Building the foundation for applied AI

Read More

Latest News from our initiative

  • Twelve IRM microtubule images, some recorded in the lab and some generated by DiffuMT

    DiffuMT Paper Accepted as Spotlight at NLDL 2027

    Our paper on evaluating mask-conditioned diffusion for microtubule microscopy has been accepted as a spotlight at the Northern Lights Deep Learning Conference (NLDL 2027).
    Read more
  • Illuminated Visions4AI letters on a stage, with the event logo projected behind them

    Manuel Weiss Presents His Research at Visions4AI 2026 in Hannover

    At Visions4AI 2026 in Hannover, Manuel presented a learning-based pipeline that lets robots detect and classify collisions using only their own joint signals.
    Read more
  • Example Raman spectra from the four application domains of RamanBench

    NeurIPS Contribution: RamanBench Sets a New Standard for AI in Raman Spectroscopy

    The paper has been accepted at NeurIPS 2026.
    Read more

  • Arnold Schwarz Presents His Research at the GCPR-2026 Conference

    We present a novel approach to infer contact forces from thermal imaging
    Read more

  • Summer School 2026 in Schorfheide

    Research, Network, Grow – The 2026 Summer School of the DFG-funded "Berlin Initiative for Applied Foundation Model Research"
    Read more
  • ICRA 2026 Vienna banner announcing Manuel Weiss's paper presentation

    Manuel Weiss Presents His Research at the ICRA-2026 Conference

    We embed a Lagrangian neural network in a computed torque controller for accurate, data-efficient robotic trajectory tracking.
    Read more
  • The article page in PLOS Computational Biology

    Microtubule Segmentation Paper Published in PLOS Computational Biology

    Our paper on human-grade microtubule segmentation with synthetic data and foundation models is out in PLOS Computational Biology. It is the result of a collaboration between HTW Berlin, BHT and the Max Planck Institute for Infection Biology.
    Read more
  • Jan Frick discussing the CliniBench poster with a visitor in the poster hall at EACL 2026

    CliniBench Presented at EACL 2026

    Our CliniBench paper, a benchmark comparing generative LLMs and encoder-based classifiers for diagnosis prediction, was accepted at EACL 2026 and presented as a poster in Rabat, Morocco.
    Read more
  • Team picture at the awarding of the right to award doctorates

    DFG visits BHT: Discussion on research and funding

    DFG representatives visited BHT to dive into our DFG-funded Research Impulse, explore advanced labs, and engage directly with researchers driving innovation across science and technology.
    Read more
  • Participants of the research meeting listening to a presentation in the robotics lab

    Research Meeting with Poster Session

    Our researchers presented their work in a poster session, and Scientific Advisory Board member Stephan Saalfeld came from the US to join us.
    Read more
  • Mario Koddenbrock holding the KI 2025 Best Paper Award certificate at the award ceremony, next to three other conference participants

    Best Paper Award at KI 2025

    Our paper on the domain robustness of contrastive vision-language models won the Best Paper Award at KI 2025, the German Conference on Artificial Intelligence.
    Read more
  • Team picture at gaining of the right to award doctorates

    BHT and HTW Berlin Receive the Right to Award Doctorates

    BHT and HTW Berlin are the first universities of applied sciences in Berlin allowed to award doctorates. All principal investigators of Appl-FM can now supervise doctorates.
    Read more
  • Group photo of the Summer School 2025 participants in front of a forest

    Summer School 2025 in Schorfheide

    On 21 and 22 May 2025, researchers from BHT and HTW Berlin met in Schorfheide for two days of talks, discussions and exchange at the Appl-FM Summer School.
    Read more
  • Felix Biessmann presenting on realistic error models at the project meeting at the Einstein Center Digital Future

    Project Meeting at the Einstein Center Digital Future

    Felix Biessmann presented on realistic error models for tabular data at the Appl-FM project meeting at the Einstein Center Digital Future in Berlin.
    Read more
See all