© 2021, Faktion. All rights reserved.
Transformers and their friends in NLP
This course is about using state-of-the-art techniques for building Natural Language Processing models
- Tokenisation, parsing, …
- Embeddings and transfer learning
- Labeling data correctly
Deep Learning Architectures
- Contextualized String Embeddings using the ELMo model
- Attention mechanisms
- Transformer architecture
- BERT model
- BERT’s derivatives and when to use them
Case studies and exercises
- E-mail automation
- Document information extraction
- NLP for search
Kaja Zupanc, PhD
Aleksandra Vercauteren, PhD
Data Scientist with a passion for NLP and a background as a researcher in Theoretical Linguistics. Specialized in formal approaches to typical aspects of linguistic interaction, such as question answering, implicit information, and discourse organization. Aleksandra loves tackling complex problems and finding patterns in unstructured data. A growing fascination for data science and its applications in Language Technology and AI led me to participate in a data science boot camp. I am an advocate of making science accessible to a broad audience and strongly believe that crowdsourcing is a crucial component of innovation.
- Python programming at the intermediate level
- Knowledge of basic TensorFlow concepts (we use the Keras API)
- Knowledge of basic Machine Learning concepts like data splitting, classification, overfitting, probabilities, ...