Anna Steinberg Schulten
NLP researcher turning long, complex documents into structured, verifiable data.
Third-year PhD Candidate · Social Data Science and AI (SODA) Lab, Department of Statistics, LMU Munich
Advisors: Prof. Frauke Kreuter (LMU Munich), Prof. Alexander Fraser (TUM Heilbronn)
Affiliations: relAI (Konrad Zuse School of Excellence in Reliable AI), MCML (Munich Center for Machine Learning)
I study how large language models, including multimodal ones, can extract structured information from long, complex documents such as corporate sustainability reports and planning documents, and how we can tell whether the extracted data is right.
Before the PhD I studied Computational Linguistics (M.Sc., LMU Munich), Mathematics (M.Sc., RWTH Aachen) and Economics (B.Sc., University of Mannheim), and worked as a software and NLP engineer, including on the open-source education platform Serlo.
Research interests
LLMs · information extraction · multimodal LLMs · uncertainty quantification
Research
Note: A. Steinberg = A. Steinberg Schulten.
Peer-reviewed publications
- 2026A Practitioner's Guide to Systematic Prompt Evaluation for Information Extraction
EMNLP 2026 Industry Track, pp. 975–987
Steinberg Schulten, A., Schierholz, M., Reichenbach, L., Kreuter, F., Fraser, A. - 2026Climatextract: A Systematic Prompt Evaluation For Climate Information Extraction
ClimateNLP 2026, pp. 29–40
Steinberg Schulten, A., Schierholz, M., Reichenbach, L., Kreuter, F., Fraser, A. [code] - 2025Addressing data gaps in sustainability reporting: A benchmark dataset for greenhouse gas emission extraction
Scientific Data, 12, 1497
Beck, J.*, Steinberg, A.*, Dimmelmeier, A. et al. (*shared first authorship) [data] - 2026Beyond Dictionaries: Machine-Learned Normalization of Emission Units in Information Extraction
ClimateNLP 2026, pp. 132–142
Reichenbach, L., Oehler, S., Schierholz, M., Steinberg Schulten, A. [code] - 2026NLP for Social Good: A Survey and Outlook of Challenges, Opportunities, and Responsible Deployment
EACL 2026 (long paper)
Karamolegkou, A. et al. (incl. Steinberg, A.)
Working papers
- 2026Beyond Aggregate Performance: A Stratified Evaluation of Multimodal LLMs for Legal-Reference Extraction from German Land-Use Plans
Working paper
Ewudor, H., Sommer, F., Steinberg Schulten, A. - 2026LMU & TUM NLP at the NTCIR-19 SciClaimEval Task: Structured Evidence and Two-Step Operation Reasoning for Cross-modal Scientific Claim Verification
Working paper (NTCIR-19 SciClaimEval Shared Task)
Steinberg Schulten, A., Fraser, A. - 2026Automated reproducibility assessments in the social and behavioral sciences using large language models
Preprint, under review at Nature Human Behaviour
Feuerriegel, S., Holtdirk, T., Marcolongo, P., Steinberg Schulten, A., et al.
Other publications
- 2026Addressing Data Gaps in Climate Change and Sustainability Statistics using Generative AI
13th Biennial IFC Conference, Bank for International Settlements
Fehr, M., Oehler, S., Reichenbach, L., Riemer, M., Steinberg Schulten, A., Schierholz, M., Dimmelmeier, A., Sommer, F.
Talks & Presentations
- 2026Poster: A Practitioner's Guide to Systematic Prompt Evaluation for Information Extraction
EMNLP 2026 Industry Track, Budapest - 2026Panelist: AI & Climate Action: Part of the Solution or Part of the Problem?
Munich Climate Week (6 Oct), hosted by Celonis, Munich - 2025ClimXtract: An Open-Source Data Extraction Pipeline for Company-Level Greenhouse Gas Emissions
DAGStat 2025, Session 24 Text Mining & NLP - 2025Errors of and Reflection in LLM Pipelines
AI, Humanity & Society: 2nd LMU-TAU Workshop for Early-Career Researchers
Education
- 2024–PhD in Statistics, Social Data Science and AI Lab, LMU Munich
Main project: information extraction of sustainability metrics from corporate sustainability reports using large language models (project website) - 2020–M.Sc. in Mathematics, RWTH Aachen
Focus: Statistics, Minor: Computer Science. Coursework completed 2023, thesis outstanding. Fall 2021: semester abroad at NaUKMA, Kyiv - 2017–2020M.Sc. in Computational Linguistics, LMU Munich
Grade: 1.2 (German scale, 1.0 = best)
Minor: Computer Science. Master thesis: Summarization of Texts using a Pretrained Language Model - 2013–2016B.Sc. in Economics, University of Mannheim
Grade: 1.9
Major: Economics, Minor: Mathematics. Exchange semester at E.N.S.A.E. ParisTech, Paris (2015–2016)
Teaching
- Nov–Dec 2026Instructor (guest lecturer), AI-Assisted Content Analysis: Methods and Application in Area Studies
Leibniz Institute for East and Southeast European Studies (IOS), Regensburg - Aug 2025, Aug 2026Instructor (co-taught), Data Science Techniques for Survey Researchers
GESIS Summer School in Survey Methodology - Apr 2025 – Mar 2026Instructor, Introduction to statistical software in R
Department of Statistics, LMU Munich - 2016–2021Tutorial instructor, Applied Stochastics (RWTH Aachen); Formal Languages and Automata (LMU Munich); Linear Algebra IIA (University of Mannheim)
Scholarships & Awards
- 2026Honourable mention, Agentic AI for Science Hackathon (Machine Learning Summer School NYC, Columbia University)
- 2021–2022Metropoles in East Europe Scholarship, German Academic Scholarship Foundation & Alfried Krupp von Bohlen und Halbach Foundation
- 2017–2023Studienstiftung des deutschen Volkes (German Academic Scholarship Foundation)
- 2016–2017Franco-German Parliamentary Scholarship, French Ministry for National Education, Higher Education and Research
Beyond research
I love learning languages: I speak five languages at least fluently and am learning Spanish and Turkish. I also love to listen to and play music: I sing in a German-Ukrainian choir and play the horn in a local Symphonic Brass Orchestra in Cologne.
I also have a soft spot for owls.
Contact
anna.steinberg@lmu.de · GitHub · Google Scholar · LinkedIn · CV (PDF)