# Knowledge Representation **Domain:** Artificial Intelligence / Logic / Information Systems **Doc Type:** Canonical Concept Node **Maturity:** Developed **Related:** [[wiki/GOFAI (Good Old-Fashioned AI)|GOFAI (Good Old-Fashioned AI)]], [[wiki/Ontology|Ontology]], [[wiki/Encoded Domain Knowledge|Encoded Domain Knowledge]], [[wiki/Semantic Network|Semantic Network]], [[wiki/Provenance|Provenance]] --- ## Definition **Knowledge representation is the design of computational structures that make entities, distinctions and relations available for reasoning, retrieval or action.** A representation determines what a system can ask, compare, infer and explain. Representations may include rules, ontologies, graphs, feature structures, vectors, examples or trained parameters. These forms are not interchangeable. An explicit taxonomy supports auditable “type of” traversal; an embedding supports graded similarity; a corpus index supports attestation; a pronunciation lexicon supports phonological operations. ## Language-Engine Context The [[wiki/Symbolic Language Engine|Symbolic Language Engine]] treats English as several overlapping representations rather than one undifferentiated word list. Pronunciations, rhyme classes, word senses, taxonomic edges, trigger relations and poetic constraints answer different questions about the same text. This approach illustrates the principle in [[wiki/Encoded Domain Knowledge|Encoded Domain Knowledge]]: intelligence depends partly on which human distinctions have been made computable. It also introduces an evidence obligation. When a system returns a suggestion, the representation that produced it should remain identifiable. ## Corpus-Governance Context [[wiki/Bidirectional Ontology|Bidirectional Ontology]] makes knowledge representation recursive. The corpus supplies evidence from which representational categories are learned, and those categories later shape how the corpus can be interpreted and extended. [[wiki/Semantic Governance|Semantic Governance]] keeps changes to that representation explicit because altering a category can change the apparent meaning of every document linked to it. ## Key Insight **Representation is not neutral storage. It defines the system's visible world and the operations that count as intelligence within it.** ## See Also [[wiki/Ontology|Ontology]], [[wiki/Bidirectional Ontology|Bidirectional Ontology]], [[wiki/Ontology Learning|Ontology Learning]], [[wiki/Semantic Governance|Semantic Governance]], [[wiki/Encoded Domain Knowledge|Encoded Domain Knowledge]], [[wiki/Semantic Network|Semantic Network]], [[wiki/Multi-Index Language Representation|Multi-Index Language Representation]], [[wiki/Word Embeddings|Word Embeddings]], [[wiki/Provenance-Sensitive Multiplicity|Provenance-Sensitive Multiplicity]] ## Relationships - **Edge Annual Question source relationship:** [[collections/Edge|Edge]] preserves the annual-question source corpus from which a proposition-specific quotation is connected to this entry. - **Edge source route:** [[collections/Edge|Edge]] connects this topic to exact Annual Question passages promoted into the Simple Reminders archive. ## Simple Reminders, Quotations, and Thoughts > "As long as something can be relayed that resolves uncertainty, that is the fundamental nature of information. While this sounds surprisingly obvious, it was an important point, given how many different languages people speak and how one utterance could be meaningful to one person, and unintelligible to another. Until Shannon's theory was formulated, it was not known how to compensate for these types of "psychological factors" appropriately. Shannon built on the work of fellow researchers Ralph Hartley and Harry Nyquist to reveal that coding and symbols were the key to resolving whether two sides of a communication had a common understanding of the uncertainty being resolved." > **— Andrew Lih**, *2012, Edge Annual Question, “What Is Your Favorite Deep, Elegant, or Beautiful Explanation?”* [[reminders/Information/Information Resolves Uncertainty Across Every Medium by Andrew Lih|Information Resolves Uncertainty Across Every Medium by Andrew Lih]] > “If the sum of all significant knowledge is finite, what proportion of it can humans, aided by intelligent machines, eventually attain?” > **— Ian McEwan**, *2018, Edge Annual Question, “WHAT IS THE LAST QUESTION?”* [[reminders/Information/How Much Knowledge Can Humans and Intelligent Machines Ultimately Attain by Ian McEwan|How Much Knowledge Can Humans and Intelligent Machines Ultimately Attain by Ian McEwan]] > “Because people understand by finding in their memories the closest possible match to what they are hearing and use that match as the basis of comprehension, any new idea will be treated as a variant of something the listener has already thought of or heard. Agreement with a new idea means a listener has already had a similar thought and well appreciates that the speaker has recognized his idea. Disagreement means the opposite. Really new ideas are incomprehensible. The good news is that for some people, failure to comprehend is the beginning of understanding. For most, of course, it is the beginning of dismissal.” > **— Roger Schank**, *2004, Edge Annual Question, “WHAT'S YOUR LAW?”* [[reminders/Information/Truly New Ideas Begin as Incomprehensible by Roger Schank|Truly New Ideas Begin as Incomprehensible by Roger Schank]] > “If we look back over recent centuries we will see the brain described as a hydrodynamic machine, clockwork, and as a steam engine. When I was a child in the 1950's I read that the human brain was a telephone switching network. Later it became a digital computer, and then a massively parallel digital computer. A few years ago someone put up their hand after a talk I had given at the University of Utah and asked a question I had been waiting for for a couple of years: "Isn't the human brain just like the world wide web?". The brain always seems to be one of the most advanced technologies that we humans currently have.” > **— Rodney A. Brooks**, *2008, Edge Annual Question, “WHAT HAVE YOU CHANGED YOUR MIND ABOUT? WHY?”* [[reminders/Information/We Keep Explaining the Brain With Our Most Advanced Technology by Rodney A. Brooks|We Keep Explaining the Brain With Our Most Advanced Technology by Rodney A. Brooks]] > “The new medium (which begins as a representation of the old medium) eclipses the old. The prize becomes more important than the achievement. The genes become more important than the organism.” > **— W. Daniel Hillis**, *2004, Edge Annual Question: “WHAT'S YOUR LAW?”* [[reminders/Cybernetics/The Representation Eclipses the Thing It Represents by W. Daniel Hillis|The Representation Eclipses the Thing It Represents by W. Daniel Hillis]] > “The universe is a quantum computer: life, sex, the brain, and human society all arise out of the ability of the universe to process information at the level of atoms, photons and elementary particles.” > **— Seth Lloyd**, *2004, Edge Annual Question: “WHAT'S YOUR LAW?”* [[reminders/Information/The Universe Is a Quantum Computer by Seth Lloyd|The Universe Is a Quantum Computer by Seth Lloyd]] > “How will people focus more on forming the right question, before rushing headlong towards the answer?” > **— Ziyad Marar**, *2018 Edge Annual Question, question* [[reminders/Information/How Can People Ask the Right Question First by Ziyad Marar|How Can People Ask the Right Question First by Ziyad Marar]] ## Sources / Provenance - Primary project account: [[projects/Ten Years Building a Symbolic Language Engine|Ten Years Building a Symbolic Language Engine]]. - Ronald Brachman and Hector Levesque, *Knowledge Representation and Reasoning* (2004). - Stuart Russell and Peter Norvig, *Artificial Intelligence: A Modern Approach*.