Welcome to the Machine Learning Human Behavior lab, where we explore the intersection of machine learning, neuroscience, and cognitive science to deepen our understanding of human behavior.
Our research focuses on understanding how humans think and act, using machine learning to model and predict behavior in real-world contexts and controlled experimental settings. By drawing inspiration from cognitive science, we aim to design AI architectures that mimic human-like learning and problem-solving. At the same time, we leverage AI tools to gain deeper insights into the complexities of human cognition and behavior.
We are also deeply committed to fostering collaborations with the humanities and social sciences. We believe that a comprehensive understanding of human behavior requires insights from these disciplines, particularly in areas such as ethics, culture, social dynamics, and human values. By partnering with researchers in psychology, philosophy, sociology, and other fields, we aim to build AI systems that are not only intelligent but also socially aware, ethically aligned, and adaptable to the diverse contexts in which humans live and interact.
- on May 24, 2025 at 9:41 pm
Die Topfavoriten Kanada und Schweden sind bereits raus, im Finale der Eishockey-WM haben die USA nun die Chance auf ihren ersten Erfolg seit 1960. Gegner Schweiz kann sogar zum ersten Mal den Titel gewinnen.
- on May 24, 2025 at 9:37 pm
Als Drittligist den DFB-Pokal holen und im nächsten Jahr als Zweitligist Europapokal spielen: Was wäre das für eine Geschichte gewesen! Ohne jeden Skrupel jedoch zerstörte Stuttgart die Fantasien der Bielefelder.
@durstewitzlab.bsky.social – DurstewitzLab Scientific AI/ machine learning, dynamical systems (reconstruction), generative surrogate models of brains & behavior, applications in neuroscience & mental health
- on May 20, 2025 at 2:15 pm
Can time series (TS) #FoundationModels (FM) like Chronos zero-shot generalize to unseen #DynamicalSystems (DS)? No, they cannot! But *DynaMix* can, the first TS/DS FM based on principles of DS reconstruction, capturing the long-term evolution of out-of-domain DS: https://arxiv.org/pdf/2505.13192v1 (1/6)
- on April 21, 2025 at 3:38 pm
I’m presenting our lab’s work on *learning generative dynamical systems models from multi-modal and multi-subject data* in the world-wide theoretical neurosci seminar Wed 23rd, 11am ET: www.wwtns.online –> incl. recent work on building foundation models for #dynamical-systems reconstruction #AI 🧪 https://www.wwtns.online/
- on February 9, 2025 at 2:18 pm
Our revised #iclr2025 paper and codebase for an architecture for foundation models for dynamical systems reconstruction is now online: https://openreview.net/pdf?id=Vp2OAxMs2s … includes additional examples of how this may be harvested for identifying drivers (control par.) of non-stationary processes. [contains quote post or other embedded content]
- on January 26, 2025 at 11:28 am
Toward interpretable #AI foundation models for #DynamicalSystems reconstruction: Our paper on transfer & few-shot learning for dynamical systems just got accepted for #ICLR2025 ! Previous version: arxiv.org/pdf/2410.04814; strongly updated version will be available soon … (1/4)
- on December 25, 2024 at 12:09 pm
That paper discusses an important issue for RNNs as used in neurosci. But we would argue that many RNN approaches do not truly reconstruct DS, for which we demand also agreement in long-term stats, attractor geometry, and generative perform. (esp. in chaotic systems, MSE as stats can be misleading). [contains quote post or other embedded content]
