Machine Learning Group – Prof. Dr. Alan Akbik
Welcome to the Machine Learning Group of the Humboldt-Universität zu Berlin!
Our group focuses on research in machine learning (ML) and natural language processing (NLP). We aim to give machines the ability to understand and use human language. To achieve this, our group develops the Flair framework for automated text analysis. It is already used in hundreds of research projects and industrial applications and is frequently listed among the most popular deep learning frameworks for NLP.
On these pages you can find more information about our group:
Flair
The Flair framework for state-of-the-art NLP
Our research
Transfer learning, universal representations, etc.
Our teaching
Information on courses for students
News
- 23.09.2024 - Senior Area Chair (NAACL 2025): Prof. Akbik will be Senior Area Chair for Information Extraction at NAACL 2025!
- 20.09.2024 - Two papers accepted (EMNLP 2024): Two papers "Less is More: Parameter-Efficient Selection of Intermediate Tasks for Transfer Learning" and "NoiseBench: Benchmarking the Impact of Real Label Noise on Named Entity Recognition" accepted to EMNLP 2024 main conference!
- 13.09.2024 - Paper accepted: Our paper "HunFlair2 in a cross-corpus evaluation of named entity recognition and normalization tools" accepted to Bioinformatics journal!
- 11.09.2024 - New paper: Our paper "TransformerRanker: A Tool for Efficiently Finding the Best-Suited Language Models for Downstream Classification Tasks" now on ArXiv!
- 28.08.2024 - New paper: Our paper "LM-Pub-Quiz: A Comprehensive Framework for Zero-Shot Evaluation of Relational Knowledge in Language Models" now on ArXiv!
- 01.07.2024 - New lab member: Dr. Pieter Delobelle -expert for Dutch-language LLMs (and beyond)- joins us as visiting researcher. Welcome to Berlin! We look forward to working together on Dutch/German/English LLMs!
- 01.07.2024 - Senior Area Chair (ACL 2025): Prof. Akbik will be Senior Area Chair for Language Modeling at ACL 2025!
- 04.06.2024 - Paper accepted (ACL 2024): Our demo paper "Fundus: A Simple-to-Use News Scraper Optimized for High Quality Extractions" accepted to ACL 2024!
- 16.05.2024 - Paper accepted (ACL 2024): Our paper "Choose Your Transformer: Improved Transferability Estimation of Transformer Models on Classification Tasks" accepted to ACL 2024!
- 05.04.2024 - New research grant (IBB): The Investitionsbank Berlin (IBB) approved a new research grant for a Forschungstransfer project to industry partner PIKK Systems!
- 27.03.2024 - New startup grant: The German Federal Ministry for Economic Affairs (BMWK) approved an EXIST startup grant for FactorizeBio!
- 21.03.2024 - Paper accepted (NAACL 2024): Our paper "OpinionGPT: Modelling Explicit Biases in Instruction-Tuned LLMs" accepted to NAACL 2024: System Demonstrations!
- 13.03.2024 - Paper accepted (NAACL 2024): Our paper "BEAR: A Unified Framework for Evaluating Relational Knowledge in Causal and Masked Language Models" accepted to NAACL 2024!
- 01.03.2024 - New research grant (BMBF): The German Federal Ministry of Education and Research (BMBF) approved a new research grant for a 3-year project together with industry partner dida Datenschmiede!
- 20.02.2024 - Paper accepted (COLING 2024): Our paper "PECC: Problem Extraction and Coding Challenges" accepted to COLING 2024!
- 31.01.2024 - New paper: Our paper "SemScore: Automated Evaluation of Instruction-Tuned LLMs based on Semantic Textual Similarity" now on ArXiv!
- 18.01.2024 - Two papers accepted (EACL 2024): Our full papers "Large-Scale Label Interpretation Learning for Few-Shot Named Entity Recognition" and "Parameter-Efficient Fine-Tuning: Is There An Optimal Subset of Parameters to Tune?" accepted to EACL 2024!