Pre-EMNLP International Networking Symposium in Darmstadt

This October, Darmstadt becomes a stopover for the international NLP community on its way to EMNLP. Zuse School ELIZA and the ELLIS Unit Darmstadt are co-organising a pre-EMNLP international networking symposium together with the Ubiquitous Knowledge Processing (UKP) Lab at TU Darmstadt, bringing leading researchers in natural language processing, machine learning, and AI to the city for two days of talks, discussion, and exchange before the main conference begins.

What is EMNLP?

The Conference on Empirical Methods in Natural Language Processing (EMNLP) is one of the two flagship annual conferences of the Association for Computational Linguistics (ACL) and, alongside ACL itself, one of the most competitive and closely watched venues in the field. Each year it draws several thousand researchers working on how machines understand, generate, and reason over human language — from foundational model architectures to real-world deployments. A pre-conference symposium in Darmstadt gives the many researchers travelling through Germany on their way to the main event a chance to meet, present early ideas, and build collaborations in a smaller, more informal setting before the conference itself begins.

Why Darmstadt

TU Darmstadt’s UKP Lab, led by Prof. Dr. Iryna Gurevych, is one of Europe’s leading research groups in natural language processing, and Darmstadt is home to both the ELLIS & ELSA Unit Darmstadt and one of the seven sites of Zuse School ELIZA’s national network. Co-organising a symposium here puts that concentration of NLP and machine-learning expertise on display to an international audience — and gives ELIZA’s own fellows and students a direct, low-barrier way to engage with the wider EMNLP community without travelling to the conference itself.

What will be discussed

Advances in large language models and machine learning. Sessions will cover the fast-moving frontier of LLM research and the machine-learning methods underpinning it.
Multimodal and multilingual AI for expert domains. A growing share of NLP research looks beyond generic text: systems that combine language with other data types (images, structured data, sensor input) and that work reliably across languages, particularly in specialised domains such as law, medicine, or the sciences, where errors carry real consequences.
AI for Science. Language and machine-learning models are increasingly used as research tools in their own right — helping scientists search literature, generate hypotheses, or process experimental data.
Societal implications of language technologies. As language technologies reach ever more people, the symposium will also make room for the harder, still-unresolved questions: how to keep these systems trustworthy, fair, and genuinely steered by human judgement rather than replacing it.
These threads come together under the symposium’s guiding theme: “The Future of Science: AI-First, Trustworthy, Human-Steered.”

Event details

When: Welcome dinner on the evening of October 21, main symposium on October 22–23, 2026
Where: TU Darmstadt, Wilhelm-Köhler-Saal (Raum 283), Hochschulstr. 1, 64289 Darmstadt
Hosted by: Prof. Dr. Iryna Gurevych (UKP Lab, TU Darmstadt)
RSVP: If you are interested in joining, please fill out this short form by September 9^th (late sign-ups possible until October 15th, as long as there are seats)