in conjunction with MICCAI
Deep learning empowers enormous scientific advances, with key applications in healthcare. It has been widely accepted that it is possible to achieve better models with growing amounts of data. However, enabling learning on these huge datasets or training huge models in a timely manner requires distributing the learning on several devices. One particularity in the medical domain, and in the medical imaging setting is that data sharing across different institutions often becomes impractical due to strict privacy regulations, making the collection of large-scale centralized datasets practically impossible.
Some of the problems, therefore, become: how can we train models in a distributed way on several devices? And is it possible to achieve models as strong as those that can be trained on large centralized datasets without sharing data and breaching the restrictions on privacy and property? Distributed machine learning, including Federated Learning (FL) approaches, could be helpful to solve the latter problem. Different institutions can contribute to building more powerful models by performing collaborative training without sharing any training data. The trained model can be distributed across various institutions but not the actual data. We hope that with FL and other forms of distributed and collaborative learning, the objective of training better, more robust models of higher clinical utility while still protecting the privacy within the data can be achieved.
Through the second MICCAI Workshop on Distributed And Collaborative Learning (DCL), we aim to provide a discussion forum to compare, evaluate and discuss methodological advancements and ideas around federated, distributed, and collaborative learning schemes that are applicable in the medical domain. Further details can be found here.
- When? 01 October 2021 (TBC)
- Where? Online
Workshop co-organisers: Shadi Albarqouni (Helmholtz AI, TU Munich), Spyridon Bakas (University of Pennsilvania), M. Jorge Cardoso (King’s College London), Bennett Landman (Vanderbilt University), Nicola Rieke (NVIDIA), Xiaoxiao Li (Princeton University), Daguang Xu (NVIDIA), Holger Roth (NVIDIA).
Featuring Applications from Health & Energy
This is the first hackathon of this kind organized by the Incubator platforms Helmholtz Imaging Platform (HIP), Helmholtz Information & Data Science School for Health (HIDSS4Health), Helmholtz Metadata Collaboration (HMC), Helmholtz Information & Data Science Academy (HIDA) and Helmholtz AI. Further information: an indico event page is in preparation, stay tuned.
- When? 28 - 30 September 2021 (TBD)
- Where? Online
- Who? Co-organized by HIP, HIDSS4Health, HMC, HIDA and Helmholtz AI.
in conjunction with MICCAI
As we witness a technological revolution that is spinning diverse research fields including healthcare at an unprecedented rate, we face bigger challenges ranging from the high cost of computational resources to the reproducible design of affordable and innovative solutions. While AI applications have been recently deployed in the healthcare system of high-income countries, its adoption in developing and emerging countries remains limited. Here, we present the first workshop aiming to i) raise awareness about the global challenges in healthcare, ii) strengthen the participation of underrepresented communities at MICCAI, and iii) build a community around Affordable AI and Healthcare in low resource settings. We are looking for contributions on (a) making AI affordable for healthcare, (b) making healthcare affordable with AI, or (c) pushing the frontiers of AI in Healthcare that enables (a) or (b).
There are formidable challenges that remain untackled in this field which are spanned by different questions: (1) How to promote knowledge sharing between AI institutions for developing countries? (2) How to build robust and affordable solutions that can be used with limited technology and poor data and communication quality? (3) How can we propel the collection and analysis of data for underrepresented populations with limited quality and annotation? (4) How can we encourage computer scientists and clinicians to identify unmet clinical needs in neglected diseases that can benefit from affordable AI solutions? -- just to raise a few.
We invite you to submit your contributions (regular long paper -- 12-page limit) that address medical problems of emerging and developing countries via algorithms spanning different sub-fields. Further details can be found here.
- When? 27 September (TBC)
- Where? Online
Workshop co-organisers: Shadi Albarqouni (HelmholtzAI, TU Munich), Bishesh Khanal (NAAMII), Islem Rekik (ITU), Nicola Rieke (NVIDIA), Debdoot Sheet (IIT Kharagpur).
We are delighted that the cooperation between McGill University in Montréal and Forschungszentrum Jülich (FZJ), one of our longest and most successful partnerships, has now been formally organized as a Helmholtz International Lab. This enables us to increase our efforts to build the next generation of multimodal human brain atlases around the successful BigBrain model. With the Helmholtz International BigBrain Analytics & Learning Laboratory (HIBALL) we develop highly detailed 3D brain models using novel AI methods and supercomputing architectures. These models will be exposed through a sustainable, transcontinental research platform for computation and data sharing, with high interoperability with the systems of large brain initiatives. With the upcoming BigBrain Workshop we aim to reach out to the international community of BigBrain users and to invite researchers from our global network (and beyond) to present their work and to discuss future prospects of the BigBrain data and tools.
The workshop will be organized as a symposium, with both invited speakers and contributed talks as well as a poster and demo session. We welcome short abstracts of current work and/or short proposals for future initiatives related to the BigBrain. The event is free of charge but prior registration is required.
- When? 22 - 23 September 2021 (TBD)
- Where? Online
- Who? The workshop is co-organized by McGill University & FZJ as part of HIBALL with support from
- Deadlines? 23 August 2021 (Abstract submission for talks)
06 September 2021 (Abstract submission for posters)
13 September 2021 (Poster upload)
20 September (Registration)
Workshop co-organisers: Alan C Evans (McGill University), Katrin Amunts (FZJ, Helmholtz AI associate), Paule-Joanne Toussaint (McGill University), Susanne Wenzel (FZJ, Helmholtz AI)
Many in the ML community wish to take action on climate change, but are unsure of the pathways through which they can have the most impact. This workshop highlights work that demonstrates that ML can be an invaluable tool in reducing greenhouse gas emissions and in helping society adapt to the effects of climate change. Climate change is a complex problem, for which action takes many forms - from theoretical advances to deployment of new technology. Many of these actions represent high-impact opportunities for real-world change, and are simultaneously interesting research problems.
This workshop is part of a series (NeurIPS 2020, ICLR 2020, NeurIPS 2019, ICML 2019) and takes place at the International Conference on Machine Learning (ICML), one of the premier conferences on machine learning. For this iteration of the workshop, the keynote talks and panel discussions will be particularly focused on connections to the Paris Agreement and reaching net-zero emission targets, but submitted papers may be on any topic of relevance to climate change and machine learning. The workshop is open to the public; it is not necessary to submit a paper to the workshop in order to attend.
- When? 23 - 24 July 2021 (TBD)
- Where? Online
- Deadlines? 28 April 2021 (Mentorship program application)
31 May 2021 (Paper/Proposal submission)
Workshop co-organisers: Hari Prasanna Das (UC Berkeley), Katarzyna (Kasia) Tokarska (ETH Zurich), Maria João Sousa (IST, ULisboa), Meareg Hailemariam (DAUST), David Rolnick (Mila, McGill), Xiaoxiang Zhu (TU Munich, Helmholtz AI, ELLIS Munich), Yoshua Bengio (Mila, UdeM)
powered by Helmholtz Information & Data Science Academy (HIDA) in cooperation with Helmholtz Artificial Intelligence Cooperation Unit (Helmholtz AI), Munich School for Data Science (MUDS), Ludwig-Maximilians-Universität München (LMU) und Munich Center for Machine Learning (MCML)
DIVE INTO MACHINE LEARNING
Building on the great response last year, the Virtual ML Summer School will be offered again in 2021. The aim is to provide an introductory course in basic techniques and concepts of supervised machine learning (ML), which has become a central part of modern data analysis.
The course is constructed holistically and as self-contained as possible, in order to cover most relevant areas of supervised ML. While the introductory parts are more aimed at a practical and operational understanding of the covered algorithms and models, we also include sound theoretical foundations and proofs in more advanced sections in order to teach ML theory as self-contained and precise as possible.
- When? 21 - 27 July 2021 (Part I) & 01 - 07 September 2021 (Part II)
- Where? Online
- Who? Co-organized by HIDA in cooperation with MUDS, LMU, MCML and
Live sessions will take place in two consecutive blocks. Participants sign up for both blocks.
Block 1: 21 - 27 July 2021 (weekend excluded), 9:30 am - 1 pm CEST
21.07. ML Basics
26.07. Classification knn
Block 2: 01 - 07 September 2021 (weekend excluded), 9:30 am - 1 pm CEST
03.09. Random Forests
07.09. Practical Advice
Trainers: Bernd Bischl (LMU München, MCML), Ludwig Bothmann (LMU München), Tobias Pielok (LMU München)
To overcome limitations in computing and data analytics related to Earth System science, the uptake of artificial intelligence (AI) and machine learning (ML) methodologies is currently being explored. Multiple initiatives are now emerging to tackle open challenges such as subscale parametrization, detection of patterns and in-situ analysis, adoption of ML for alternative process models, or dedicated fast prediction systems to address specific end-user needs.
Following the first workshop held in February 2020, this second workshop will provide an update on the state of the art in applying and extending AI/ML techniques in topics relevant to Earth System science, from integrating ML with models to deriving new insights from observational data, to extending AI/ML to impact models and approaches. We call for contributions to the sessions, offering presentation and lightning talk slots for those interested in submitting an abstract. Contributions are welcome from all participants, from updates on ongoing research activities and future plans to technical insights worthwhile to share.
- When? 03 - 04 May 2021 (14:30-18:00 CEST)
- Where? Online
- Who? The workshop is co-organized by DKRZ, HZG and GERICS with support from Helmholtz AI.
Workshop co-organisers: Laurens Bouwer (GERICS), Christopher Kadow (DKRZ), Tobias Weigel (DKRZ/Helmholtz AI), Eduardo Zorita (HZG)
You would like to contribute to medically relevant research questions with Data Science? Then take part in the first international COVID Data Challenge. The topic will focus on challenges in medical imaging in the context of the COVID pandemic.
ELLIS Genoa teamed up with HIDA, ELLIS Munich and the Israel Data Science Initiative for a two-day-virtual data science challenge, which is supported by Helmholtz AI. The task will be based on multimodal data including images and other variables. Sign up - and use your AI and Data Science skills to help doctors and scientists fight COVID.
How it works
Save your spot today and join us from anywhere with your computer and a stable internet connection in April. You’ll be able to team up and collaborate with data scientists across borders during the virtual event. The winning team will be awarded an attractive prize - of course related to data science! Please note that the times refer to CET.
We’re looking forward to seeing you there!
- When? 28 - 29 April 2021 (11:00-17:30 CEST)
- Where? Online
- Who? The workshop is co-organized by ELLIS Genoa, HIDA, ELLIS Munich and Israel Data Science Initiative with support from Helmholtz AI.
In a data-based future, it will be key to democratise access to AI to maximise research impact. We want to show you how.
Join us to meet method and domain specialists with a shared interest in AI, learn more about the Helmholtz AI initiatives, discuss use cases in applied AI/ML and expand your network.
- When? 14 - 15 April 2021 (afternoon CET)
- Where? Online
Additionally, you can join the discussion on Twitter through the hashtag #HelmholtzAIcon.
Access to computing resources is key for the Helmholtz AI community to accelerate innovative AI applications. Therefore, the “Helmholtz AI computing resources” (HAICORE) was initiated at the end of 2019 as a one-off measure within the framework of the Helmholtz Incubator. HAICORE comprises Helmholtz Association of German Research Centres funding of 2.5 million euro, mainly for GPU hardware, which is implemented at the Forschungszentrum Jülich (FZJ) and Karlsruhe Institute of Technology (KIT). The resources should be open to the entire AI community within the Helmholtz Association.
This workshop is designed to guide you through the first steps of using supercomputing machines for your own AI application. We will try to get your code and workflow up and running, and aim to make getting started on a supercomputer as smooth as possible. After this course, you will not only be ready to use HAICORE, but you will have taken the first step to unlocking compute resources at the largest scale with a compute time application at the Gauss Supercomputing Center.
The workshop will be held in a small group size to ensure that your questions will be addressed. Please let us know what topics you are interested in and we will try to adjust.
- When? 06 - 07 April 2021 (9:00-13:00 CEST)
- Where? Online
- Who? The workshop is co-organized by FZJ & KIT with support from Helmholtz AI.
Workshop co-organisers: Stefan Kesselheim (FZJ) & Markus Götz (KIT)
GPU Hackathons are five day intensive hands-on events designed to help computational scientists port their applications to GPUs using libraries, OpenACC, CUDA and other tools by pairing participants with dedicated mentors experienced in GPU programming and development. Representing distinguished scholars and preeminent institutions around the world, these teams of mentors and attendees work together to realize performance gains and speedups by taking advantage of parallel programming on GPUs.
A collection of GPU lectures, tutorials, and labs are available for all participants at no fee.
This event is jointly organized by Helmholtz-Zentrum Dresden-Rossendorf (HZDR) and Jülich Supercomputing Centre (JSC) in association with the Helmholtz Federated IT Services Software Cluster (HIFIS).
- When? 15, 22, 23, 24 March 2021
- Where? Online
23 SEPT 2020 // Helmholtz virtual data science career day 2020
Are you a data scientist looking to contribute to solving major challenges facing society, science and the economy? Then visit the Helmholtz Virtual Data Science Career Day and jump-start the next step in your career. Whether you’re looking for exciting PhD subjects, postdoc positions or other positions in applied data science, this fair will get you in touch with your future employer.
Visit our Helmholtz AI booth and learn about our community and the data science jobs it may offer in these fields: Energy, Earth and environment, Health, Aeronautics, space and transport, Matter and Information.
21 SEPT - 02 OCT 2020 // Helmholtz ML virtual summer school 2020
powered by Helmholtz Information & Data Science Academy (HIDA) in cooperation with Helmholtz Artificial Intelligence Cooperation Unit (Helmholtz AI), Munich School for Data Science (MUDS), Ludwig-Maximilians-Universität München (LMU) and Munich Center for Machine Learning (MCML)
DIVE INTO MACHINE LEARNING
Core program of the Virtual ML Summer School 2020 is an introductory course to fundamental techniques and concepts of supervised machine learning (ML), which has become a central part of modern data analysis.
In particular non-linear and non-parametric methods have been used successfully in uncovering complex patterns and relationships by computer scientists and statisticians.
|Sep 21||ML Basics |
|Sep 22||Supervised regression|
|Sep 23||HIDA virtual career day|
|Sep 24||Supervised Classification|
|Sep 25||Supervised Classification |
Keynote by Bernd Bischl on mlr3
|Sep 29||Helmholtz AI showcase - Dominik Thalmeier: "Using anomaly detection to identify mutations that effect hearing behaviour" |
|Sep 30||Trees |
|Oct 1||Helmholtz AI showcase - Christian Müller: "Sparse predictive modeling of microbiome data" |
|Oct 2||Practical Advice|
The focus of the course is to give a basic understanding of the different algorithms, models and concepts while explaining the necessary mathematical foundation.
Participants will acquire theoretical as well as practical competencies regarding some fundamental models of learning from data. Also participants will be enabled to conduct a data analysis project, including understanding and interpreting the data, in order to critically judge the advantages and disadvantages of the different methods.
Virtual meetings and group work (except 23 Sep) take place from 1-4:30pm.
You are required to work through course materials (videos, quizzes, online exercises) in prepartion for the live sessions by yourself and at your own pace. The live sessions will be used to put the concepts you learned about into practice.
You will have access to the course materials a couple of weeks prior to the starting date. It is up to you if you prepare topics in the mornings before each session in the afternoon or in the weeks before the course starts.
The course is targeted at ML beginners with a basic, university level, education in maths and statistics:
- Basic linear algebra: vectors, matrices, determinants
- Simple calculus: derivatives, integrals, gradients
- Some probability theory: probability, random variables, distributions
- Basic statistics knowledge: descriptive statistics, estimators.
(Linear) modelling from a statistics perspective will help, but is not required.
- Working knowledge of R
Bernd Bischl (LMU München, MCML); Tobias Pielok (LMU München); Heidi Seibold (Helmholtz AI)
Christian Müller (Helmholtz AI), Dominik Thalmeier (Helmholtz AI)
The program of the Helmholtz Virtual ML Summer School builds on the course program Introduction to Machine Learning (I2ML), which was developed by Bernd Bischl, Fabian Scheipl, Heidi Seibold, Christoph Molnar and Daniel Schalk. Concept and materials are accessible and licensed under Creative Commons Attribution 4.0 International (CC BY 4.0). If you use the course, the initiators are looking forward to your feedback!
The deadline for registration was 31 July 2020.
Registrations that we received afterwards were automatically put on the waiting list. The demand was very high, so we will try to offer more events of this kind in the future.
MUDS, Helmholtz AI and HIDA are part of the Helmholtz Incubator Information & Data Science.
05 MAR 2020 // Helmholtz AI kick-off meeting
Helmholtz AI will enable the efficient and agile development and implementation of applied artificial intelligence (AI) and machine learning (ML) assets across the whole Helmholtz Association. To realize this mission, we strongly support networking between all Helmholtz centres and the official Helmholtz AI kick-off meeting is the perfect opportunity to meet method and domain specialists with a shared interest in AI.
Representatives of all Helmholtz centres will learn more about Helmholtz AI initiatives, discuss use cases in applied AI/ML and expand their network.
- When? 05 March 2020
- Where? Lenbach Palais, Munich
To make the Helmholtz AI Kick-off event accessible to our entire community, we will provide live streaming of the event on our website. You are most welcome to join us!
Additionally, you can join the discussion in Twitter through the hashtag #helloHelmholtzAI.
19-21 FEB 2020 // Hacking for Health: Helmholtz AI health hackathon
UPDATE 14.02.2020 >> REGISTRATION CLOSED - Thank you!
In the face of increasingly complex health issues, here at Helmholtz AI central unit - Munich we want to bring together motivated people across the Helmholtz Association to solve real-world challenges through cross-disciplinary collaboration and innovative AI/ML methods. Are you up for the challenge? Here's what you need to do:
> First step: Fill in the FREE registration form before Friday, February 14
> Second step: Take your laptop
> Last but not least: Join us in Helmholtz Zentrum München for a three-day event (February 19 - 21, 2020) in which you'll be part of a team with other research talents; together you'll try to find an innovative solution to a challenge proposed by a health-focused Helmholtz Center.
Wait, there's more!
> Each team will be mentored by an expert scientist from the Helmholtz Center that proposed the challenge.
> Relevant datasets will already be preprocessed so that you can focus on the challenge as much as possible.
> At the end of the Helmholtz AI health hackathon every team gets the chance to present its solutions and even publish the results!
Of course you'll also enjoy delicious food, have fun, get the chance to know each other better and chat with inspiring experts on applied AI and ML. Join us!