The April edition of SEA involves IR & NLP in Gov in, with one speaker. This is a hybrid event.
Location: Lab42, Science Park Amsterdam, Room L3.36.
The Zoom link, in case you want to join online, will be available when you RSVP.
Speaker: Anne Schuth
Title: Automated Extraction of Machine-Executable Legislation
Abstract: The transformation of legal texts into executable code has traditionally been a manual, error-prone process performed by programmers with limited legal expertise. This disconnect between legislation and implementation creates significant challenges for government IT systems. In this talk, I'll present our information retrieval approach to automatically extracting machine-executable specifications from legal texts. Using LLMs, we convert natural language legislation into declarative specifications that can be directly executed by our interpretation engine. I'll demonstrate our system's capability to automatically identify cross-references between laws, extract logical conditions and requirements, extract scenarios, and generate executable representations that preserve legal semantics. Our evaluation across multiple Dutch laws shows how this approach creates traceable information flows from legal source texts to execution outcomes. For the IR community, this work highlights novel challenges in specialized domain extraction: dealing with legal references and dependencies, mapping natural language conditions to logical operations, and maintaining semantic fidelity while creating executable models
Bio: Anne Schuth is an Engineer at the Bureau Architecture for the Dutch Government, where he focuses on AI and engineering solutions for digital government. Prior to his government work, Anne held leadership positions at Spotify as ML Engineering Manager for Search and at DPG Media as Head of News Personalisation. He was a research scientist at Google and an ML Engineer at Blendle. He received his PhD from the University of Amsterdam under the supervision of prof.dr. Maarten de Rijke, with his research focusing on online learning to rank for information retrieval. Anne is a familiar face at the SEA, having previously organized the meetups.
Counter: SEA Talk #279.
