Tuesday seminars
The Tuesdays seminar series is devoted to various topics relevant for the Integreat centre´s research focus. Presenters from the Integreat community and beyond have 40 minutes to present, followed by a group discussion. The seminars are held on Tuesdays during the weeks with even numbers.
Seminars are open for attendance in person at the Integreat premises and online (Zoom) for all.
Upcoming
Minimax-entropy neurosymbolic models
The evolving science of AI evaluation
Previous
Graph neural networks: asymptotics and unbounded expressivity
Building Biology In Silico: Spiking Neural Networks as Windows into Disease
Computing attribution scores for classification in explainable AI
Perception, reasoning and planning with multimodal models
How aligned language models break
From logical inference to reasoning in vector spaces
Belief change and ontology repair: two sides of the same coin?
Towards machine learning models that know what they do not know
Optimizing decision-making for long-term outcomes without delay
Data thinning and beyond
On space folding by neural networks
Reasoning under uncertainty with subjective logic
Inverse Inverse Reinforcement Learning
Efficient Likelihood-free Monte Carlo Sampling for Generalized Bayesian Inference Using Probabilistic Surrogate Loss
Simulation-based model selection for causal effect estimation
Shaping GNNs with dynamical systems
Using formal verification for text-to-SQL
Generalised mixed effects models for changepoint analysis of biomedical time series
High dimensional regression methods for inhomogeneous Poisson processes via split-variational inference
Causal inference in high-dimensional exploratory experiments
Register always matters: analysis of LLM pretraining data through the lens of language variation
The unseen frontier: pushing the limits of LLM sparsity with surrogate-free ADMM
Explainable AI for time series - a signal processing approach