ConceptNet
In plain wordsA free, crowd-built network of everyday facts — 'a knife is used for cutting', 'rain makes things wet' — in many languages, which AI systems use to make sense of ordinary situations.
An open, multilingual knowledge graphknowledge graph, n.: A huge web of facts stored as connections between things — 'Paris is the capital of France', 'Marie Curie won a Nobel Prize' — that computers can search, combine and reason over. of everyday knowledge in which the nodes are ordinary words and phrases rather than logical symbols, linked by labeled relations such as IsA, PartOf, UsedFor, CapableOf, Causes, HasPrerequisite and AtLocation — for example, (knife, UsedFor, cutting). It merges volunteer contributions, above all from the Open Mind Common Sense project, with assertions gathered through online games and imported from resources such as WordNetWordNet, prop. n.: A hand-made map of English vocabulary in which words sharing a meaning are bundled together and the bundles are linked — recording, say, that a dog is a kind of animal and a wheel is part of a car., Wiktionary and DBpedia, weighting each by confidence. Unlike CycCyc, prop. n.: An AI system built by specialists who spent decades typing in millions of everyday truths, such as 'people cannot be in two places at once', hoping to give computers common sense., it uses informal natural-language nodes rather than a logical language, trading precision for coverage. ConceptNet has been used to build word embeddings that combine graph and text (ConceptNet Numberbatch), to construct commonsense reasoningcommonsense reasoning, n.: The everyday background knowledge and reasoning people take for granted — dropped glasses break, hungry people eat, nobody is in two places at once — which machines long struggled to match. benchmarks such as CommonsenseQA, and to supply background knowledge to question-answering and dialogue systems.
Origin
Grew out of the Open Mind Common Sense project, launched at the MIT Media Lab in 1999 by Push Singh and colleagues. Hugo Liu and Singh described ConceptNet in 2004; later versions by Catherine Havasi, Robyn Speer and collaborators extended it across languages (ConceptNet 3, 2007) and merged in other open resources (ConceptNet 5, from 2012).
Related
- commonsense reasoningcommonsense reasoning, n.: The everyday background knowledge and reasoning people take for granted — dropped glasses break, hungry people eat, nobody is in two places at once — which machines long struggled to match.
- knowledge graphknowledge graph, n.: A huge web of facts stored as connections between things — 'Paris is the capital of France', 'Marie Curie won a Nobel Prize' — that computers can search, combine and reason over.
- CycCyc, prop. n.: An AI system built by specialists who spent decades typing in millions of everyday truths, such as 'people cannot be in two places at once', hoping to give computers common sense.
- WordNetWordNet, prop. n.: A hand-made map of English vocabulary in which words sharing a meaning are bundled together and the bundles are linked — recording, say, that a dog is a kind of animal and a wheel is part of a car.
- WikidataWikidata, prop. n.: A free, shared collection of facts about millions of things — people, places, books, chemicals — kept by volunteers in the Wikipedia family and laid out so programs can use it as easily as people.
- knowledge graph embeddingknowledge graph embedding, n.: A way of turning the things and relations in a knowledge graph into lists of numbers, so a computer can guess missing facts from patterns, such as predicting where someone was born.
Sources
- Liu, H. & Singh, P. (2004). ConceptNet — A Practical Commonsense Reasoning Tool-Kit. BT Technology Journal 22(4), 211–226.
- Speer, R., Chin, J. & Havasi, C. (2017). ConceptNet 5.5: An Open Multilingual Graph of General Knowledge. Proceedings of AAAI-17.