Natural Language Understanding Lab

Natural Language Understanding (NLU) will be key for making the most of AI and enables tasks as wide ranging as using voice assistants at home to a reliable automatization of certain social services. To accelerate innovation, development of applications, and implementation in this area, NLU therefore constitutes one of AI Sweden’s strategic initiatives.

Do you want access to GPT-SW3?

AI Sweden together with RISE and WASP WARA Media & Language have built the first truly large-scale generative language model for the Nordic languages, and primarily Swedish.

Access GPT-SW3 here

If you are interested in becoming a partner of AI Sweden, and getting access to the partner benefits, including the Labs, please feel free to reach out.

What is Natural Language Understanding?

Natural Language Understanding (NLU) is a collective term for the applied aspects of how computer programs or electronic devices can analyze, produce, modify or respond to human text and speech. Natural Language Processing (NLP) is a core element of this area. NLP algorithms (language models) learn from language data. They enable machine understanding and machine representation of natural (human) language. Using NLP, applications and tools can be used to identify patterns and provide insights hidden in large amounts of language data, too vast for humans to process. Based on patterns and insights, unstructured data can be turned into actionable information, allowing us to make data-driven and well-informed decisions. In order for Sweden to benefit from the rapid development in the applied language technology area, language models need to be trained on datasets for Swedish and minority languages spoken in Sweden.

Insights from behind the scenes of AI

The AI Sweden NLU team’s aim with writing blog posts is to be as open and transparent as possible with the development process, and to invite several different perspectives on this type of model — both our own as well as those who may be more critical of this development.

Go to the blog

Projects

A woman standing by a desk in fromt of her computer

A shared digital assistant for the public sector

AI Sweden, in collaboration with partners, is developing a digital assistant for the public sector based on AI technology. The objective is to streamline text-based tasks, freeing up time for more...

Data Readiness Lab

The Data Readiness Lab provides a collaborative environment and contributes skills, tools, frameworks and resources to increase data readiness and data maturity among actors in both the public and...

DeployAI

The project DeployAI will build, deploy, and launch a fully operational AI-on-demand platform (AIoDP) promoting trustworthy, ethical, and transparent European AI solutions for use in the industry...

Photo in black and white of trucks on a parking lot seen from above

FormAI

FormAI explores the use of generative AI coupled with formal verification for developing secure and trustworthy software in the automotive domain.

GPT-SW3

AI Sweden, together with RISE and WASP WARA Media & Language, have developed a large-scale generative language model for the Nordic languages, primarily Swedish.

GPT-SW3 validation project

Can a large-scale Swedish generative language model reduce the threshold for using advanced language models to solve text processing tasks across a variety of sectors and use cases?

Interdisciplinary expert pool for NLU

This project aims to establish and test a new form of cross-disciplinary collaboration for accelerating applied AI in the context of AI-driven language technology: The Interdisciplinary Expert Pool...

Language Models for Swedish Authorities

The project develops large-scale language models for the Swedish public sector, enabling a number of applications with specific relevance to Swedish authorities.

An illustrative composition featuring multiple screens arranged in a circular formation, radiating light from the center

Multimodal language model

AI Sweden's language team is taking the next major step by initiating the development of Sweden's first large multimodal language model. The new model is expected, just like GPT-SW3, to become an...

A picture from a library setting showing shelves with books

Pilot study for federated language models in Swedish

AI Sweden joined the National Library of Sweden and Scaleout Systems for a pilot on federatively trained language models. This was the first federative, large-scale modeling of artificial neural...

A tranquil scene capturing a close-up of a serene lake and mountain view, with a dog peacefully lying on a wooden dock in the foreground

SuperLim

The SuperLim project forms a Swedish version of the English benchmarking platform (Super)GLUE. The first version of an evaluation framework for Swedish language models has now been released with a set...

Landscape image of a lake and red Swedish houses

Swedish Language Data Lab

A thoroughly developed base for Natural Language Processing (NLP) is one of the cornerstones of successful AI applications. NLP is one of AI Sweden's strategic areas, and the Swedish Language Data Lab...

Photo of a large amounts of text documents laying on a table, seen from above. AI-generated

TrustLLM

TrustLLM will develop European large language models (LLMs) on an unprecedented scale, trained on the largest amount of text so far in European AI, covering a range of underrepresented languages, and...

NLP Seminar Series

NLP Seminar Series is a bi-weekly forum for people who work with or are interested in Natural Language Processing (NLP) and language technologies. The seminars are organized by RISE NLP group and AI Sweden. Head on over to our channel on YouTube to watch the latest uploads.