Deduction
Reasoning from general rules to specific facts; the inverse of induction.
As defined in Pedro Domingos on Machine Learning and the Master Algorithm
Other definitions
Reasoning from general principles to specific conclusions that must follow from them.
#15 Leonardo da Vinci: The Biography
Where it comes up · 5
- AI Today and Tomorrow2 claims
- #15 Leonardo da Vinci: The Biography1 claim
- Pedro Domingos on Machine Learning and the Master Algorithm1 claim
- 335 | Andrew Jaffe on Models, Probability, and the Universe1 claim
- 343 | Tom Griffiths on The Laws of Thought1 claim
Most important claims · 6 of 6Search all claims for “Deduction” →
- Deduction goes from true premises to true conclusions with certainty but is in a sense 'boring' because the conclusions merely follow from definitions, whereas induction requires a leap of probability, so a model can only ever be known to correspond to the world probabilistically.335 | Andrew Jaffe on Models, Probability, and the Universe
- Causality, at the end of the day, just means being able to predict the effects of your actions; a notion of causality that does not involve predicting the effects of your actions would be indistinguishable from correlation, and machine learning can be used to predict the effects of your actions.Pedro Domingos on Machine Learning and the Master Algorithm
- Deduction is reasoning from certain premises to certain conclusions, whereas induction is reasoning from things we know that may not fully determine the conclusion, yet still making a reasonable inference on that basis—the kind of reasoning scientists use to discover principles of the world.343 | Tom Griffiths on The Laws of Thought
- Leonardo preferred to induce from experiments rather than deduce from theoretical principles, consulting experience first and only then using reasoning to show why the experience must operate as it does.#15 Leonardo da Vinci: The Biography
- Connectionist methods (based on induction) have dominated AI research publications overwhelmingly compared to symbolic methods (based on deduction), with symbolic methods only briefly dominating between the two schools of thought.AI Today and Tomorrow
- Symbolic methods require writing down formal logical systems and deducing from them, exemplified by calculators which cannot make arithmetic mistakes because humans explicitly programmed the rules of arithmetic into them.AI Today and Tomorrow