Scholar
Michael Fromm
Google Scholar ID: NL5yVhYAAAAJ
Fraunhofer IAIS
Citations & Impact
Citations
433
H-index
10
i10-index
10
Publications
20
Co-authors
10
list available
Publications
Resume (English only)
Academic Achievements
- Current Foundation Model Projects:
- - TrustLLM: Democratize Trustworthy and Efficient Large Language Model Technology for Europe
- - openGPT-X: Development of Large European Language Models
- - EuroLingua-GPT: One Model for all European Languages
- Selected Research Articles:
- - Computer Vision:
- * Automated detection of conifer seedlings in drone imagery using convolutional neural networks
- * Unsupervised anomaly detection for X-ray images
- * XD-STOD: Cross-Domain Superresolution for Tiny Object Detection
- - Natural Language Processing:
- * Argument Mining Driven Analysis of Peer-Reviews
- * TACAM: topic and context aware argument mining
- * Diversity Aware Relevance Learning for Argument Search
- * Active Learning for Argument Strength Estimation
- * Cross-Domain Argument Quality Estimation
Research Experience
- - Principal Research Scientist for Foundation Models, Fraunhofer IAIS, 2023 - Present
- - Postdoc, LMU Munich, 2022 - 2023
- - Innovation Lab Coordinator, LMU Munich, 2020 - 2023
- - Research Scientist, LMU Munich, 2016 - 2023
- - Intern Software Development, Dassault Systemes, 2015 - 2016
Education
- LMU Munich
- - PhD in CS, 2018 - 2022
- - MSc in CS, 2016 - 2018
- - BSc in CS, 2012 - 2016
Background
- Research Interests: Multilingual Large-Language Models, Multimodal Foundation Models, Computer Vision, Remote Sensing.
Miscellany
- Skills: Data Processing with e.g. Polars, Pytorch, DataTrove; Programming in Python, Java, C++; Building foundation models in PyTorch and our own Framework.
- Teaching Experiences:
- - Introduction to Programming
- - Argument Mining
- - Big Data Management and Analytics
- - Practical Machine Learning
- - Seminar Machine Learning
- - Deep Learning and Artificial Intelligence
- - Artificial Intelligence for Games
- - Machine Learning with limited labeled Data
- - Supervision of dozens of bachelor & master thesis
