معرفی
Justus-Jonas Erker is a PhD Candidate in Natural Language Processing at the UKP Lab, Technische Universität Darmstadt, supervised by Dr. Nils Reimers and Prof. Dr. Iryna Gurevych. His research focuses on distributed representation learning, particularly in dialogue systems and retrieval-augmented generation. Key areas include sentence-transformers, emergent properties via cosine similarity, and efficient NLP techniques.
Research Interests:
- Representation Learning
- Dialogue Modeling
- Retrieval Augmentation
- Efficient NLP Algorithms
Recent Work: His 2024 ACL paper introduced Triple-Encoders for efficient dialogue representation, while his 2023 work on Curved Contrastive Learning demonstrated novel approaches to dialogue planning. In 2022, he contributed a foundational cancel culture corpus for sociolinguistic NLP analysis.
Affiliations: Active contributor to open-source projects (GitHub: 15 repos), including libraries for sentence-transformers and dialogue modeling. Engaged in academic outreach via Twitter/X and LinkedIn.
