DOC PREVIEW
Berkeley COMPSCI 182 - Lecture Notes

This preview shows page 1-2-3-4-27-28-29-30-55-56-57-58 out of 58 pages.

Save
View full document
View full document
Premium Document
Do you want full access? Go Premium and unlock all 58 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 58 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 58 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 58 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 58 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 58 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 58 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 58 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 58 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 58 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 58 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 58 pages.
Access to all documents
Download any document
Ad free experience
Premium Document
Do you want full access? Go Premium and unlock all 58 pages.
Access to all documents
Download any document
Ad free experience

Unformatted text preview:

Five levels of Neural Theory of LanguageSlide 3Reflexive ReasoningSlide 5How fast is reflexive reasoning?How can a system of slow and simple neuron-like elementsCharacterization of reflexive reasoning?ShrutiReflexive Reasoning representational and processing issuesDynamic representation of relational instancesSlide 12LearningRelation focal-clusterEntity, category and relation focal-clustersSlide 16Focal-cluster of a relational schemaSlide 18Focal-cluster of an entitySlide 20Slide 21Slide 22Encoding “slip => fall” in ShrutiSlide 24A Metaphor for ReasoningFocal-clusters with intra-cluster linksSlide 27Linking focal-clusters of types and entitiesFocal-clusters and context-sensitive priors (T-facts)Focal-clusters and episodic memories (E-facts)Explaining away in ShrutiOther features of ShrutiUnification in Shruti : merging of phasesEntity instantiation in ShrutiEncoding “fall => hurt” in ShrutiSlide 36Slide 37Slide 38Slide 39Support for ShrutiNeurophysiological evidence for synchronyPredictions: constraints on reflexive inferencePredictions: Constraints on reflexive reasoningMassively Parallel InferenceProbabilistic interpretation of link weightsSlide 46Encoding X-schemaSlide 48Slide 49Slide 50Slide 51Representing belief and utility in ShrutiEncoding “Fall => Hurt”Focal-clusters augmented to encode belief and utilityBehavior of augmented ShrutiShruti suggests how different sorts of knowledge may be encoded within neurally plausible networksCurrent status of learning in Shruti…current status of learning in ShrutiSlide 59Lokendra Shastri Lokendra Shastri ICSI, Berkeley ICSI, BerkeleyFive levels of Neural Theory of LanguageCognition and LanguageComputationStructured ConnectionismComputational NeurobiologyBiologySHRUTISHRUTIabstractionLokendra Shastri Lokendra Shastri ICSI, Berkeley ICSI, Berkeley“John fell in the hallway. Tom had cleaned it. He got hurt.” Tom had cleaned the hallwaythe hallway. The hallway floor was wetThe hallway floor was wet.. John slipped and fell on the wet floorJohn slipped and fell on the wet floor. JohnJohn got hurt as a result of the fallas a result of the fall.such inferences establish referential and causal coherence.Lokendra Shastri Lokendra Shastri ICSI, Berkeley ICSI, BerkeleyReflexive Reasoning Ubiquitous Automatic, effortless Extremely fast --- almost a reflex response of our cognitive apparatusLokendra Shastri Lokendra Shastri ICSI, Berkeley ICSI, BerkeleyReflexive ReasoningNot all reasoning is reflexiveContrast with reflective reasoningdeliberateinvolves explicit consideration of alternativesrequire props (paper and pencil)e.g., solving logic puzzles … differential equationsHow fast is reflexive reasoning?•We understand language at the rate of 150-400 words per minute Reflexive inferences required for establishing inferential and causal coherence are drawn within a few hundred millisecond Lokendra Shastri Lokendra Shastri ICSI, Berkeley ICSI, BerkeleyHow can a system of slow and simple neuron-like elements•encode a large body of semantic and episodic knowledge and yet•perform a wide range of inferences within a few hundred milliseconds? Lokendra Shastri Lokendra Shastri ICSI, Berkeley ICSI, BerkeleyCharacterization of reflexive reasoning?•What can and cannot be inferred via reflexive processes? Lokendra Shastri Lokendra Shastri ICSI, Berkeley ICSI, BerkeleyShrutihttp://www.icsi.berkeley.edu/~shastri/shruti•Lokendra Shastri•V. Ajjanagadde (Penn, ex-graduate student)•Carter Wendelken (UCB, ex-graduate student) •D. Mani (Penn, ex-graduate student)•D.J. Grannes (UCB, ex-graduate student) •Jerry Hobbs, USC/ISI (abductive reasoning)•Marvin Cohen, CTI (metacognition; belief and utility)•Bryan Thompson, CTI (metacognition; belief and utility)Lokendra Shastri Lokendra Shastri ICSI, Berkeley ICSI, BerkeleyReflexive Reasoningrepresentational and processing issues•Activation-based (dynamic) representation of events and situations (relational instances)Lokendra Shastri Lokendra Shastri ICSI, Berkeley ICSI, BerkeleyDynamic representation of relational instances“John gave Mary a book”giver: Johnrecipient: Marygiven-object: a-bookgivera-bookMaryrecipientJohngiven-object*Lokendra Shastri Lokendra Shastri ICSI, Berkeley ICSI, BerkeleyReflexive Reasoning•Expressing dynamic bindings •Systematically propagating dynamic bindings•Computing coherent explanations and predictions–evidence combination–dynamic instantiation and unification of entities Requires compatible neural mechanisms for:All of the above must happen rapidlyLokendra Shastri Lokendra Shastri ICSI, Berkeley ICSI, BerkeleyLearning•one-shot learning of events and situations (episodic memory)•gradual/incremental learning of concepts, relations, schemas, and causal structuresLokendra Shastri Lokendra Shastri ICSI, Berkeley ICSI, BerkeleyRelation focal-cluster + - ? fall-pat fall-locFALLFALLLokendra Shastri Lokendra Shastri ICSI, Berkeley ICSI, BerkeleyEntity, category and relation focal-clusters + - ? fall-pat fall-locFALLFALLLokendra Shastri Lokendra Shastri ICSI, Berkeley ICSI, BerkeleyEntity, category and relation focal-clusters + - ? fall-pat fall-locFALLFALLFunctional nodes in a focal-cluster [collector (+/-), enabler (?), and role nodes] may be situated in different brain regionLokendra Shastri Lokendra Shastri ICSI, Berkeley ICSI, BerkeleyFocal-cluster of a relational schemaFALLFALL + - ? fall-pat fall-locfocal-clusters of motor schemasassociated with fallfocal-clusters of lexical know-ledge associated with fallfocal-clusters of perceptual schemas and sensory representations associated with fallfocal-clusters of otherrelational schemascausally related to fallepisodicmemories offall eventsLokendra Shastri Lokendra Shastri ICSI, Berkeley ICSI, BerkeleyFocal-clustersNodes in the fall


View Full Document

Berkeley COMPSCI 182 - Lecture Notes

Documents in this Course
Load more
Download Lecture Notes
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Lecture Notes and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Lecture Notes 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?