1 / 28

The Logic of Intelligence

The Logic of Intelligence Pei Wang Department of Computer and Information Sciences Temple University Logic, Cognition, and AI In general, “logic” is the study of valid reasoning , and reasoning plays an important role in cognition.

betty_james
Télécharger la présentation

The Logic of Intelligence

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. The Logic of Intelligence Pei Wang Department of Computer and Information Sciences Temple University

  2. Logic, Cognition, and AI In general, “logic” is the study of valid reasoning, and reasoning plays an important role in cognition. The study of artificial intelligence aims at computer systems with the cognitive capacity of the human mind. “Intelligence” can be understood as “rationality” and “validity” --- “to do the right thing”.

  3. Reasoning System A computerized reasoning system typically consists of the following major components: • a formal language, • a semantic theory, • a set of inference rules, • a memory structure, • a control mechanism. The first three are usually called a “logic”, and the last two use the logic in a computer system.

  4. Traditional Theories • language and inference rules: first-order predicate calculus, • semantics: model theory, • memory: relational or object-oriented database, • inference control: theory of computation (algorithm, computability, and computational complexity).

  5. Problems of Traditional Theories • uncertainty: fuzzy concepts, changing meanings and truth values, plausible results, conflicting evidence, nondeterministic inference process, … • semantic justification of non-deductive inference: induction, abduction, analogy, … • counter-intuitive results: sorites paradox, implication paradox, confirmation paradox, Wason’s selection task, … • computability and complexity: termination problem, combinatorial explosion, …

  6. non-monotonic logic paraconsistent logic relevance logic probabilistic logic fuzzy logic inductive logic temporal logic modal logic situation calculus possible world theory Proposed Solutions • mental logic • mental model • case-based reasoning • Bayesian network • neural network • genetic algorithm • heuristic algorithm • learning algorithm • anytime algorithm … …

  7. Common Root of the Problems The traditional theories were developed in the study of the foundation of mathematics, while the problems appear outside math. The logic of mathematics may be different from the logic of cognition. In mathematical reasoning, the knowledge and resources are assumed to be sufficient (with respect to the tasks).

  8. Different Types of Systems “pure-axiomatic system”: the system’s knowledge and resources are assumed to be sufficient. “semi-axiomatic system”: certain aspects (but not all) of the knowledge and resources are assumed to be sufficient. “non-axiomatic system”: the knowledge and resources of the system are assumed to be generally insufficient.

  9. Non-Axiomatic Reasoning System NARS is a reasoning system. It uses a formal logic (language, semantics, inference rules) and is implemented in a computer system. NARS is fully based on the assumption of insufficient knowledge and resources, in the sense of being a finite, real time, open, and adaptive system. NARS is different from traditional theories in all major components.

  10. bird animal Inheritance is reflexive and transitive. Inheritance Based Representation S  P : there is an inheritance relation from term S to term P. S is aspecialization of P; P is a generalization of S.

  11. TI TE T Extension and Intension For a given term T, its extension TE = {x | xT} its intension TI = {x | T x} Theorem: (S P)  (SE PE) (PI  SI) Therefore, “Inheritance” means “inheritance of extension/intension”.

  12. Æ Å S P Å Æ Evidence Positive evidence of S P : {x | x  (SE PE)  (PI SI)} Negative evidence of S P : {x | x  (SE–PE)  (PI–SI)} Amount of evidence: positive: w+ = | SE PE| + | PI SI| negative: w– = | SE–PE| + | PI–SI| total: w = w+ + w– = | SE| + | PI|

  13. S P [f, c] Truth Value In NARS, the truth value of a statement is a pair of numbers, and measures the evidential support to the statement. S P [f, c] f: frequency, w+/w c: confidence, w / (w +1)

  14. Experience-Grounded Semantics The truth value of a statement is defined according to certain “idealized experience”, consisting of a set of binary inheritance statements. The meaning of a term is defined by its extension and intension, according to certain “idealized experience”.

  15. Syllogistic Inference Rules A typical syllogistic inference rule takes a pair of premises, and produces a conclusion. The truth value of the conclusion is calculated by a function according to the semantics of NARS. Different combinations of premises trigger different rules (with different truth-value functions).

  16. To Design Truth-value Functions 1. Treat all involved variables as Boolean (binary) variables; 2. For each value combination in premises, decide the values in conclusion; 3. Build Boolean functions among the variables; 4. Extend the functions to real-number: not(x) = 1 – x; and(x, y) = x * y; or(x, y) = 1 – (1 – x) * (1 – y).

  17. M S P Deduction bird  animal [1.00, 0.90] robin bird [1.00, 0.90] ¾¾¾¾¾¾¾¾¾¾ robin  animal [1.00, 0.81] M  P [f1, c1] S  M [f2, c2] ¾¾¾¾¾¾¾ S  P [f, c] f = f1 * f2 c = c1 * c2 * f1 * f2

  18. M S P Induction swan bird [1.00, 0.90] swan  swimmer [1.00, 0.90] ¾¾¾¾¾¾¾¾¾¾¾ bird  swimmer [1.00, 0.45] M  P [f1, c1] M  S [f2, c2] ¾¾¾¾¾¾¾ S  P [f, c] f = f1 c = f2 * c1 * c2 / (f2 * c1 * c2 + 1)

  19. M S P Abduction seabird swimmer [1.00, 0.90] gull  swimmer [1.00, 0.90] ¾¾¾¾¾¾¾¾¾¾¾¾¾ gull  seabird [1.00, 0.45] P M [f1, c1] S  M [f2, c2] ¾¾¾¾¾¾¾ S  P [f, c] f = f2 c = f1 * c1 * c2 / (f1 * c1 * c2 + 1)

  20. P S Revision bird  swimmer [1.00, 0.62] bird  swimmer [0.00, 0.45] ¾¾¾¾¾¾¾¾¾¾¾ bird  swimmer [0.67, 0.71] f1* c1* (1 - c2) + f2* c2* (1 - c1) ¾¾¾¾¾¾¾¾¾¾ c1* (1 - c2) + c2* (1 - c1) c1* (1 - c2) + c2* (1 - c1) ¾¾¾¾¾¾¾¾¾¾¾¾¾ c1* (1 - c2) + c2* (1 - c1) + (1 - c2) * (1 - c1) S  P [f1, c1] S  P [f2, c2] ¾¾¾¾¾¾¾ S  P [f, c] f = c =

  21. Other Inference Rules M  P [f1, c1] S  M [f2, c2] ¾¾¾¾¾¾¾ S  P [f, c] analogy union P  M [f1, c1] S  M [f2, c2] ¾¾¾¾¾¾¾¾¾ (S  P)  M [f, c] implication B  C [f1, c1] A  B [f2, c2] ¾¾¾¾¾¾¾ A  C [f, c]

  22. Other Relations and Inheritance An arbitrary statement R(a, b, c) can be rewritten as inheritance relations with compound terms: • (*, a, b, c) R “The relation among a, b, c is a kind of R.” • a(/, R, _, b, c) “ais such an xthat satisfies R(x, b, c).” • b(/, R,a, _, c) “bis such an xthat satisfies R(a, x, c).” • c(/, R,a, b, _) “cis such an xthat satisfies R(a, b, x).”

  23. gull swimmer robin [1.00, 0.90] [1.00, 0.90] [1.00, 0.90] feathered_creature [0.00, 0.90] [1.00, 0.90] [1.00, 0.90] [1.00, 0.90] [1.00, 0.90] crow bird swan Memory as a Belief Network The knowledge of the system is a network of beliefs among terms. A term with all of its beliefs is a concept. Cbird

  24. Inference Tasks The current version of NARS accepts two types of inference tasks: • questions to be answered, • new knowledge to be absorbed. For both types of tasks, the system let them interact with available beliefs. For questions, the interactions produce answers or derived questions. For new knowledge, the interactions add or modify the relevant beliefs. This process usually generates new tasks recursively. A task is also stored in the corresponding concepts.

  25. Inference Process When running, NARS repeats the following cycle: • Choose a concept within the memory; • Choose a task within the concept; • Choose a belief within the concept; • Use applicable rules to produce new tasks; • Return the used items to memory; • Add the new (input and derived) tasks into the memory. Provide an answer if available.

  26. Control Strategy NARS maintains priority distributions among concepts, tasks, and beliefs, uses them to make choice, and adjusts them after each step. Factors influence priority: • quality of the item, • usefulness of the item in history, • relevance of the item to the current context.

  27. Unified Solutions • The truth value uniformly represents various kinds of uncertainty. • The truth value depends on both positive and negative evidence. • The non-deductive inference rules is justified according to the semantics. • The meaning of a term is determined by its experienced relations with other terms. • With syllogistic rules, the premises and conclusions must be semantically related. • The inference processes in NARS does not follow predetermined algorithms.

  28. Conclusions It is possible to build a reasoning system that adapts to its environment, and works with insufficient knowledge and resources. Such a system provides a unified solution to many problems in cognitive sciences. There is a logic of intelligence, though it is fundamentally different from the logic of mathematics.

More Related