This edition of Supercomputing Thursdays will showcase how synthetic data can be used to train Natural Language Processing algorithms in areas lacking enough original data.
Training artificial intelligence (AI) models requires large amounts of data. This might be a limiting factor in fields where data is scarce, of poor quality or difficult to use due to privacy protection legislation. One way of overcoming these difficulties is using synthetic data – artificial data generated from original data that should deliver very similar results.
Our guest speakers from Slovakia’s National Competence Centre for Supercomputing will showcase how synthetic data can be used to train and enhance Natural Language Processing algorithms.
They will focus on three concrete use cases:
- the development of chatbots
- the development of voicebots
- religious text analysis