830 likes | 846 Views
9.012 Brain and Cognitive Sciences II. Part VIII: Intro to Language & Psycholinguistics - Dr. Ted Gibson. Presented by Liu Lab. Fighting for Freedom with Cultured Neurons. Distributed Representations, Simple Recurrent Networks, And Grammatical Structure Jeffrey L. Elman (1991)
E N D
9.012Brain andCognitive Sciences II Part VIII: Intro to Language & Psycholinguistics - Dr. Ted Gibson
Presented by Liu Lab Fighting for Freedom with Cultured Neurons
Distributed Representations, Simple Recurrent Networks, And Grammatical Structure Jeffrey L. Elman (1991) Machine Learning Nathan Wilson
Distributed Representations/ Neural Networks • are meant to capture the essence of neural computation: many small, independent units calculating very simple functions in parallel.
Distributed Representations/ Neural Networks: EXPLICIT RULES?
Distributed Representations/ Neural Networks: EXPLICIT RULES?
Distributed Representations/ Neural Networks: EXPLICIT RULES? EMERGENCE!
Distributed Representations/ Neural Networks • are meant to capture the essence of neural computation: many small, independent units calculating very simple functions in parallel.
FeedForward Neural Network (from Sebastian’s Teaching)
FeedForward Neural Network (from Sebastian’s Teaching)
Why Apply Network / Connectionist Modeling to Language Processing? • Connectionist Modeling is Good at What it Does • Language is a HARD problem
What We Are Going to Do • Build a network
What We Are Going to Do • Build a network • Let it learn how to “read”
What We Are Going to Do • Build a network • Let it learn how to “read” • Then test it!
What We Are Going to Do • Build a network • Let it learn how to “read” • Then test it! • Give it some words in a reasonably grammatical sentence • Let it try to predict the next word, • Based on what it knows about grammar
What We Are Going to Do • Build a network • Let it learn how to “read” • Then test it! • Give it some words in a reasonably grammatical sentence • Let it try to predict the next word, • Based on what it knows about grammar • BUT: We’re not going to tell it any of the rules
What We Are Going to Do • Build a network
FeedForward Neural Network (from Sebastian’s Teaching)
Methods > Network Implementation > Structure 0000000000001 OUTPUT 100100100100100100100100 HIDDEN 1000000000000 INPUT
What We Are Going to Do • Build a network • Let it learn how to “read”
Methods > Network Implementation > Training Words We’re going to Teach it: • - Nouns: • boy | girl | cat | dog | • boys | girls | cats | dogs • - Proper Nouns: • John | Mary • “Who” • - Verbs: • chase | feed | see | hear | walk | live | • chases | feeds | sees | hears | walks | lives • “End Sentence”
Methods > Network Implementation > Training 1. Encode Each Word with Unique Activation Pattern
Methods > Network Implementation > Training 1. Encode Each Word with Unique Activation Pattern • - boy => 000000000000000000000001 • girl => 000000000000000000000010 • feed => 000000000000000000000100 • -sees => 000000000000000000001000 • . . . • who => 010000000000000000000000 • End sentence => • 100000000000000000000000
Methods > Network Implementation > Training 1. Encode Each Word with Unique Activation Pattern • - boy => 000000000000000000000001 • girl => 000000000000000000000010 • feed => 000000000000000000000100 • -sees => 000000000000000000001000 • . . . • who => 010000000000000000000000 • End sentence => • 100000000000000000000000 2. Feed these words sequentially to the network (only feed words in sequences that make good grammatical sense!)
Methods > Network Implementation > Structure 1000000000000 INPUT
Methods > Network Implementation > Structure HIDDEN 1000000000000 INPUT
Methods > Network Implementation > Structure 100100100100100100100100 HIDDEN 1000000000000 INPUT
Methods > Network Implementation > Structure OUTPUT 100100100100100100100100 HIDDEN 1000000000000 INPUT
Methods > Network Implementation > Structure 0000000000001 OUTPUT 100100100100100100100100 HIDDEN 1000000000000 INPUT
Methods > Network Implementation > Training 1. Encode Each Word with Unique Activation Pattern • - boy => 000000000000000000000001 • girl => 000000000000000000000010 • feed => 000000000000000000000100 • -sees => 000000000000000000001000 • . . . • who => 010000000000000000000000 • End sentence => • 100000000000000000000000 2. Feed these words sequentially to the network (only feed words in sequences that make good grammatical sense!)
Methods > Network Implementation > Structure 0000000000001 OUTPUT 100100100100100100100100 HIDDEN 1000000000000 INPUT
What We Are Going to Do • Build a network • Let it learn how to “read”
Methods > Network Implementation > Structure 0000000000001 OUTPUT 100100100100100100100100 HIDDEN 1000000000000 INPUT
Methods > Network Implementation > Structure 0000000000001 OUTPUT 100100100100100100100100 HIDDEN If learning word relations, need some sort of memory from word to word! 1000000000000 INPUT
FeedForward Neural Network (from Sebastian’s Teaching)
Methods > Network Implementation > Structure 0000000000001 OUTPUT 100100100100100100100100 HIDDEN 1000000000000 INPUT
Methods > Network Implementation > Structure 0000000000001 OUTPUT 100100100100100100100100 HIDDEN 1000000000000 100100100100100100100100 INPUT CONTEXT
Methods > Network Implementation > Structure 0000000000001 OUTPUT 100100100100100100100100 HIDDEN 1000000000000 100100100100100100100100 INPUT CONTEXT
Methods > Network Implementation > Structure 0000000000001 OUTPUT 100100100100100100100100 HIDDEN 1000000000000 100100100100100100100100 INPUT CONTEXT
Methods > Network Implementation > Structure 0000000000001 OUTPUT 100100100100100100100100 HIDDEN 1000000000000 100100100100100100100100 INPUT CONTEXT
Methods > Network Implementation > Structure BACKPROP! 0000000000001 OUTPUT 100100100100100100100100 HIDDEN 1000000000000 100100100100100100100100 INPUT CONTEXT
What We Are Going to Do • Build a network • Let it learn how to “read” • Then test it! • Give it some words in a reasonably grammatical sentence • Let it try to predict the next word, • Based on what it knows about grammar • BUT: We’re not going to tell it any of the rules
Results > Emergent Properties of Network > Subject-Verb Agreement • After Hearing: • “boy….” • Network SHOULD predict next word is: • “chases” • NOT: • “chase” • Subject and verb should agree!
Results > Emergent Properties of Network > Noun-Verb Agreement • After Hearing: • “boy….” • Network SHOULD predict next word is: • “chases” • NOT: • “chase” • Subject and verb should agree!
Results > Emergent Properties of Network > Noun-Verb Agreement boy….. End of Sentence “Who” Plural Verb, DO Impossible Plural Verb, DO Required Plural Verb, DO Optional What Word Network Predicts is Next Single Verb, DO Impossible Single Verb, DO Required Single Verb, DO Optional Plural Noun Single Noun 0.0 0.2 0.4 0.6 0.8 1.0 Activation