All Blogposts

Here you’ll find all blog posts from the Lamarr Institute – across all topics and categories. Our authors share how AI is researched, applied, and discussed. Enjoy exploring and learning!

BlogApplication

CapyMOA: Efficient Machine Learning for Evolving Data Streams

Dr. Heitor Murilo Gomes,

25. February 2026

Real-time data keeps changing – your models should too. CapyMOA brings high-performance stream learning and concept drift handling to Python, bridging MOA efficiency with practical usability....

Illustration in a Van-Gogh-inspired style showing a calm humanoid robot observing a flowing stream of data that transforms into adaptive decision trees. Swirling blue and yellow brushstrokes depict real-time data streams, changing graphs, and evolving tree structures, symbolizing Streaming Gradient Boosting and online learning under concept drift.

BlogResearch

Streaming Gradient Boosting: Pushing Online Learning Beyond its Limits

Dr. Nuwan Gunasekara,

04. February 2026

Streaming Gradient Boosting (SGB) brings gradient boosting to evolving data streams—adapting models to concept drift in real time with strong accuracy and efficiency across classification and regression tasks....

00 Blog Beitragsbild HEROS - Lamarr Institute for Machine Learning (ML) and Artificial Intelligence (AI)

BlogResearch

Green Online Learning with Heterogeneous Ensembles

Kirsten Köbschall,

29. January 2026

What if greener AI could also be more accurate? Discover how heterogeneous online ensembles like HEROS boost performance, cut resource costs, and adapt instantly to evolving data streams....

Generated by ChatGPT via the prompt “A digital scene captures a humanoid robot, with a smooth white body and dark accents, operating a wooden waterwheel at a quaint watermill beside a flowing river under a clear sky. The landscape around it is painted in bold, swirling strokes reminiscent of Van Gogh's Post-Impressionist style, featuring lush greenery, rolling hills, and a winding path, with the robot and the rotating waterwheel as the focal points.”

BlogBasics

Learning From Data Streams: Foundations of Stream Learning

Dr. Sebastian Buschjäger,

21. January 2026

Data never stops — so how do models keep up? This opening post introduces stream learning, concept drift, and real-time adaptation for continuously evolving data streams....

BlogApplication

Fine-Tuning a Low-Resource Language: Training the Whisper Model on Marathi

Dr. Rishikesh Pandit,

12. November 2025

A wide gap still separates how well an English or German ASR model would transcribe especially on heavy accented or non-native speech and similar in case of Marathi or other low resource languages. Can we overcome this shortcoming? If yes, how?...

Digital representation of warehouse augmented with virtual blueprints of logistics systems

BlogApplication

From Sketches to Digital Warehouse Layout with the WarehouseDesigner

Christian Pionzewski, Maximilian Otten, Anike Murrenhoff,

29. October 2025

AI is changing warehouse planning: WarehouseDesigner transforms simple sketches into digital, optimized warehouse layouts in seconds, accelerating the planning of modern warehouse systems....

00 Blog Beitragsbild Fine Tuning LRL 1 - Lamarr Institute for Machine Learning (ML) and Artificial Intelligence (AI)

BlogResearch

Why Fine-Tuning ASR Still Fails Low-Resource Languages – and How We Can Fix It

Dr. Rishikesh Pandit,

15. October 2025

Fine-tuning ASR models remains complex for low-resource languages. Using Marathi, we explore the linguistic, technical, and ethical barriers – and how to overcome them for more inclusive speech technology....

AI label for MobileNetV3Small (rated A) shown next to a symbolic image of a green tree growing from a computer chip. The AI label highlights energy efficiency, robustness, and accuracy, promoting sustainable and trustworthy AI. ____ KI-Label für MobileNetV3Small (mit der Bewertung A) neben einem symbolischen Bild eines grünen Baumes, der aus einem Computerchip wächst. Das KI-Label hebt Energieeffizienz, Robustheit und Genauigkeit hervor und fördert damit eine nachhaltige und vertrauenswürdige KI.

BlogResearch

AI Labels – How They Can Provide Guidance and Build Trust

Katharina Poitz, Dr. Raphael Fischer,

23. July 2025

AI labels can bridge communication gaps and promote trust in AI. A user study shows that they are clear, comparable, and accessible to experts and people without a technical background, helping them to make sustainable, informed decisions....

IceCube astrophysics neutrinos lamarr - Lamarr Institute for Machine Learning (ML) and Artificial Intelligence (AI)

BlogBasics

How AI Helped to Spot Neutrinos from our Galaxy

Jana Konrad,

09. July 2025

The IceCube Neutrino Observatory could finally detect neutrinos from the Milky Way. That was possible thanks to new AI tools, which enabled a more accurate analysis of the observed events. ...