MPI-IS, Tuebingen, Germany
    [5] Postdoctoral Scholars in Computational Brain Science – Brown Institute for Brain Sciences, Providence, RI
    [6] Postdoctoral Research Associate in Neural Engineering, University of Essex, UK
    PhD and Research Fellows Positions in Machine Learning / Deep Learning, Hochreiter Group, Linz - Austria
    [7] 10 PhD Positions in Machine Learning / Deep Learning, Hochreiter Group, Linz, Austria
    [8] Five Research Fellows in Machine Learning / Deep Learning, Hochreiter Group, Linz, Austria
    [9] Two Research Fellows in Machine Learning / Deep Learning, Hochreiter Group, Linz, Austria
    [10] Post doc position in comparative cognition (4-yrs), Vienna, Austria
    [11] PhD call for applications PROPOSAL OF SUBJECT OF THESIS - New Stochastic Optimization Approaches to the Aerospace Vehicle Spatial Design Problem


    [6] De: Luca Citi <lciti[at]essex.ac.uk>

    Postdoctoral Research Associate in Neural Engineering
    University of Essex - School of Computer Science and Electronic Engineering

    The Essex Brain-Computer Interfaces and Neural Engineering laboratory is happy to announce a postdoctoral position in the MURI project "Closed-Loop Multisensory Brain-Computer Interface for Enhanced Decision Accuracy" (more information at: https://sites.usc.edu/muri-project/ ).
    The project is a very exciting one and we have teamed up with outstanding partners from USC, Harvard, UCL, Berkeley, Imperial, among others.

    The Essex team's work on the project focuses on brain-computer interfacing, on algorithms for signal processing and extraction of information from EEG and other physiological signals, on behavioural and neuro-physiological investigations of multisensory feature binding and integration, as well as methods for predicting the level of attention and confidence in decision making of a participant from behavioural, physiological and neural data in real time.
    The duties of the role include conducting research, development and dissemination of neural engineering techniques within the MURI project.

    Applicants are expected to hold a PhD (or be very close to submitting their PhD thesis) in Biomedical Engineering, Brain-computer Interfaces, Neural Engineering, Electronic Engineering, Statistics, Physics, Computer Science or a closely related discipline. The ideal candidate will have significant experience in signal processing, statistical modelling of neural signals and processes. Applicants are also expected to have a strong publication record (relative to their career stage) as first author, ideally including publications in 1st quartile journals in relevant areas.

    The successful applicant will join the Essex team - formed by Dr Luca Citi (PI), Prof Riccardo Poli (Co-I and UK team leader), Dr Caterina Cinel (named Research Fellow) - and will be part of the Essex BCI-NE Lab, today the UK's largest research group in brain-computer interfaces.

    This post is initially fixed-term until the 31st of October 2019 but may be extended for two more years if further funding is approved.

    Appointment will be made as Senior Research Officer.

    Closes: 30th June 2018
    Job Ref: REQ01468
    Salary: £32,548 - £34,521 per annum

    Further information and application instructions:
    https://www.jobs.ac.uk/job/BKF396


    [7] De : Ulrich Bodenhofer <bodenhofer[at]bioinf.jku.at>

    10 PhD Positions in Machine Learning / Deep Learning, Hochreiter Group, Linz, Austria

    Johannes Kepler University Linz (JKU), Austria, is looking for research assistants and junior researchers in the area of machine learning and deep learning with Sepp Hochreiter. These fully-funded positions will be affiliated with the LIT AI Lab and the Institute for Machine Learning andCandidates are expected to enroll into JKU’s PhD programme, and have to option to complete their PhD under Sepp Hochreiter’s supervision.

    Job description:
    • conduct research in machine learning / deep learning with the aim to obtain a PhD within four years,
    • publish in renowned international journals and conferences,
    • work in research projects at the LIT AI Lab or in collaboration with partners.

    Requirements:
    • MSc degree or equivalent,
    • strong interest and previous education in machine learning (e.g. deep learning, reinforcement learning, kernel methods, probabilistic modeling, meta-learning, attention models); a track record in the field is a plus (e.g. MSc thesis, publications, project experience),
    • knowledge in one or more of the following application domains is a plus: signal processing, vision, speech, natural language processing, physics, bio-/chemoinformatics, computational medicine, autonomous driving.

    About the group: Within the last years, Sepp Hochreiter (who is best known for the invention of LSTM, for the Vanishing Gradient problem, “Flat Minima”, and “Learning to Learn”) has built up a dynamic team of more than20 researchers. The group has recently achieved widely acclaimed contributions and successes, such as, winning the NIH’s Tox21 toxicity prediction challenge with deep learning, the invention of the ELU activation function, Self-Normalizing Networks, and providing a convergence proof for GAN learning. The group has many international collaborations and receives funding from national and international research programs (like the EU) as well as from companies, such as, Johnson&Johnson, Merck, Bayer, Zalando and from joined labs like the Audi.JKU Deep Learning Center.

    About the location: The area offers an excellent quality of living in the heart of Europe – close to the alps between Vienna, Salzburg, Prague and Munich. Linz provides a superb cultural environment, most famous for the Ars Electronica Festival, the Brucknerfest, and the nearby Salzburg Festival. The picturesque and versatile landscape provides countless options for recreation and sports in nature (skiing, hiking, climbing, cycling, and many more).

    If you have questions, please contact: Prof. Dr. Sepp Hochreiter, +43 732 2468 4520, recruitment[at]bioinf.jku.at.


    [8] De : Ulrich Bodenhofer <bodenhofer[at]bioinf.jku.at>

    Five Research Fellows in Machine Learning / Deep Learning, Hochreiter Group, Linz, Austria

    Johannes Kepler University Linz (JKU), Austria, is looking for five post-doctoral research fellows to advance machine learning and deep learning in close collaboration with Sepp Hochreiter. These positions are affiliated with the LIT AI Lab and the Institute for Machine Learning.

    Job description:
    • conduct independent research in the field,
    • collaborate in machine learning and deep learning projects,
    • publish in renowned international journals and conferences.

    Requirements:
    • PhD degree,
    • track record in machine learning (e.g. deep learning, reinforcement learning, kernel methods, probabilistic modeling, meta-learning, attention models),
    • knowledge in one or more of the following application domains is a plus: signal processing, vision, speech, natural language processing, physics, bio-/chemoinformatics, computational medicine, autonomous driving,
    • willingness and ability to work in a team.

    About the group: Within the last years, Sepp Hochreiter (who is best known for the invention of LSTM, for the Vanishing Gradient problem, “Flat Minima”, and “Learning to Learn”) has built up a dynamic team of more than20 researchers. The group has recently achieved widely acclaimed contributions and successes, such as, winning the NIH’s Tox21 toxicity prediction challenge with deep learning, the invention of the ELU activation function, Self-Normalizing Networks, and a convergence proof for GAN learning. The group has many international collaborations and receives funding from national and international research programsas well as from companies, such as, Johnson&Johnson, Merck, Bayer, Zalando and from joined labs like the Audi.JKU Deep Learning Center.

    About the location: The area offers an excellent quality of living in the heart of Europe – close to the alps between Vienna, Salzburg, Prague and Munich. Linz provides a superb cultural environment, most famous for the Ars Electronica Festival, the Brucknerfest, and the nearby Salzburg Festival. The picturesque and versatile landscape provides countless options for recreation and sports in nature (skiing, hiking, climbing, cycling, and many more).

    If you have questions, please contact: Prof. Dr. Sepp Hochreiter, +43 732 2468 4520, recruitment[at]bioinf.jku.at.


    [9] De : Ulrich Bodenhofer <bodenhofer[at]bioinf.jku.at>

    Two Research Fellows in Machine Learning / Deep Learning, Hochreiter Group, Linz, Austria

    Johannes Kepler University Linz (JKU), Austria, is looking for two post-doctoral research fellows to advance machine learning and deep learning research with Sepp Hochreiter. These six year positions are affiliated both with the newly established LIT AI Lab and the Institute for Machine Learning.

    Job description:
    • conduct independent research in the field,
    • collaborate in machine learning and deep learning projects,
    • publish in renowned international journals and conferences,
    • supervise students; prepare and hold lectures; support study programs.

    Requirements:
    • PhD degree,
    • track record in machine learning (e.g. deep learning, reinforcement learning, kernel methods, probabilistic modeling, meta-learning, attention models),
    • knowledge in one or more of the following application domains is a plus: signal processing, vision, speech, natural language processing, physics, bio-/chemoinformatics, computational medicine, autonomous driving,
    • willingness and ability to work in a team and to support students and junior researchers.

    About the group: Within the last years, Sepp Hochreiter (who is best known for the invention of LSTM, for the Vanishing Gradient problem, “Flat Minima”, and “Learning to Learn”) has built up a dynamic team of more than 20 researchers. The group has recently achieved widely acclaimed contributions and successes, such as, winning the NIH’s Tox21 toxicity prediction challenge with deep learning, the invention of the ELU activation function, Self-Normalizing Networks, and providing a convergence proof for GAN learning. The group has many international collaborations and receives funding from national and international research programs as well as from companies, such as, Johnson&Johnson, Merck, Bayer, Zalando and from joined labs like the Audi.JKU Deep Learning Center.

    About the location: The area offers an excellent quality of living in the heart of Europe – close to the alps between Vienna, Salzburg, Prague and Munich. Linz provides a superb cultural environment, most famous for the Ars Electronica Festival, the Brucknerfest, and the nearby Salzburg Festival. The picturesque and versatile landscape provides countless options for recreation and sports in nature (skiing, hiking, climbing, cycling, and many more).

    If you have questions, please contact: Prof. Dr. Sepp Hochreiter, +43 732 2468 4520, recruitment[at]bioinf.jku.at.

    Prospective applicants interested in these positions are requested to electronically send an application via the online portal http://jku.at/application. Please include “Job Reference Number 3619” (deadline: July 30, 2018) or “Job Reference Number 3578” (deadline: July 4, 2018).


    [10] De: Isabelle Charrier <isabelle.charrier[at]u-psud.fr>

    Post doc position in comparative cognition (4-yrs), Vienna, Austria

    At the Unit of Comparative Cognition, Messerli Research Institute, Vienna, Austria, we are seeking a postdoctoral researcher who is eager to investigate cognitive and emotional processes in non-human animals, especially dogs. At the Clever Dog Lab we are committed to researching the behavioural, physiological (including neuronal) and genetic underpinnings of dog cognition. The successful candidate will have the opportunity to develop her/his own research agenda, using a large repertoire of state-of-the-art techniques and methodologies (including fMRI, eye-tracking, touch screens, automatized video analysis and behaviour annotation) and benefiting from administrative and technical support from members of the unit (including a lab manager, mechanical and electronic technicians and IT personnel).

    Please see the full advertisement here:
    http://www.vetmeduni.ac.at/fileadmin/v/z/mitteilungsblatt/stellen/2017_2018/20180615_Postdoc_Comparative_Cognition.pdf

    More details about the unit of Comparative Cognition, Vienna:http://www.vetmeduni.ac.at/en/messerli/science/cognition

    Informal enquiries about this post may be directed to Professor Ludwig Huber, ludwig.huber[at]vetmeduni.ac.at

    Application deadline: 8th July 2018.

    --
    Prof. Ludwig Huber, PhD
    Head of Comparative Cognition
    Messerli Research Institute

    University of Veterinary Medicine, Vienna (Vetmeduni Vienna)
    Veterinaerplatz 1, 1210 Vienna, Austria
    T +43 1 25077-2680
    M +43 664 60257-6250
    ludwig.huber[at]vetmeduni.ac.at
    www.vetmeduni.ac.at/messerli

    Partner institutions of the Messerli Research Institute:
    Messerli-Foundation, University of Veterinary Medicine Vienna, Medical University of Vienna, University of Vienna


    [11] De : Rachid Chelouah <rc[at]eisti.eu>

    PhD call for applications PROPOSAL OF SUBJECT OF THESIS - Reference: TIS-DTIS-2018-15

    Title: New Stochastic Optimization Approaches to the Aerospace Vehicle Spatial Design Problem

    The host laboratory at ONERA : Domain : TIS
    Department: Information Processing and Systems Department - Unit: Design and Evaluation of Aerospace Vehicles
    Location (ONERA center): Palaiseau

    ONERA responsibles: Tel. 01 80 38 66 30 Email :
    Romain Wuilbercq romain.wuilbercq[at]onera.fr
    Karim Dahia karim.dahia[at]onera.fr
    Arnault Tremolet arnault.tremolet[at]onera.fr

    The host laboratory at Paris-Seine : Quartz
    Thesis supervisor: Rachid Chelouah Email : rachid.chelouah[at]eisti.eu Tel. :+33 (0)1 34 25 84 20

    Co-encadrant: Stefan Borhofen Email : stefan.bornhofen[at]eisti.eu
    Address: University of Paris-Seine/EISTI, 95011 Cergy-Pontoise Cedex

    Description of the subject
    The pre-project phase for the development of an aerospace vehicle is one that is likely to bring out innovative configurations. In a general way, after a parallel evaluation of several topologies, a particular configuration is chosen on the basis of selection criteria resulting from the optimization of a mission. At present no systematization of the process seems to be possible without having a function of planning in the sense of geometric placement of objects. The problem of the geometric arrangement is defined by the ability to place different objects without interpenetration in an envelope. The generalization of the problem consists in taking into account their functional aspect which can contribute to prohibit or force their relative placement.

    The work focuses on objects and an external topology of 3D vehicle. To solve the problem of geometrical arrangement, two large families stand out, one of excluding all solutions with interpenetrations which supposes the entire evaluated, one speaks of legal placement, the other by authorizing them but by affecting a penalty function according to the degree of interpenetration, this is called relaxed placement. The latter approach, adopted as part of the proposed work, was exploited in [1] highlighting the interest of the coupling between robust multi-objective optimization techniques and a separation method.
    The first part of the work consists in appropriating methods of modeling and rapid evaluation of collisions between objects. Given the chosen relaxed placement, an assessment of this collision should be made by estimating the interpenetration depth, for example. The physical aspects (eg electromagnetic field, radiative radiation) and functional aspects can also be introduced by the notion of region of influence [2]. The consideration of spatial collisions and regions of influence in a planning process will be explored during this thesis.

    Such a problem of placement of objects is characterized by a strong combinatorics. The result is the need to explore a vast space of solutions. The number of variables associated with the quantity of objects to be placed is also important. Finally, the integration of constraints, to be defined exhaustively, is the heart of the problem because of their diverse nature (geometric, functional, thermal, ... etc). Stochastic optimization algorithms therefore appear relevant in order to explore a vast space of solutions, containing many constraints and many variables (complete Np problem) [3, 4, 5, 6, 7, 8, 9, 10, 11 , 12, 13]. The challenge of this thesis will be to adapt optimization algorithms to the specificity of the problem by introducing for example an algorithmic overlay so as to deal with all constraints.

    The developments of this thesis will be carried out in a capitalization environment such as the ACADIA platform developed at ONERA. In particular, they will be applied to multidisciplinary optimization problems using the OpenMDAO framework (Python, Cython, C ++). The methods and algorithms developed will be applied to various aerospace applications or road and rail transport that will verify their genericity.

    References :
    [1] G. Jacquenot, « Méthode générique pour l’optimisation d’agencement géométrique et fonctionnel », 18 Janvier 2010.
    [2] Joaquim P. L. Viegas, Susana M. Vieira, Joao M. C. Sousa and, Elsa M. P. Henriques, « Metaheuristics for the 3D Bin Packing Problem in HAPE3D the Steel Industry », 2014 IEEE Congress on Evolutionary Computation (CEC) July 6-11, 2014, Beijing, China.
    [3] Xiao LIU, Jia-min LIU, An-xi CAO, Zhuang-le YAO, « - a new constructive algorithm for the 3D irregular packing problem », Front Inform Technol Electron Eng 16(5):380-390, 2015
    [4] Marouene Kefi, Paul Richard, Thuong Hoang, Takehiko Yamaguchi and Vincent Barichard, « Using Constraint Solver for 3D Layout Assistance in Human-scale Virtual Environment », HUCAPP 2017 - International Conference on Human Computer Interaction Theory and Applications, 2017.
    [5] Giorgio Fasano, « A global optimization point of view to handle non-standard object packing problems », J Glob Optim, 2012.
    [6] C. Monjaret, « Introduction aux méthodes d’optimisation pour l’aménagement spatial », RT 1/14706 DPRS, 2010.
    [7] C. Leboucher, R. Chelouah, H.S. Shin, S. Le Ménec, P. Siarry, A. Tsourdos and A. Kotenkoff, An Enhanced Particle Swarm Optimisation Method Integrated With Evolutionary Game Theory, IEEE Transactions on Computational Intelligence and AI in Games, Janv 2018
    [8] Peio Loubiere, Astrid Jourdan, Patrick Siarry and Rachid Chelouah, A modified sensitivity analysis method for driving a multidimensional search in the Artificial Bee Colony algorithm, IEEE Congress on Evolutionary Computation, IEEE CEC 2016,
    [9] C. Leboucher, H.S. Shin, P. Siarry, S. Le Ménec, R. Chelouah, and A. Tsourdos, Convergence Proof of an Enhanced Particle Swarm Optimisation Method Integrated with Evolutionary Game Theory, Information Sciences, ScienceDirect, Elsevier, DOI doi:10.1016/j.ins.2016.01.011, pp. 389-411 2016:
    [10] James Kennedy, Rachid Chelouah, Maurice Clerc et Patrick Siarry, Swarm Intelligence Research édité en deux tomes par IGI Publishing ISSN 1947-9263 et ISSN 1947-9271, IJSIR, 2012.

    Profile of candidate du (de la) candidat (e) :
    Education: Student in Master 2 or engineering school
    Desired Specificities: Mathematical Modeling, Stochastic Optimization, Artificial Intelligence, Computer Science.

    Person to contact :
    Rachid Chelouah
    Laboratoire Quartz
    Université Paris-Seine / EISTI
    Mail : rc[at]eisti.eu

    Karim Dahia
    ONERA
    Laboratoire : Conception et Évaluation de Véhicules Aérospatiaux
    mail : karim.dahia[at]onera.fr


    posted in Jobs offers - Call for projects read more
  • Geo-Sci-Info

    PhD psotitions : 6 offers

    [1] Research Associate /Senior Research Associate positions, University of Bristol, UK
    [2] Postdoc position in Computational Neuroscience at the CRM, Barcelona, Spain
    [3] 1 research associate and 2 funded PhD positions on the evolution of neural learning and plasticity, Loughborough University, UK
    [4] Two PhD Student Positions in Computer Vision for Learnings Systems at MPI-IS, Tuebingen, Germany
    [5] Postdoctoral Scholars in Computational Brain Science – Brown Institute for Brain Sciences, Providence, RI
    [6] Postdoctoral Research Associate in Neural Engineering, University of Essex, UK
    PhD and Research Fellows Positions in Machine Learning / Deep Learning, Hochreiter Group, Linz - Austria
    [7] 10 PhD Positions in Machine Learning / Deep Learning, Hochreiter Group, Linz, Austria
    [8] Five Research Fellows in Machine Learning / Deep Learning, Hochreiter Group, Linz, Austria
    [9] Two Research Fellows in Machine Learning / Deep Learning, Hochreiter Group, Linz, Austria
    [10] Post doc position in comparative cognition (4-yrs), Vienna, Austria
    [11] PhD call for applications PROPOSAL OF SUBJECT OF THESIS - New Stochastic Optimization Approaches to the Aerospace Vehicle Spatial Design Problem


    [1] De : Jeffrey Bowers <J.Bowers[at]bristol.ac.uk>

    Research Associate /Senior Research Associate positions, University of Bristol
    Salary: £32,548 to £41,212 per annum
    Deadline: 20th August 2018

    We are seeking to appoint talented computational modelers to join an ERC-funded project entitled “Generalisation in Mind and Machine”.  The project explores how well neural networks support human-like generalisation across a range of tasks.  Two postdoctoral fellows and two PhD students are working on project, and we are looking to hire two more persons with PhDs (or potentially BSc) in computer science or cognitive psychology (or a related field) with experience working with neural networks.

    In contrast with the large community of researchers focused on improving the performance of deep networks for applied reasons, the goal of this project is to explore what neural networks tell us about how the brain works.  More specifically, the team will be working on questions of generalisation in the domains of object identification, word identification, short-term memory, and games, amongst other areas.  A core issue is whether networks need to be include symbolic computations to succeed on some forms of generalisation. For overview of some of the issues we are exploring in the project see: Bowers, J. S. (2017). Parallel Distributed Processing Theory in the Age of Deep Networks. Trends in Cognitive Sciences, 12, 950-961.  For more details regarding the project see our ERC website:http://mindandmachine.blogs.bristol.ac.uk/

    This post is available from Oct 1st 2018 (but somewhat flexible about start date) and currently has funding secured until Oct 1st 2021.  To apply go to: https://www.jobs.ac.uk/job/BKR397/erc-research-associate-senior-research-associate/  If you have any questions, please get in touch with Jeff Bowers at: j.bowers[at]bristol.ac.uk

    Jeff Bowers


    [2]  De : Klaus Wimmer <kwimmer[at]crm.cat>

    A fully funded POSTDOC POSITION in computational neuroscience is available in the lab of Klaus Wimmer, Computational Neuroscience Group at the CRM, Barcelona, Spain.

    Profile
    We are looking for an enthusiastic and scientifically curious researcher with a strong interest in computational neuroscience. The perfect candidate has a strong mathematical, physical or engineering background, scientific programming skills (Matlab, Python), and a keen interest in biological neural systems. Knowledge in computational neuroscience, dynamical systems, machine learning or advanced statistics is a plus. Good team spirit is a must.

    Research topic
    The neural basis of decision making and working memory has been studied extensively, yet our understanding of how distributed circuits in the brain perform these cognitive functions is only at the beginning. Models of cortical circuits can shed light on the underlying neural network dynamics. The Postdoc will work on building such models and on cutting-edge analysis of large-scale neural activity recordings (neuronal population recordings, fMRI, EEG).

    The lab
    The Computational Neuroscience Group is based at the Centre de Recerca Matemàtica at the campus of the Universitat Autònoma de Barcelona. It is a joint effort of Alex Roxin and Klaus Wimmer, and forms part of a larger network of theoretical and systems neuroscience labs in Barcelona. The successful candidate will benefit from a vibrant and stimulating research community and will have the opportunity of enjoying a lively city.

    More information can be found at: https://sites.google.com/view/wimmerlab

    How to apply
    Interested candidates should e-mail their application as a single pdf file to Klaus Wimmer, kwimmer[at]crm.cat, with the subject “Postdoc 2018”. The application should include: (1) CV with publication list, (2) a brief description of research experience and interests, (3) contact information for two references.
    The position is available immediately and applications will be accepted until it is filled. Informal inquiries are welcome.

    -- 
    Klaus Wimmer
    Ramón y Cajal researcher at
    Centre de Recerca Matemàtica
    Campus de Bellaterra, Edifici C
    08193 Bellaterra (Barcelona)
    Spain
    Tel. +34 935 86 85 15
    https://sites.google.com/view/wimmerlab


    [3] De : Andrea Soltoggio <A.Soltoggio[at]lboro.ac.uk>

    1 research associate and 2 funded PhD. postions on the evolution of neural learning and plasticity

    One research associate and two funded Ph.D. positions are available at the Computer Science Department, School of Science, Loughborough University, UK, on the topics of the evolution of lifelong learning in neural networks.

    Research.  The aim is to develop new neuroevolution algorithms for lifelong learning. The objectives are to devise machine learning systems that autonomously adapt to changing conditions such as variation of the data distribution, variation of the problem domain or parameters, with minimal human intervention. The approach will use neuroevolution, neuromodulation, and other methodologies to continuously discover and update learning strategies, implement selective plasticity, and achieve continual learning.
    For an overview of the research direction, see the paper: Born to Learn: the Inspiration, Progress and Future of Evolved Plastic Artificial Neural Networkshttps://www.researchgate.net/publication/315710249_Born_to_Learn_the_Inspiration_Progress_and_Future_of_Evolved_Plastic_Artificial_Neural_Networks
    Application areas include a variety of automation and machine learning problems, e.g. vision, control, and robotics, with a particular focus on resilience and autonomy.

    Working environment. The research associate and Ph.D. students, based at the Computer Science Department, will work in an international team with opportunities for collaboration and travel. They will have access to a number of robotic platforms such as mobile and flying robots, manufacturing robots, High Performance Computing clusters, and GPU computing. The Computer Science Department and robotics laboratories have ongoing collaborations with large industries and programs to promote start-ups. 
    Loughborough University is ranked 7th in the UK in the 2019 League Table Ranking http://www.thecompleteuniversityguide.co.uk/loughborough/performance ), and is located in Loughborough, a town well connected to London by a 1h20m journey by train.

    Requirements.
    Postdoc: A Ph.D. in Computer Science or related with a strong publication record, coding abilities, predisposition to work in a team and independence, passion for science, solid work ethics.  
    Ph.D. students: The ideal candidate holds (or is about to obtain) a first-class honour undergraduate/postgraduate degree (or equivalent) in Computer Science, Mathematics, Statistics, Electrical or Electronic Engineering, or has authored publications in recognised conferences/journals. Independent working skills are valued as well as the capability of working in a team. Collegiality and interpersonal skills are essential. 
    Excellent English language skills are also essential (see requirements herehttp://www.lboro.ac.uk/international/englang/index.htm)

    Period and salaries. 
    Postdoc position: until June 2020 (with possible extension) with a competitive salary at Grade 6 (http://www.lboro.ac.uk/services/hr/benefits/pay-rewards/)
    Start: as soon as possible. 
    Ph.D. studentships:
    Scholarship: £14,777 per annum plus tuition fees at the UK/EU rate. 
    Start: August 2018 or shortly after.
    Duration: 3.5 years.

    Enquiries and applications. Interested candidates are invited to send preliminary enquiries to a.soltoggio[at]lboro.ac.ukincluding a CV, a university transcript of marks, a list of references, and a statement of about 300 words motivating their interest in this area of research.

    -- 
    Dr.  Andrea Soltoggio
    Lecturer in Artificial Intelligence

    Department of Computer Science,
    Centre for Data Science, 
    Centre for Information Management 
    Haslegrave Building, N.2.03
    Loughborough University
    LE11 3TU, UK

    Phone: +44 (0) 1509 635748
    Email: a.soltoggio[at]lboro.ac.uk
    Twitter: [at]asoltoggio
    Web: http://www.lboro.ac.uk/departments/compsci/staff/dr-andrea-soltoggio.html


    [4] De : Joerg Stueckler <joerg.stueckler[at]tuebingen.mpg.de>

    Two PhD Student Positions in Computer Vision for Learnings Systems 
    within the Embodied Vision Group at the Max Planck Institute for Intelligent Systems in Tuebingen, Germany 
    https://ev.is.mpg.de/jobs/2-phd-student-positions-in-computer-vision-for-learning-systems

    Research questions 
    The Embodied Vision Group at the Max Planck Institute for Intelligent Systems investigates novel methods for autonomous systems to learn dynamic scene understanding and to use this understanding to perform complex tasks such as navigation or object manipulation. We aim at systems that learn from raw sensor measurements like images or tactile information and through action within their environment. A research focus in this context is on computer vision topics, including

    • Physical and 3D dynamic scene understanding
    • Learning of predictive environment models
    • Self-supervised and online visual and multi-modal learning
    • Vision-based interactive perception and learning for object manipulation
    • Vision-based navigation for drones and mobile robots
    • Deep reinforcement learning

    The positions 
    We are looking for two PhD students who are holding an outstanding Master’s degree in the computer or natural sciences, electrical or control engineering or applied mathematics. The PhD students will conduct research in one or several of the above topic areas.

    • Candidates should have studied areas related to computer vision and machine learning.Areas of particular interest for us at the moment are deep learning, visual scene understanding, visual/visual-inertial simultaneous localization and mapping, 3D scene reconstruction, robot vision, robot learning and deep reinforcement learning.
    • Successful candidates will typically have ranked at or near the top of their classes and be highly proficient in written and spoken English.
    • Excellent computer science skills as well as a strong mathematical background are required.
    • Prior research experience in computer vision, deep learning, robotic object manipulation or autonomous navigation is a plus.

    The PhD students will receive a PhD funding contract with an initial duration of 3 years. The position is funded for 3-4 years. Salary will be based on previous experience according to guidelines of the German Collective Wage Agreement for the Public Service (TVöD). The earliest start date is August 1st, 2018.

    The Max-Planck Society is committed to increasing the number of individuals with disabilities in its workforce and therefore encourages applications from such qualified individuals. The Max-Planck Society seeks to increase the number of women in those areas where they are underrepresented and therefore explicitly encourages women to apply.

    The group 
    The Embodied Vision Group (https://ev.is.mpg.de) is a newly established research group at the Max Planck Institute for Intelligent Systems in Tuebingen, Germany, and is lead by Dr. Joerg Stueckler. The institute is a world-class center for foundational research in machine learning, computer vision, robotics and material science. Tübingen is a scenic medieval university town, cradled in what is simultaneously one of Germany’s most beautiful landscapes in one of Europe’s most economically successful areas. The working language at the institute is English.

    How to apply 
    Applications and inquiries should be sent quoting reference number 33.18 to Dr. Joerg Stueckler (see contact details below). Applications must be submitted by email as a single pdf (max. 10 MB) and include a CV, motivation letter with research statement, publication list, transcripts of BSc and MSc degrees, and contact details of 2-3 references. Optionally, up to 2 selected own publications or theses can be included in a second pdf (max. 5 MB). Applications should also indicate earliest date of availability.

    There is no fixed application deadline; applications are considered until the positions have been filled or are no longer needed. Preference will be given to applications received before July 15th, 2018.

    For further details on the positions and how to apply, please visit 
    https://ev.is.mpg.de/jobs/2-phd-student-positions-in-computer-vision-for-learning-systems

    Please liberally forward this announcement and share to possibly interested candidates or persons who might know suitable candidates.

    Best regards, 
    Joerg Stueckler

    -- 
    Dr. Joerg Stueckler 
    Max Planck Research Group Leader

    Embodied Vision Group 
    Max-Planck-Institute for Intelligent Systems

    Max-Planck-Ring 4 
    72076 Tübingen 
    Tel. +49 (0)7071 601-385 
    Email: joerg.stueckler[at]tue.mpg.de 
    http://ev.is.mpg.de


    [5] De: Thomas Serre <thomas_serre[at]brown.edu>

    The Frank and Serre labs at Brown university are seeking applicants for the Paul J. Salem Postdoctoral Scholarships in Brain Science. The postdoctoral fellow will lead an exciting new project at the interface between machine learning and neuroscience. In particular, we are looking for computational neuroscience and machine learning experts interested in the intersection between vision, memory and reinforcement learning. Relevant projects in the two groups can be seen in the following example works:

    • Franklin, N.T. & Frank, M.J. (2018). Compositional clustering in task structure learning. PLOS Computational Biology, 14(4): e1006116. http://ski.clps.brown.edu/papers/FranklinFrank_Compositional18.pdf
    • Nassar, M.R., Helmers, J. & Frank, M.J. (2018). Chunking as a rational strategy for lossy data compression in visual working memory. Psychological Review. http://ski.clps.brown.edu/papers/NassarHelmersFrank_chunking.pdf
    • Drew Linsley, Junkyung Kim, Vijay Veerabadran, Thomas Serre.  Learning long-range spatial dependencies with horizontal gated-recurrent units. 2018 https://arxiv.org/abs/1805.08315v1
    • Drew Linsley, Dan Scheibler, Sven Eberhardt, Thomas Serre. Global-and-local attention networks for visual recognition. 2018 https://arxiv.org/abs/1805.08819v1
    Candidates are expected to have a solid background in one or more of the following domains: modern machine learning, computational models of neural dynamics underlying perceptual or cognitive processes, signal processing. In addition, to conducting primary research with neural networks, candidates will be involved in the mentoring of students, and will participate in workshops and challenges at the interface between machine learning and neuroscience (see e.g.,  http://compneuro.clps.brown.edu/datathon_2017/and http://compneuro.clps.brown.edu/2018-modeling-competition/).

    The initial appointment is for 12 months, renewable for another year, and potentially longer depending on funding. The start date is negotiable though an early start is strongly preferred. Salary is commensurate with experience and is competitive. We encourage Salem Scholars to seek external funding during their appointment, as a critical component in their professional development.

    Requirements:
    Candidates must have received their PhDs within 3 years of the application deadline, and will work under the supervision of Drs Frank and Serre who are affiliated with the Carney Initiative for Computation in Brain and Mind.   They must have a strong background in computational neuroscience and machine learning, with a track record of relevant publications at top venues (such as NIPS, ICML, PLOS Computational Biology, etc). Excellent programming skills are required (e.g., C/C++/Matlab/Python/R).

    Application:
    Please send your applications by email to michael_frank[at]brown.edu thomas_serre[at]brown.edu. Please include a brief statement of interests, a curriculum vita, a list of publications and the name of 2-3 reference writers (no letter needed at this stage). There is no deadline for the application but applicants are encouraged to apply as soon as possible as the position will be filled as soon as a suitable applicant is found.

    The Carney Initiative for Computation in Brain and Mind (CICBM; http://compneuro.clps.brown.edu), which began Fall 2013 as a component of BIBS, is an energetic and enthusiastic effort that fosters synergistic collaborations across departments. Groups affiliated with the initiative work on two core levels of computation. The first level focuses on theoretical neuroscience, including computational perception, control over action and learning, and fundamental questions in neuronal networks (synaptic plasticity, circuits, networks, oscillations). The second level focuses on applications and neurotechnology, including brain-machine interfaces, advanced neural data analysis, computer vision, computational psychiatry, and robotics. CICBM has 16 core computational faculty (http://compneuro.clps.brown.edu/people/) spanning six departments, and many more faculty who incorporate computation for theory development, analysis, or both. Computational neuroscience tools at Brown have been applied in projects including brain-machine control of robotic arms in paralyzed humans; models of visual systems in biological organisms and their innovative application for classifying animal behavioral patterns; predicting and quantifying effects of genetics, disease, medications, and brain stimulation on motor and cognitive function; identification of the source of neural rhythms and their roles in sensorimotor function; development of fundamental theories of brain plasticity, and learning; state-of-the art models of machine learning and reinforcement learning in computer science.

    The Carney Institute for Brain Science at Brown University advances multidisciplinary research, technology development, and training in the brain sciences and works to establish Brown University as an internationally recognized leader in brain research. The institute was just endowed with a new $100 million gift. CIBS unites more than 100 faculty from a diverse group of departments at Brown, spanning basic and clinical departments, and physical and biological sciences. CIBS provides a mechanism to advance interdisciplinary research efforts among this broad group.  CIBS provides  essential support to obtain and administer multi-investigator grants for research, infrastructure, and training. The Institute actively seeks new training funds to support interdisciplinary education that transcends that available in individual academic departments.

    -- 
    Thomas Serre | GMT -5:00  EDT | T: +1 401-484-0750 
    Associate Professor of Cog Ling & Psych Sciences | Brown University
    URL: goo.gl/G69SaF | Google Talk: tserre | Skype: thomas.serre

    posted in Jobs offers - Call for projects read more
  • Geo-Sci-Info

    Capture du 2018-06-26 12-48-30.png

    OFFICIAL WEBPAGE - INFORMATIONS - DOWNLOADS

    Capture du 2018-06-26 12-51-03.png

    Contact us by email at admin [at] ramp.studio if you are interested in using the platform in your classroom, as an internal tool for prototyping in your data science team, or to launch a data challenge. Consider joining our slack team if you would like to be part of the growing community of rampers.

    Paris-Saclay Center of Data Science: RAMPs are organized by the Paris Saclay Center for Data Science: A multi-disciplinary initiative to define, structure, and manage the data science ecosystem at the University Paris-Saclay.

    What is a RAMP?

    A RAMP is a collaborative data challenge. See here for more details.

    Bibliography:

    • The RAMP framework: from reproducibility to transparency in the design and optimization of scientific workflows, Kégl, Boucaud, Cherti, Kazakçı, Gramfort, Lemaitre, Van den Bossche, Benbouzid, Marini PREPRINT

    Team

    Alumni

    posted in GSI FORGE read more
  • Geo-Sci-Info

    Capture du 2018-06-26 09-25-12.png

    OFFICIAL WEBSITE - PROGRAM

    POSTER

    The main goal of this program is to gather the community of researchers working on questions that relate in some way statistical physics and high dimensional statistical inference. The format will be several (~10) 3h introductory lectures, and about twice as many contributed invited talks. The topics include:

    • Energy/loss landscapes in disordered systems, machine learning and inference problems
    • Computational and statistical thresholds and trade-offs
    • Theory of artificial multilayer neural networks
    • Rigorous approaches to spin glasses and related models of statistical inference
    • Parallels between optimisation algorithms and dynamics in physics
    • Vindicating the replica and cavity method rigorously
    • Current trends in variational Bayes inference
    • Message passing algorithms
    • Applications on machine learning in condensed matter physics
    • Information processing in biological systems

    Deadline Registration : February 28, 2018

    Lecturers

    • Gerard Ben Arous (Courant Institute)
    • Giulio Biroli (CEA Saclay, France)
    • Nicolas Brunel (Duke University)
    • Yann LeCun (Courant Institute and Facebook)
    • Michael Jordan (UC Berkeley)
    • Stephane Mallat (ENS et college de France)
    • Andrea Montanari (Stanford)
    • Dmitry Panchenko (University of Toronto, Canada)
    • Sundeep Rangan (New York University)
    • Riccardo Zecchina (Politecnico Turin, Italy)

    Speakers

    • Antonio C Auffinger (Northwestern University, USA)
    • Afonso Bandeira (Courant Institute, USA)
    • Jean Barbier (Queens Mary, UK)
    • Quentin Berthet (Cambridge UK)
    • Jean-Philippe Bouchaud (CFM Paris, France)
    • Silvio Franz (Paris-Orsay, France)
    • Surya Ganguli (Stanford, USA)
    • Alice Guionnet (ENS Lyon, France)
    • Aukosh Jagganath (Harvard, USA)
    • Yoshiyuki Kabashima (Tokyo Tech, Japan)
    • Christina Lee (Microsoft Research, USA)
    • Marc Lelarge (ENS Paris, France)
    • Marc Mezard (ENS Paris, France)
    • Leo Miolane (ENS Paris, France)
    • Remi Monasson (ENS Paris, France)
    • Giorgio Parisi (Roma La Sapienza, Italy)
    • Will Perkins (University of Birmingham, UK)
    • Federico Ricci-Tersenghi (Roma La Sapienza, Italy)
    • Cythia Rush (Columbia, USA)
    • Levent Sagun (CEA Saclay, France)
    • Samuel S. Schoenholz (Google Brain, USA)
    • David Jason Schwab (CUNY, USA)
    • Guilhem Semerjian (ENS Paris, France)
    • Alexandre Tkatchenko (University of Luxembourg)
    • Naftali Tishby (Hebrew University, Israel)
    • Pierfrancesco Urbani (CEA Saclay, France)
    • Francesco Zamponi (ENS Paris, France)

    Organizing Committee

    • Florent Krzakala (ENS & UPMC, Paris)
    • Lenka Zdeborova (CEA & CNRS, Saclay)

    Capture du 2018-06-26 11-59-19.png

    posted in International School - Statistical Physics and Machine Learning back together read more
  • Geo-Sci-Info

    Capture du 2018-06-21 16-35-53.png

    OFFICIAL WEBSITE - REGISTRATION

    Presentation

    This Summer School will consist in two courses given by professors Sergey Bobkov (Minneapolis) and Mokshay Madiman (Delaware) on Information Theory and Convex Analysis. The aim is to bring researchers from different communities (Probability, Analysis, Computer science) in the same place.
    Participation of postdocs and PhD students is strongly encouraged. The school has some (limited number of) grants for young people (see the "registration" link above).

    Abstracts        
    Sergey Bobkov (Minneapolis) : Strong probability distances and limit theorems
    Abstract: The lectures explore strong distances in the space of probability distributions, including total variation, relative entropy, chi squared and more general Renyi/Tsallis informational divergences, as well as relative Fisher information. Special attention is given to the distances from the normal law. The first part of the course is devoted to the general theory, and the second part to the behavior of such informational distances along convolutions and associated central limit theorem.

    Mokshay Madiman (Delaware): Entropy power and related inequalities in continuous and discrete settings
    Abstract: The lectures explore the behavior of Renyi entropies of convolutions of probability measures for a variety of ambient spaces. The first part of the course focuses on Euclidean spaces, beginning with the classical Shannon-Stam entropy power inequality and the closely related Brunn-Minkowski inequality, and developing several of the generalizations, variants, and reversals of these inequalities. The second part of the course focuses on discrete abelian groups, where one sees close connections to additive combinatorics.

    Related Materials

    (1) Survey on entropic limit theorems (M. Madiman) :

    • Lecture 1. Introduction - What is information theory? The first question that we want to address is: “What is information?” Although there are several ways in which we might think of answering this question, the main rationale behind our approach is to distinguish information from data. We think of information as something abstract that we want to convey, while we think of data as a representation of information, something that is storable/communicable. This is best understood by some examples.
    • Lecture 2. Basics / law of small numbers. Due to scheduling considerations, we postpone the proof of the entropic central limit theorem. In this lecture, we discuss basic properties of the entropy and illustrate them by proving a simple version of the law of small numbers (Poisson limit theorem). The next lecture will be devoted to Sanov’s theorem. We will return to the entropic central limit theorem in Lecture 4.
    • Lecture 3. Sanov’s theorem. The goal of this lecture is to prove one of the most basic results in large deviations theory. Our motivations are threefold: 1. It is an example of a probabilistic question where entropy naturally appears. 2.The proof we give uses ideas typical in information theory. 3. We will need it later to discuss the transportation-information inequalities (if we get there).
    • Lecture 4. Entropic CLT (1). The subject of the next lectures will be the entropic central limit theorem (entropic CLT) and its proof.
    • Lecture 5. Entropic CLT (2). The goal of this lecture is to prove monotonicity of Fisher information in the central limit theorem. Next lecture we will connect Fisher information to entropy, completing the proof of the entropic CLT.
    • Lecture 6. Entropic CLT (3). In this lecture, we complete the proof of monotonicity of the Fisher information in the CLT, and begin developing the connection with entropy. The entropic CLT will be completed in the next lecture.
    • Lecture 7. Entropic CLT (4). This lecture completes the proof of the entropic central limit theorem.
    • Lecture 8. Entropic cone and matroids. This lecture introduces the notion of the entropic cone and its connection with entropy inequalities.
    • Lecture 9. Concentration, information, transportation (1). The goal of the next two lectures is to explore the connections between concentration of measure, entropy inequalities, and optimal transportation.
    • Lecture 10. Concentration, information, transportation (2) Recall the main proposition proved in the previous lecture, which is due to Bobkov and Götze (1999).

    (2) A survey on forward and reverse entropy power inequalities, 2017.

    Organizers

    Sponsors

    posted in Summer School - Information theory: Inequalities distances and analysis read more
  • Geo-Sci-Info

    Applications are invited for a 4-year PhD studentship in Neuroscience and Machine Learning at the University of Oxford. The successful candidate will join the Oxford Interdisciplinary Bioscience DTP and will work with Andrew Saxe, Tim Behrens and Christopher Summerfield on understanding the computational mechanisms of learning in biological systems and artificial agents, and will be primarily based in the Department of Experimental Psychology. The successful candidate will have the opportunity to collaborate with an industry partner.

    Applicants should have strong quantiative skills and a background in computer science, computational neuroscience, statistics, or a related discipline. Applicants with a joint interest in both machine learning/AI research and computational neurobiology are particularly encouraged to apply.

    In the first instance please contact Dr Andrew Saxe (asaxe [at] fas.harvard.edu) to discuss your suitability for the project. The application deadline is Friday 13th July 2018 (see http://www.biodtp.ox.ac.uk/how-apply/bbsrc-artificial-intelligence-npif-studentships-2018.html for details). This project is funded for four years by the Biotechnology and Biological Sciences Research Council BBSRC and as such is only available to UK and EU candidates. Successful students will receive a stipend of no less than the standard RCUK stipend rate, currently set at £14,777 per year.

    posted in Jobs offers - Call for projects read more

Internal error.

Oops! Looks like something went wrong!