Make it (net)work! - MLA2S Networking Seminar #09

Monday 11 May 2026, 13:00 → 16:55 Europe/Vienna

9th Network Seminar of the Thematic Platform MLA²S. 
Seminarrooms 2 and 3 (formerly 1 and 2, also known as 3A.1 and 3A.2), PSK 3rd floor

 

Zoom connection:
https://oeaw-ac-at.zoom.us/j/65470482668?pwd=KVfK11wgchXcwVHRUFWKVNXNt09xmU.1

Meeting ID: 654 7048 2668
Passcode: 19B7NV

 

 

To register: click "register" below here. 

13:00 → 13:15 News from MLA2S and Make it (net)work!

Including a short round table news update.

Speakers: Dr Claudius Krause (MBI Vienna (ÖAW)), Claus Trost (Erich Schmid Institute of Materials Science of theAustrian Academy of Sciences), Jan Odstrčilík (IMAFO), Kati Heinrich (IGF | ÖAW), Nicki Holighaus (Acoustics Research Institute, Austrian Academy of Sciences)

13:15 → 13:40 Channelling the power of LLMs toward the study of antiquity

In the fall of 2025, the Decoding Antiquity project was launched by Anna Dolganov in collaboration with Mistral AI and SAIL Reply with the aim of building advanced AI systems for ancient languages. We have started with a multimodal LLM for Ancient Greek optimized for the restoration of fragmentary texts, and our next goal will be to train a Vision Language model capable of transcribing hundreds of thousands of undeciphered Ancient Greek papyri. We are also preparing to train a multimodal LLM for Latin on top of our Greek model to create a bilingual AI system. I present a preliminary report of our progress and the broader vision of the project.

Speakers: Anna Dolganov (ÖAI | ÖAW), David Smith (Khoury College of Computer Sciences, Northeastern University Boston)

13:40 → 14:05 Machine Learning–Enabled Identification of Archaeological Objects from Carnuntum (Austria): The Heritage Science Austria 2.0 LEGION Project

The project LEGION (machine LEarninG-enabled Identification of archaeological Objects in the middle daNube river basin) develops an AI-based framework for the automated identification and classification of archaeological objects, focusing on Roman common-ware pottery from Roman Carnuntum, part of the UNESCO World Heritage “Danube Limes.” Funded through the Heritage Science Austria 2.0 programme, the project builds on a unique legacy dataset of tens of thousands digitized pottery profile drawings produced during decades of archaeological research, which are systematically enriched with structured archaeological metadata including typology, manufacturing techniques, decoration, chronology, and fragment characteristics. By integrating state-of-the-art ML approaches with eXplainable AI (XAI) and human-in-the-loop (HITL) validation, LEGION aims to establish transparent and reproducible workflows for the analysis of large archaeological datasets. In doing so, the project demonstrates the potential of AI for the digital reuse and standardization of legacy archaeological documentation and opens new perspectives for large-scale analyses of pottery production, distribution, and consumption patterns in the Roman Middle Danube region.

Speakers: Dominik Hagmann (ÖAI | ÖAW), Irene Ballester-Campos (CVL | TU Vienna)

14:05 → 14:20 Break to walk a few steps and get some fresh air in.

14:20 → 14:45 Vibe Coding and Automated Data Harvesting: Bridging Archaeological Intuition with LLMs

In the evolving landscape of Digital Archaeology, Large Language Models (LLMs) are shifting from text generators to active co-developers. This presentation explores the practical application of "Vibe Coding" and AI-driven automation within the scientific workflow of the Digital Archaeology and Classics (DAC) department. We demonstrate how natural language descriptions can rapidly prototype specialized research software, as seen in our open-source tools on GitHub – including STOCHASI for artifact simulation, CombiTab for seriation analysis, and the HarrisMatrixEditor (HME) for stratigraphic visualization.
Beyond development, we address the "daily bread" of AI in research: the high-speed extraction and transformation of data. By scraping fragmented information from the web and distilling structured datasets from legacy publications, we transform unstructured knowledge into FAIR-compliant formats. This integration of AI-assisted methods lowers technical barriers while maintaining the rigorous standards required for modern archaeological data modelling.

Speaker: Christian Gugl (ÖAI | ÖAW)

14:45 → 15:30 The Evolution of Genome Regulation

Every cell in our body contains the same genetic information, yet multicellular species exist as highly complex organisms. This is driven by the dynamic control of DNA, largely mediated by proteins called histones, which assemble with DNA to form structures called nucleosomes. The positioning and composition of nucleosomes influence how the genome is read, and their conservation across the tree of life makes them a powerful window into how different species regulate their genomes. Using synthetic biology and machine learning, we build genome regulation systems from scratch to reconstruct two billion years of cellular history and generate insights with biotechnological and therapeutic potential.

Speaker: Irwin Nicholas (GMI | ÖAW)

15:30 → 16:30 Networking with refreshments / further discussions