Dear dida follower,

We are excited to announce the upcoming second dida conference, scheduled for May 31st at B-Part Berlin. After the success of last year’s first dida conference, we are looking forward to seeing you again at a day full of machine learning and networking, focusing on topics such as Large Language Models and MLOps. Please register early, as the available free tickets are limited.

Apart from our dida conference, we have of course been busy developing new machine learning solutions, and would like to mention two new project case studies that you might find interesting:

In our latest project, "Anomaly Detection in Track Scenes", we have partnered with Deutsche Bahn as part of their "Initiative Digitale Schiene", focusing on ML-based solutions for their automated driving system for trains, in particular ensuring safety through the automatic detection of anomalous or hazardous objects. Unlike conventional systems that identify objects within some predefined classes (trees, persons, cars), our goal was to create a model capable of detecting any object and assessing its level of threat. Read the full project case study

here.

In our second project, "Automated Detection and Analysis of Tailings", we used various satellite images, a suite of computer vision models and mathematical modeling to detect, segment and analyze mining tailings with respect to their volume and mineralogical content. Our approach offers a promising alternative to traditional, labor-intensive exploration methods for the identification and analysis of remote mining locations. Furthermore, it is able to analyze mineral configurations by applying machine learning methods to multispectral satellite data. You can read the full project case study [here](https://dida.do/de/projekte/automatisierte-erkennung-und-analyse-von-minenschutthalden).

Furthermore, Philipp Jackmuth, our managing director, recently appeared on "The Data Science Podcast" by INWT Statistics. He and Dr. Amit Ghosh, managing director of INWT Statistics, had an interesting conversation about what makes machine learning projects successful. You can find this episode here.

Our CTO, Dr. Lorenz Richter, followed the invitation to speak at the Measure Transport, Diffusion Processes and Sampling Workshop at the Flatiron Institute in New York City. The workshop brought together experts on mathematical aspects of diffusion based generative modeling and was a great opportunity to exchange with colleagues. Lorenz talked about our works on optimal control and generative modeling.

Last but not least, let’s talk about Large Language Models. At dida, we more and more focus on the development of LLM-based solutions for our clients and partners and constantly follow the latest research developments. As it is likely that the topic of LLMs is currently also on your radar or even already onto your road map, we would like to invite you to have an initial 30min conversation with us, to discuss your LLM thoughts and give a first evaluation of your potential use cases. For a first LLM overview, you can also visit our introductory article here.

We hope you liked reading about our progress. All the best for you and your plans.

Best regards,
Lorenz and Philipp