Natural Language Generation – Research Hub on NLG-Wiki.org
Welcome to the Natural Language Generation (NLG) research portal hosted at NLG-Wiki.org. This site is dedicated to collecting, organising and sharing up-to-date knowledge on how computers generate human-readable text from structured data, meaning representations, and other inputs. Whether you are a researcher, student or practitioner, you'll find comprehensive material covering theories, architectures and case studies in NLG.
What is NLG?
Natural Language Generation is the branch of artificial intelligence and computational linguistics that focuses on creating systems capable of producing fluent, coherent and context-appropriate natural language output. It involves converting internal representations — such as tables, semantic structures or numbers — into readable text in English or other human languages.
Unlike Natural Language Understanding (NLU), which tries to interpret human language, NLG must decide how to express known content in a way humans understand.
Why this portal exists
NLG research spans rule-based templates, statistical modelling, neural sequence-to-sequence systems and hybrid architectures. Researchers face issues such as content selection, lexical choice, linguistic realisation, style variation and handling of factuality or “hallucinations”.
NLG-Wiki.org compiles:
-
Surveys of state-of-the-art NLG methods
-
Links to downloadable NLG systems and datasets
-
Case studies of applications in industry, media, healthcare and chatbots
-
Bibliographies and conference proceedings such as INLG
Who should use this site
-
Computer scientists and linguists researching text generation
-
Engineers developing systems for summarising data, reporting, or dialogue
-
Students looking for reading lists, datasets or system descriptions
-
Practitioners applying NLG in business domains and seeking methodological insight
How to navigate
On NLG-Wiki.org you will find clear entry points:
-
Overview pages explaining core tasks and historically-important systems
-
Resources sections with links to tools, datasets, corpora and downloadable engines
-
Research threads addressing open challenges like controllability, factual accuracy, multilinguality and style transfer
-
Application domains covering areas from weather report generation to image-captioning and large-language-model output
Join the NLG community
NLG-Wiki.org is maintained with an open spirit. Contributors are welcome to add new systems, report on datasets, and discuss evaluation protocols or ethical issues such as bias and “hallucination” in generated text. This continually evolving platform ensures that researchers are connected to the latest developments in this fast-moving area.
Start exploring at NLG-Wiki.org — your gateway to the research, tools and discussions on natural language generation.