SUMO final review meeting KNMI

Posted by Frank Selten on April 5th, 2014

On June 3td 2014 the final SUMO review meeting will take place in De Bilt in the Netherlands, hosted by the Royal Netherlands Meteorological Institute. At the meeting an overview will be presented on the achievements of the SUMO project, the latest developments and an outlook for future developments. Here's a tentative agenda.


Time Title Presenter
08:30 - 08:40 Welcome and introduction Kocarev, MANU
08:40 - 09:10 WP1 Potentials and limits of SUMO Kurths, PIK
09:10 - 09:40 WP2 Learnability and performance estimation Wiegerinck, RU
09:40 - 10:10 WP7 Learning to interconnect ODE models Dzeroski, RU
10:10 - 10:40 WP3 Connecting intermediate complexity climate models Duane, MANU
10:40 - 11:00 Coffee break
11:00 - 11:30 WP4a WP4b SUMO applied in complex weather models Berner NCAR, Selten KNMI
11:30 - 12:00 WP5 SUMO applied in complex climate models Keenlyside, Shen, UiB
12:00 - 12:30 WP8 Connecting state-of-the-art climate models Kirtman, Miami
12:30 - 12:45 WP6 Outreach, achievements, beyond SUMO Kocarev, MANU
12:45 - 14:00 Lunch break
14:00 - 15:00 Reviewers meeting with program officer
15:00 - 15:15 Coffee break
15:15 - 16:00 Reviewers feedback and discussion

SUMO scientific spring meeting Ljubljana

Posted by Frank Selten on May 13th, 2013

On 4-5 May 2013 a scientific SUMO project meeting took place in Ljubljana, hosted by the Institut Jozef Strephan. The aim was to exchange the progress in the understanding of the behavior of weakly coupled models, discuss the results of the WP5 super model after learning, discuss attractor learning approaches versus vector field learning and discuss the cooperation with the interactive ensemble infrastructure of Ben Kirtmann of University of Miami.

SUMO project review meeting Potsdam

Posted by Frank Selten on Nov 16th, 2012

On 19-20 Nov 2012 the second annual review meeting took place at PIK in Potsdam. On monday a project meeting was held at PIK, on tuesday the project members presented the second year results to the review board. The review board consists of prof. dr. Klaus Hasselmann, prof. dr. Michael Ghil and prof .dr. Witold Dzwinel.


Time Title Presenter
08:45 - 09:00 Reviewers briefing (restricted reviewers meeting) --
09:00 - 09:20 SUMO - overview Kocarev, MANU
09:20 - 09:50 WP1 Connectivity and optimization Kurths, PIK
09:50 - 10:20 WP2 Quality of local optima and global optimization methods Wiegerinck, RU
10:20 - 10:50 Coffee break
10:50 - 11:10 extra Forecast with imperfect models Basnarkov, MANU
11:10 - 11:40 WP7 Generating diverse models to include in supermodels Dzeroski, IJS
11:40 - 12:10 WP3 Supermodeling with QG channel models Duane, MANU
12:10 - 14:10 Lunch break
14:20 - 14:50 WP4 Guidelines for a climate super model Selten, KNMI
14:50 - 15:10 WP5 Super climate model behaviour after learning Keenlyside, UiB
15:10 - 15:30 WP8 Development of a trans-Atlantic climate supermodel Keenlyside, UiB
15:30 - 16:30 Coffee break and reviewers debriefing (restricted)
16:30 - 17:30 Reviewers feedback to the consortium

SUMO scientific spring meeting Bergen

Posted by Frank Selten on May 22nd, 2012

On 22-24 May 2012 a scientific SUMO project meeting took place in Bergen. The aim was to discuss about learning and connection strategies for the state-of-the-art climate models in WP5. The program consisted of scheduled talks by the different working groups and plenty of time for discussions.

SUMO project review meeting Skopje

Posted by Frank Selten on May 10th, 2011

On 23-24 Nov 2011 the first annual review meeting took place in Skopje. On wednessday a project meeting was scheduled at MANU, on thursday the project members presented the first year results to the review board. The review board consists of prof. dr. Klaus Hasselmann, prof. dr. Michael Ghil and prof .dr. Witold Dzwinel.


Time Title Presenter
08:45 - 09:00 Reviewers briefing (restricted reviewers meeting) --
09:00 - 09:30 SUMO - overview Kocarev and Duane, MANU
09:30 - 10:00 WP1 Models and metrics Kurths and Fujiwara, PIK
10:00 - 10:30 WP2 Learning low complexity and intermediate complexity models Wiegerinck, RU
10:30 - 11:00 WP3 Connections learning approaches for PDEs Duane, MANU
11:00 - 11:30 WP4 Connections in climate models of intermediate complexity Selten and Hiemstra, KNMI
11:30 - 12:00 WP5 Super climate model with manually chosen connections Keenlyside, UIB
12:00 - 12:30 WP7 Learning complete supermodels Saso Dzeroski, IJS
12:30 - 14:00 Lunch break
14:00 - 14:30 On the limits of super modeling Greg Duane, MANU
14:30 - 15:00 Coupling, synchronisation and averaging Wim Wiegerinck, RU
15:00 - 15:30 Modeling the mechanisms of major climate shift with machine learning approaches Saso Dzeroski, IJS
15:30 - 15:45 coffee break
15:45 - 16:30 Reviewers debriefing (restricted)
16:30 - 17:00 Reviewers feedback to the consortium
20:00 - Dinner

FET11 meeting in Budapest

Posted by Frank Selten on May 10th, 2011

On 4-6 May 2011 members of the SUMO project participated in the FET11 meeting in Budapest. A poster was presented with the objectives and first results of the project. There was interest in the super modeling approach from people in the field of robotics in the learning of complex tasks given a set of available action-response models. The SUMO project members discussed the hierarchy of models suitable for developing the super modeling approach and different learning approaches.

SUMO kick-off meeting in Skopje

Posted by Frank Selten on Nov 10th, 2010

On 3 and 4 November 2010 a group of 20 scientists gathered at the Macedonian Academy of Sciences and Arts in Skopje to exchange knowledge and ideas relevant for the SUMO project. After a word of welcome by the project coordinator Ljupco Kocarev, the work package leaders presented an overview of the background knowledge relevant for the work in SUMO and the models, methods and strategies that will be employed to develop the supermodeling concept. In addition Saso Dzeroski from the Josef Stefan Institute presented results of data-driven modeling studies that could be relevant for SUMO and Naoya Fujiwara from Potsdam Institute presented a study that showed how synchronization errors can be used in parameter learning. On the second day organizational and science issues were discussed in a smaller group meeting.