- Machine-learning-nederland

Annual meeting of the EMS TAG SciML, MOX, Department of Mathematiks, Politecnico di Milano, Italy, March 24-26, 2025
by Alexander Heinlein 21 Dec '24

21 Dec '24

Dear Colleagues, We are delighted to invite you to the Annual Meeting of the Topical Activity Group for Scientific Machine Learning of the European Mathematical Society (EMS TAG SciML) which will take place at MOX, Department of Mathematics, Politecnico di Milano, Milan, Italy from March 24-26, 2025; see https://www.mate.polimi.it/events/EMS-TAG-SciML-25/index.php The meeting will bring together researchers and practitioners from across Europe and beyond to discuss the latest developments and future directions in the field of scientific machine learning. The program includes the following features * Invited Lectures by leading experts in the field from academia and industry * Tutorial Session introducing to topics of scientific machine learning * Poster Presentations showcasing cutting-edge research * Panel Discussion/Round Table on emerging trends and challenges * Networking Opportunities to connect with peers and collaborators We encourage contributions to the poster session from all areas of scientific machine learning; the deadline for poster submissions is January 10, 2025. Participation in the workshop is free of charge but registration is required (https://www.mate.polimi.it/events/EMS-TAG-SciML-25/index.php) We look forward to welcoming you to Milan. Best regards, On behalf of the EMS TAG SciML Organizing and Scientific Committees Axel Klawonn (spokesperson of the EMS TAG SciML) P.S.: For more information on the EMS TAG SciML, see https://ems-tag-sciml.github.io/ -- --- Alexander Heinlein Delft University of Technology (TU Delft) Delft Institute of Applied Mathematics (DIAM) Mekelweg 4, 2628CD Delft, Netherlands https://searhein.github.io

1 0

BeNeRL Reinforcement Learning Seminar: Hojoon Lee from KAIST AI (December 19, this Thursday)
by Yang, Z. (Zhao) 16 Dec '24

16 Dec '24

Dear colleagues, Our next BeNeRL Reinforcement Learning Seminar (Dec. 19) is coming: Speaker: Hojoon Lee (https://joonleesky.github.io<https://joonleesky.github.io/>), PhD student from KAIST AI. Title: Designing Neural Network Architecture for Deep Reinforcement Learning Date: December 19, 16.00-17.00 (CET) Please find full details about the talk below this email and on the website of the seminar series:https://www.benerl.org/seminar-series The goal of the online BeNeRL seminar series is to invite RL researchers (mostly advanced PhD or early postgraduate) to share their work. In addition, we invite the speakers to briefly share their experience with large-scale deep RL experiments, and their style/approach to get these to work. We would be very glad if you forward this invitation within your group and to other colleagues that would be interested (also outside the BeNeRL region). Hope to see you on December 19! Kind regards, Zhao Yang & Thomas Moerland VU Amsterdam & Leiden University —————————————————————— Upcoming talk: Date: December 19, 16.00-17.00 (CET) Speaker: Hojoon Lee (https://joonleesky.github.io<https://joonleesky.github.io/>) Title: Designing Neural Network Architecture for Deep Reinforcement Learning Zoom: https://universiteitleiden.zoom.us/j/65411016557?pwd=MzlqcVhzVzUyZlJKTEE0Nk… Abstract: While scaling laws have accelerated breakthroughs in computer vision and language modeling, their effects are less predictable in reinforcement learning (RL), where simply “scaling up” data, parameters, and computations rarely guarantees better results. In this talk, I will explore the barriers that make scaling challenging in RL and introduce new architectural designs that alleviate these challenges. I will also discuss future research opportunities that can improve scaling laws in RL. Bio: Hojoon is a Ph.D. student at KAIST AI, advised by Professor Jaegul Choo. He previously received his M.S. from KAIST and his B.S. from Korea University. In 2024, he interned with the GranTurismo team at Sony AI, mentored by Takuma Seno and Professor Peter Stone. In 2022, he interned with the RL team at Kakao, developing a new RL framework Jordly. His research focuses on designing network architectures and algorithms for RL that can continually learn, adapt, and generalize in dynamic environments.

1 0

Spring School on Control Theory and Reinforcement Learning, 17-21 March 2025, at CWI, Amsterdam, NL
by Aditya Gilra 13 Dec '24

13 Dec '24

Dear all, We invite registrations for the Spring School on Control Theory and Reinforcement Learning from 17-21 March 2025, at the Centrum Wiskunde & Informatica (CWI, research institute for mathematics and computer science in the Netherlands). This spring school emphasizes connections across control theory, reinforcement learning and stochastic approximation, enabling students to access these broader themes and start to work on cross-cutting projects. The school will be at a preparatory PhD level, suitable for advanced Master's and starting PhD students in these areas. The Spring School is part of a Research Semester Programme at CWI, with further workshops planned in 2025. Please see the full programme here: https://www.cwi.nl/en/events/cwi-research-semester-programs/control-theory-… Please apply for the spring school here: https://www.cwi.nl/en/events/cwi-research-semester-programs/spring-school-c… If you have any questions about our Spring School, please contact events(a)cwi.nl . We look forward to seeing you. On behalf of the organizers of the Research Semester Programme: Control Theory and Reinforcement Learning: Connections & Challenges.

1 0

Fwd: RL Workshop in Mannheim
by Tim van Erven 13 Dec '24

13 Dec '24

Forwarding on behalf of Mathias Staudigl (note the excellent line-up of speakers): -------- Forwarded Message -------- Subject: RL Workshop in Mannheim Date: Tue, 10 Dec 2024 07:24:20 +0100 From: Mathias Staudigl <mathias.staudigl(a)uni-mannheim.de> To: Tim van Erven <tim(a)timvanerven.nl> Dear colleagues, On Feb 7th 2025 we are running a workshop on reinforcement learning (sample based stochastic control) at the University of Mannheim. Speakers are • Peter Dayan (MPI Tübingen) • Daniele Calandriello (Deepmind) • Kassraie Parnian (ETH Zürich) • Alberto Maria Metelli (Politecnico di Milano) • Michael Muehlebach (MPI Tübingen) • Gergely Neu (Pompeu Fabra Barcelona) • Ciara Pike-Burke (Imperial College London) • Davide Tateo (TU Darmstadt) This one-day workshop will give interested researchers the possibility to get an overview of the current state of the field. Everyone is welcome to attend! https://www.wim.uni-mannheim.de/doering/conferences/rl-2025 No workshop fees are charged. Best wishes Leif Döring Professor Mathias Staudigl, PhD Chair for Mathematical Optimization Universität Mannheim Fakultät für Wirtschaftsinformatik und Wirtschaftsmathematik B6, C 306 | 68131 Mannheim Tel: +49 621 181 2545 E-Mail: mathias.staudigl(a)uni-mannheim.de Web: https://sites.google.com/view/mathiasstaudigl

1 0

Special issue SPL on e-values - deadline extended to February 1st
by Heide, Rianne de (UT-EEMCS) 02 Dec '24

02 Dec '24

Call for papers: deadline extended Special issue in Statistics and Probability Letters Guest editors: Prof. Jelle Goeman and Dr. Rianne de Heide Statistics and Probability Letters is a venue for very short papers up to six pages, short notes that may be a difficult fit with other journals. When preparing a submission, you can use the provided elsarticle latex template with setting final. E-values and Multiple Testing E-values have recently emerged as a new perspective on statistical hypothesis testing, allowing anytime-valid inference. But what if we have more than one hypothesis of interest? What multiple testing methods translate well (or not so well) from p-values to e-values? How can the e-value perspective enrich multiple testing? Though some seminal contributions have already made in this area, much remains to be explored. We accept submissions about e-values with a tangential link to multiple testing. Manuscript Submission Deadline: February 1, 2025 https://www.sciencedirect.com/journal/statistics-and-probability-letters/sp…

1 0

Call for submissions: LeaLog@CSL workshop
by Balder ten Cate 29 Nov '24

29 Nov '24

*Call for Submissions* *CSL 2025 Workshop on Learning and Logic (LeaLog@CSL)* https://sites.google.com/view/lealog25 Amsterdam, Netherlands 10 February 2025 Co-located with the 33rd EACSL Annual Conference on Computer Science Logic (CSL 2025) =================== About the workshop =================== The CSL 2025 Workshop on Learning and Logic is an on-site event happening on the 10th of February 2025 in Amsterdam, Netherlands. It will take place as part of the 33rd EACSL Annual Conference on Computer Science Logic (CSL 2025). The workshop brings together researchers who are working on topics at the intersection of learning and logic, ranging from the logical foundations of learnability and computational learning theory to logical analyses of machine learning models and applications of machine learning in knowledge representation and reasoning. The workshop will consist of invited talks and contributed talks. It does not have any proceedings, and therefore previously published or ongoing work are both encouraged to be presented. Topics for the presentation at the workshop include, but are not limited to, the following: - Logical analysis of machine learning architectures - Techniques for learning logical concepts - Computational learning theory - Logic for formal learning theory - Informational complexity of learning - Graph learning - Applications of ML in knowledge representation and data management - Neuro-symbolic integration - Statistical relational AI - Logical and epistemic aspects of distributed learning - Logical aspects of learning in multi-agent systems - Logical analysis of (iterated) belief dynamics and information change - Logical techniques for explainable AI - Data-driven techniques for temporal logic specification and verification ============================ Submissions ============================ Submissions consist of a title, a short abstract, and an extended abstract in the form of a PDF file (one page, excluding references). The link to the submission form can be found on the workshop website. ============================ Important dates and information ============================ Submission deadline: 8 January 2025 (Anywhere on Earth) Notification: 15 January 2025 * Event: 10 February 2025 * accepted submissions will receive a chance to register for the workshop and/or for CSL by January 19 without paying late registration fee. ============= Invited Speakers ============= - Alexandru Baltag (ILLC, University of Amsterdam) - Johan van Benthem (University of Amsterdam, Tsinghua University, and Stanford) - Dana Fisman (Ben-Gurion University) - Martin Grohe (RWTH Aachen) - Kristin Yvonne Rozier (Iowa State University) ============= Organisers ============= Steffen van Bergerem (Humboldt University of Berlin) Balder ten Cate (ILLC, University of Amsterdam) Aybüke Özgün (ILLC, University of Amsterdam) Sonja Smets (ILLC, University of Amsterdam) -- Tradeable cent : Traceable dent : Tentacle beard : Tentacled bear A better candle : Bed-care talent : Celebrated tan : Tractable need Decent, able rat : Able, tender cat : Central debate : Balder ten Cate

1 0

BeNeRL Reinforcement Learning Seminar: Tal Daniel from Technion (November 14, this Thursday)
by Yang, Z. (Zhao) 11 Nov '24

11 Nov '24

Dear colleagues, Our next BeNeRL Reinforcement Learning Seminar (Nov. 14) is coming: Speaker: Tal Daniel (https://taldatech.github.io<https://taldatech.github.io/>), PhD student from Technion. Title: Particles to Policies: Object-Centric Learning in Pixel-Based Decision Making Date: November 14, 16.00-17.00 (CET) Please find full details about the talk below this email and on the website of the seminar series: https://www.benerl.org/seminar-series The goal of the online BeNeRL seminar series is to invite RL researchers (mostly advanced PhD or early postgraduate) to share their work. In addition, we invite the speakers to briefly share their experience with large-scale deep RL experiments, and their style/approach to get these to work. We would be very glad if you forward this invitation within your group and to other colleagues that would be interested (also outside the BeNeRL region). Hope to see you on November 14! Kind regards, Zhao Yang & Thomas Moerland VU Amsterdam & Leiden University —————————————————————— Upcoming talk: Date: November 14, 16.00-17.00 (CET) Speaker: Tal Daniel (https://taldatech.github.io<https://taldatech.github.io/>) Title: Particles to Policies: Object-Centric Learning in Pixel-Based Decision Making Zoom: https://universiteitleiden.zoom.us/j/65411016557?pwd=MzlqcVhzVzUyZlJKTEE0Nk… Abstract: In visual multi-object manipulation decision-making tasks, understanding object interactions from pixel observations is a foundational challenge for real-world applications of reinforcement learning and robotics. Recent advances in unsupervised object-centric representations offer a promising path, as they decompose the image-based inputs into interpretable structures that naturally align with multi-object manipulation. In this talk, we explore recent advances that utilize object-centric representations for more efficient policy learning. We begin by introducing Deep Latent Particles (DLP), a representation that leverages keypoint-based object discovery to capture spatial and appearance features in a compact, interpretable form. Next, we discuss the application of these representations in model-free reinforcement learning, demonstrating how an Entity-Centric RL approach enables efficient learning for tasks requiring multi-object interaction and the emergence of compositional generalization. Extending these ideas, we present EC-Diffuser, a diffusion-based behavioral cloning framework to learn policies from demonstrations, achieving zero-shot generalization across complex object configurations. Finally, we introduce Deep Dynamic Latent Particles (DDLP), which extends DLP to video prediction, providing a bridge to model-based RL by predicting object dynamics over time. Bio: Tal Daniel is a final-year Ph.D. student in the Electrical and Computer Engineering faculty at the Technion, where he earned his B.Sc. and M.Sc., under the supervision of Prof. Aviv Tamar, and an incoming postdoc at Carnegie Mellon University (CMU). He is the winner of The Miriam and Aaron Gutwirth Memorial and The Irwin and Joan Jacobs Ph.D fellowships. His research interests include unsupervised representation learning, generative modelling and reinforcement learning.

1 0

CfP: International Symposium on Intelligent Data Analysis (IDA 2025)
by Krempl, G.M. (Georg) 08 Nov '24

08 Nov '24

[ If you receive this e-mail multiple times, please feel free to submit multiple papers ] ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ Call for Ideas The 23rd International Symposium on Intelligent Data Analysis (IDA 2025) taking place May 7th-9th, 2025, in Konstanz, Germany. Submission deadline: November 22nd, 2024 Notification of acceptance: January 31st, 2025 Camera ready submission: February 21st, 2025 All dates are specified as 23:59 SST (Standard Samoa Time / Anywhere on Earth) Lake Constance, Germany http://ida2025.org ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ Advancing Intelligent Data Analysis requires novel, potentially game-changing ideas. IDA’s mission is to promote ideas over performance: a solid motivation can be as convincing as an exhaustive empirical evaluation. Therefore, IDA accepts inspiring papers for both presentation and publication. The IDA symposium is intentionally small-scale and single-track to create an open atmosphere that encourages discussion. In addition, IDA awards the Frontier Prize every year, recognizing the most visionary contributions. Submissions will undergo a single-blind review process; author identities are known to reviewers. The conventional reviewing process favors incremental advances on established work and can discourage the kinds of papers that IDA 2025 aims to publish. The reviewing process will address this issue explicitly: referees will evaluate papers based on novelty, technical quality, potential impact, and clarity. Furthermore, each submission will be reviewed by one of the senior program committee (SPC) members. Any paper for which an SPC makes a convincing argument about how it addresses the symposium’s goals will be accepted independent of the overall review score. All accepted paper contributions will be published in the Lecture Notes in Computer Science series by Springer-Verlag. --- FOLLOW IDA --- Website: https://www.ida2025.org Instructions: https://ida2025.org/regular-paper-track/ Submission: https://cmt3.research.microsoft.com/IDA2025 Facebook: https://www.facebook.com/IDAsymposia/ Twitter: https://twitter.com/@ida_symposia

1 0

CWI seminar Jelmer Wolterink
by Wouter Edeling 07 Nov '24

07 Nov '24

Dear all, The next talk in our seminar series on machine learning and UQ in scientific computing will take place on Tuesday November 19 at 11h00 CET. Here, Jelmer Wolterink from the University of Twente will present his work on incorporating symmetries into Scientific Machine Learning models for hemodynamics. For those at CWI, the location will be L016, and for online attendees a zoom link is found below. Feel free to share this link with others. Kind regards, Marius Kurz, Wouter Edeling Topic: CWI-SC seminar Jelmer Wolterink Time: Nov 19, 2024 11:00 AM Amsterdam Join Zoom Meeting https://cwi-nl.zoom.us/j/85486177209?pwd=4BLXpgKgPgThRm9uJjYeZbs8bMdxL9.1 Meeting ID: 854 8617 7209 Passcode: 813381 Exploiting Symmetries for Personalized Hemodynamics Modeling in Cardiovascular Disease Large-scale population studies, interactive visualization and diagnosis in the clinic, and medical device design require fast, reliable, and differentiable models for estimation of hemodynamic parameters such as velocity, pressure, and wall shear stress. Commonly used computational fluid dynamics (CFD) approaches are prohibitively time-consuming and complex, leading to interest in the use of deep learning-based methods as a surrogate for CFD. I will discuss how symmetries in imaging and hemodynamics allow us to train deep learning models for cardiovascular segmentation and personalized prediction of hemodynamics efficiently with small real-world data sets. I will touch upon several of our recent works in scale invariant and rotation equivariant artery segmentation for hemodynamics estimation, surrogate modeling for hemodynamic fields estimation, as well as practical considerations in working with small and diverse real-world data sets for learning and validation.

1 0

Postdoctoral position in mathematical statistics
by Pires da Silva Castro, Rui 04 Nov '24

04 Nov '24

Dear colleagues, The mathematical statistics group of the Department of Mathematics and Computer Science is searching for an excellent candidate for a two-year Postdoctoral position: https://tinyurl.com/yc2ccv5v Description: The mathematical statistics group of the Department of Mathematics and Computer Science is searching for an excellent candidate for a two-year Postdoctoral position. The Mathematical Statistics group, which is part of the Statistics, Probability and Operations Research (SPOR)<https://spor.win.tue.nl/> cluster, focusses on developing novel statistical models and methodology endowed with theoretical guarantees, for the data-driven analysis of complex and high-dimensional systems. Our research topics range from heavily data-driven to fundamental and methodological aspects, including causal learning, dependence structure models, inference in complex networks, high-dimensional and non-parametric statistics and sequential decision making and data collection. The group also collaborates closely with the Teaching & Research Institute for Data Science Analytics (TRI-DSA). As the successful candidate, you are expected to pursue independent research within your own agenda, provided these strengthen and complement the expertise in the group. As the position is not attached to a specific project there is considerable freedom to independently pursue your own research agenda, as long as it has a good embedding with the research activities of the group. The position comes with light educational duties for about 25% of your working time (this percentage includes preparation time as well). Your educational tasks might include tutorial sessions and office hours, grading exams and homework, student supervision, and lecturing. Job requirements: As an ideal candidate you should hold a PhD degree in mathematics or mathematical statistics, clearly demonstrating your ability to conduct independent research on topics in mathematical statistics and/or learning theory. A different PhD background might be considered, provided you can showcase your ability to conduct high quality academic research, reflected in demonstrable outputs in mathematical statistics and/or learning theory venues. Strong interest in teaching and education is also a requirement for this position. Prior teaching experience is desirable, but not specifically required. As an ideal candidate, you must have good organization and communication skills, be highly self-motivated and independent, and be able to work in a team. Proficiency (written and verbal) in English is also required. Application and further information: https://tinyurl.com/yc2ccv5v Best regards, Rui Castro

1 0

2024

2023

2022

2021

2020

Machine-learning-nederland