About Advances in Natural Language Generation

Language generation (LG) is a crucial technology for human-machine communication. A good performance of language generation tasks implies (1) that machines are equipped with world knowledge that can require multimodal processing and reasoning (e.g. textual, visual and auditory inputs, or sensory data streams), and (2) the study of strong, novel machine learning (ML) methods (e.g. structured prediction, generative models), since virtually all state-of-the-art natural language processing (NLP) models are learned from data.

The aim of this collection is to bring together recent works related to developments and improvements within the field of language generation in a bid to advance scientific knowledge and encourage research into this field. The collection focuses on articles that are related but not limited to research concerning the following topics:

• Multimodal language generation
• Multilingual language generation
• Multitask language generation
• Creative language generation
• Paraphrase generation
• Translation of expressions, multiwords and phrasal units
• Linguistic resources such as datasets of paraphrases
• Association of expressions with identical or similar meaning
• Effective ways to add value to the MT technology
• Applications of language generation: machine translation, paraphrasing, text rewriting, cross-analysis of language varieties, summarisation, automatic feedback generation, text simplification, language teaching, etc.
• Grounded multimodal reasoning and generation
• Efficient machine learning algorithms, methods, and applications to language generation
• Dialogue, interaction and conversational language generation applications
• Large knowledge bases and graphs that can be used for language generation
• Common sense reasoning in language generation
• Applications of language generation in industry and society
• ChatGPT: opportunities, challenges, and threats
• Large language models and their generative power

The Collection has been developed by the Multi3Generation COST Action network (CA18231) but is open to submissions from Horizon projects.

Open Research Europe requires open access to research data supporting articles under the principle ‘as open as possible, as closed as necessary’. All articles should include citations to repositories that host the data underlying the results, together with any information needed to replicate, validate, and/or reuse the results/your study and analysis of the data. We recognize there may be exceptions due to ethical, data protection, or confidentiality considerations, or because the data have been obtained from a third party and access restrictions apply. Only research funded by Horizon 2020 and/or Horizon Europe is eligible for publication on Open Research Europe. All article processing charges will be covered centrally by the European Commission.