Language technology

Language technology is a collective term for various technologies that use natural language as data.

Applications of language technology

We work with a number of language technology applications, such as

• deriving specific information from large volumes of text (text mining),
• classifying and categorising text documents,
• de-identifying sensitive personal information in texts,
• interactive systems, e.g. chatbots,
• speech technology, e.g. automated transcription,
• automated translation from one language to another.

Language technology models

One of our focus areas is developing language technology models from modest amounts of training data. This is particularly relevant for Norwegian language technology where language resources are scarce and international development is predominantly in English.

In addition, the technologies are increasingly used to analyse data that isn’t derived from languages, such as logs or industrial data.

ChatGPT

ChatGPT and other large language models (LLM) have been a significant part of our research since 2018, and we have extensive experience in training, adapting and fine-tuning neural language models to accommodate different tasks and demands.

Social robots

We often utilise social robots to evaluate and test prototypes using appropriate methodologies, and to measure learning experiences and effects. We can also examine various ethical issues and delve into deep neural networks that govern the way the robots process language.

Issues with language technology models

Language technology presents a range of challenges that require attention and resolution. Follow the links below to read more about the work we do at NR.