abductive reasoning
In plain wordsReasoning from an observation to its most likely explanation — the grass is wet, so it probably rained — the kind of educated guess that doctors and detectives make.
Inference that runs from effects back to possible causes: faced with observations, it proposes a hypothesis that would account for them and, when several compete, prefers the one that is simplest, covers the most evidence and is most plausible beforehand. True premises do not guarantee its conclusion: from 'if it rained, the grass is wet' and 'the grass is wet' it proposes rain, though a sprinkler would also explain the observation. In AI, abduction is formalized as finding a set of assumptions that, together with background knowledge, entails the observations while remaining consistent — the basis of model-based diagnosis, plan and intention recognition, and abductive logic programminglogic programming, n.: A style of programming in which you write down facts and rules in logic and then ask questions; the computer works out the answers itself by logical deduction.; in probabilistic terms it corresponds to finding the most probable explanation in a Bayesian networkBayesian network, n.: A diagram of causes and effects with probabilities attached — say, how rain and a sprinkler can each make grass wet — that lets a computer update its beliefs about hidden causes when it sees evidence.. Scientific hypothesis generation, debugging and the interpretation of language all involve abduction, and it is a common test of 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. reasoningreasoning, n.: Working out conclusions step by step from facts, rules or evidence, as in solving a math problem or planning a trip. Whether today's AI truly reasons, or imitates reasoning, is debated..
The monitoring agent abduced a failing disk as the simplest explanation for both the latency spike and the checksum errors.
Origin
Named by the philosopher Charles Sanders Peirce, who from 1878 distinguished it from deduction and induction, calling it 'hypothesis' and later 'abduction' or 'retroduction'; Gilbert Harman's 1965 phrase 'inference to the best explanation' became a common synonym. Harry Pople's 1973 work on mechanizing medical diagnosis brought it into AI.
Usage note
Some philosophers separate abduction as generating hypotheses from inference to the best explanation as selecting among them; AI usage usually merges the two.
Related
- deductive reasoningdeductive reasoning, n.: Reasoning in which true starting points guarantee a true conclusion: if every square is a rectangle and this shape is a square, then it must be a rectangle.
- inductive reasoninginductive reasoning, n.: Reasoning from particular examples to a general rule or prediction — every swan I have seen is white, so swans are probably white — which is useful but can always turn out wrong.
- reasoningreasoning, n.: Working out conclusions step by step from facts, rules or evidence, as in solving a math problem or planning a trip. Whether today's AI truly reasons, or imitates reasoning, is debated.
- Bayesian networkBayesian network, n.: A diagram of causes and effects with probabilities attached — say, how rain and a sprinkler can each make grass wet — that lets a computer update its beliefs about hidden causes when it sees evidence.
- logic programminglogic programming, n.: A style of programming in which you write down facts and rules in logic and then ask questions; the computer works out the answers itself by logical deduction.
- 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.
- deductive reasoningdeductive reasoning, n.: Reasoning in which true starting points guarantee a true conclusion: if every square is a rectangle and this shape is a square, then it must be a rectangle.
- inductive reasoninginductive reasoning, n.: Reasoning from particular examples to a general rule or prediction — every swan I have seen is white, so swans are probably white — which is useful but can always turn out wrong.
Sources
- Peirce, C. S. (1878). Deduction, Induction, and Hypothesis. Popular Science Monthly 13, 470–482.
- Josephson, J. R. & Josephson, S. G. (eds.) (1994). Abductive Inference: Computation, Philosophy, Technology. Cambridge University Press.
Mentioned in
deductive reasoningdeductive reasoning, n.: Reasoning in which true starting points guarantee a true conclusion: if every square is a rectangle and this shape is a square, then it must be a rectangle., inductive reasoninginductive reasoning, n.: Reasoning from particular examples to a general rule or prediction — every swan I have seen is white, so swans are probably white — which is useful but can always turn out wrong.