natural language generation
In plain wordsGetting computers to produce human language — from turning a spreadsheet into a weather report to writing an essay or a chatbot's reply.
The subfield of natural language processingnatural language processing, n.: The field that teaches computers to handle human language — to read, interpret, translate, summarize and produce text and speech. It is one of the oldest and largest areas of artificial intelligence. concerned with producing fluent, appropriate text (or speech) from some input: structured data, a knowledge base, a meaning representation, another text or a prompt. Classical NLG systems used a pipeline of content determination (what to say), document and sentence planning (how to organize and package it, including word choice and referring expressions) and surface realization (producing grammatical sentences), and were deployed for weather forecasts, financial reports and medical summaries where accuracy and control mattered. Neural sequence-to-sequencesequence-to-sequence model, n.: A model that turns one sequence into another, such as a French sentence into English or speech into text, by first reading the whole input and then writing the output step by step. models and then large language modelslarge language model, n.: A computer program trained on enormous amounts of text to predict the words that come next. At vast scale, that simple skill lets it chat, answer questions, translate, summarize and write code. replaced most pipelines with end-to-end generation, shifting the field's problems toward factualityfactuality, n.: How far what an AI says is actually true about the world. A factual model gets facts, numbers, names and dates right and does not make things up., faithfulness to the input, controllability, diversity and evaluation, for which BLEUBLEU, abbr.: An automatic score for machine translation that checks how many short runs of words in the computer's translation also appear in human reference translations., ROUGEROUGE, abbr.: An automatic score for summaries that checks how much of the wording in human-written reference summaries also appears in the computer-generated one., human judgment and model judgesLLM-as-a-judge, n.: Using one AI language model to grade or compare the answers of another, as a cheaper and faster stand-in for human reviewers. are used.
The bank's natural language generation system writes 30,000 portfolio summaries overnight from the same planning rules.
Origin
Research on generating text from semantic representations dates from the 1960s and 1970s, and by the 1980s NLG had its own international workshops. Ehud Reiter and Robert Dale's 'Building Natural Language Generation Systems' (2000) codified the classic pipeline.
Related
- natural language understandingnatural language understanding, n.: The part of language AI concerned with working out what a piece of text means — who did what to whom, what the writer wants and what follows from it — rather than with producing text.
- natural language processingnatural language processing, n.: The field that teaches computers to handle human language — to read, interpret, translate, summarize and produce text and speech. It is one of the oldest and largest areas of artificial intelligence.
- large language modellarge language model, n.: A computer program trained on enormous amounts of text to predict the words that come next. At vast scale, that simple skill lets it chat, answer questions, translate, summarize and write code.
- text summarizationtext summarization, n.: Using a computer to shorten a document, meeting or set of articles into a brief version that keeps the main points.
- machine translationmachine translation, n.: Software that renders writing or speech in one language into another language, as online translation services and phone translation apps do.
- decodingdecoding, n.: The method a language model uses to turn its predictions into actual text — for example, always picking the most likely next word, or choosing somewhat at random among the likely ones.
- factualityfactuality, n.: How far what an AI says is actually true about the world. A factual model gets facts, numbers, names and dates right and does not make things up.
- natural language understandingnatural language understanding, n.: The part of language AI concerned with working out what a piece of text means — who did what to whom, what the writer wants and what follows from it — rather than with producing text.
Sources
- Reiter, E. & Dale, R. (2000). Building Natural Language Generation Systems. Cambridge University Press.
- Gatt, A. & Krahmer, E. (2018). Survey of the State of the Art in Natural Language Generation: Core Tasks, Applications and Evaluation. Journal of Artificial Intelligence Research 61, 65–170.
Mentioned in
natural language processingnatural language processing, n.: The field that teaches computers to handle human language — to read, interpret, translate, summarize and produce text and speech. It is one of the oldest and largest areas of artificial intelligence., conversational AIconversational AI, n.: The technology that lets computers hold conversations with people in everyday language, by text or voice. It powers chatbots, voice assistants and automated customer-service lines., natural language understandingnatural language understanding, n.: The part of language AI concerned with working out what a piece of text means — who did what to whom, what the writer wants and what follows from it — rather than with producing text., text summarizationtext summarization, n.: Using a computer to shorten a document, meeting or set of articles into a brief version that keeps the main points.