Events

HIDA Lecture: Agentic AI - Safe Autonomous Agents

Toward the end of 2025, autonomous AI agents made significant leaps in capability. Smarter coding agents, a growing ecosystem of integrations exposed as skills and Model Context Protocol (MCP) servers, and projects like OpenClaw have demonstrated how these components can be combined into self-evolving AI systems. 

Explanation-Aware Attacks and the Limits of using XAI

Modern (deep) learning methods used to lack understandable reasoning in their decision process, making crucial decisions less trustworthy. The research field of "Explainable AI" (XAI) has turned the tables, though, enabling precise relevance attribution of input features for otherwise opaque models.

BLISS AI Speaker Series #35: "Beyond Patches: Learning Dense Visual Features"

Modern vision foundation models are highly capable and used as feature extractors in virtually any modern computer vision project. However, their patchified outputs inherently limit performance on dense, pixel-wise tasks. This talk presents strategies to learn and optimize dense visual features beyond these standard patches. We first introduce DIY-SC, a framework that leverages pseudo-labels to significantly improve pretrained models for correspondence tasks, while maintaining the original backbone's generalizability. 

07/2026 BIFOLD Colloquium

Gustavo Alonso, full professor at the Department of Computer Science at ETH Zürich, will talk about the role that data streaming plays in architecture and system design, especially in the context of heterogeneous architectures.

Lunch Talk: Dominik Naumann & Theekshana Dissayanake

In this Lunch Talk on "Toward AI-Enabled Cardiovascular Phenotyping in the Vehicle: The AutoHealth Study" Dominik and Theekshana will present the AutoHealth study. This study shows how in-car data collected across five cardiovascular cohorts contains sufficient physiological information to distinguish disease groups which enables AI-driven risk assessment.

Molecular Simulation in the Age of Artificial Intelligence

Machine learning and deep neural networks boost the efficiency of ab-initio molecular dynamics simulations, making possible first-principles predictions where previously only simplified semiempirical models could be used. Two examples at vastly different physical scales will be employed to illustrate the concept.

08/2026 BIFOLD Colloquium

Recent advances in remote sensing have led to an increase in the number of available foundation models; each trained on different modalities, datasets, and objectives, yet capturing only part of the vast geospatial knowledge landscape. While these models show strong results within their respective domains, their capabilities remain complementary rather than unified. 

Cedalion Workshop

This scientific workshop introduces researchers to Cedalion, a Python-based toolbox for advanced fNIRS and DOT analysis in naturalistic settings. Built on HOMER3 and AtlasViewer, it integrates machine learning and multimodal data processing for cutting-edge neuroimaging research. 

BEADS Summit

Join us for the BEADS summit on October 21st, 2026. Ten leading scientists from academia (Aalborg, CWI, Purdue, TU München) and industry (Databricks, DuckDB, Oracle, Relational AI, Snowflake) will offer talks on current developments in data systems.