Distance A New Class of Methods Ronald Tolley AI Assessment - - PowerPoint PPT Presentation

distance
SMART_READER_LITE
LIVE PREVIEW

Distance A New Class of Methods Ronald Tolley AI Assessment - - PowerPoint PPT Presentation

Distance A New Class of Methods Ronald Tolley AI Assessment [Various aspects of] artificial [intelligence] have skewed off to find specialized niches Text recognition and document scanning are beginning to provide a


slide-1
SLIDE 1

Distance

A New Class of Methods Ronald Tolley

slide-2
SLIDE 2

AI Assessment

“[Various aspects of] artificial [intelligence] … have skewed off … to find specialized niches … “Text recognition and document scanning are … beginning to provide a significant new input medium for computer systems. “… the original vision of creating a true, humanlike intelligence that started so much of this research remains as unrealized as ever.”

Hogan, Mind Matters, p. 199

slide-3
SLIDE 3

Distance Assessment

  • Overall AI assessment
  • FH domain

– Match / Merge Consolidation

  • Non-FH domains
  • Contrast FH and classical AI applications
  • Contrast machine and human methods
  • Corridor methods
slide-4
SLIDE 4

Distance Example 1

KELLOGG Moses b b Massachusetts m Lydia KELLOGG m about 1748 m d d KELLOGG Moses b b m Mary SHELDON m 30 Apr 1740 m d d Massachusetts

slide-5
SLIDE 5

Distance Example 2

FISHER William b b Devon, England m Sarah Warren m 1 Apr 1849 m d d Nephi, Utah FISHER William b b Devon, England m Sarah Gadd m 11 Jan 1869 m d d probably Idaho

slide-6
SLIDE 6

Family History versus Classical AI

  • Recorded with intent
  • No resampling possible
  • Missing / occulted data
  • Definitive structure

– complexity in resolving issues

  • Back story

… back story

… back story

slide-7
SLIDE 7

Three Images

slide-8
SLIDE 8

Three Images

slide-9
SLIDE 9

Three Images

slide-10
SLIDE 10

Three Top Strips

slide-11
SLIDE 11

Three Middle Strips

slide-12
SLIDE 12

Three Bottom Strips

slide-13
SLIDE 13

Short Image Sequence

slide-14
SLIDE 14

Long Sequence

slide-15
SLIDE 15

Missing Elements: Occultation

  • Human visual field

– unifying fragments

  • McCloud

– closure

  • Restak

– fill-in

  • Hogan

– emergent properties

slide-16
SLIDE 16

Missing Elements: Closure

  • Human visual field

– unifying fragments

  • McCloud

– closure

  • Restak

– fill-in

  • Hogan

– emergent properties

slide-17
SLIDE 17

Compare: machine, human

Classical AI

  • High Leverage
  • Strong Methods
  • Very Precise Criteria
  • Exacting Evaluation
  • Reductivistic

– simplicity – Occam

  • Uncertainty

– handled as defect Classical Human

  • Low Leverage
  • “Weak” Methods
  • Imprecise Criteria
  • Arbitrary Evaluation
  • Non-reductivistic

– complexity – Rube Goldberg

  • Uncertainty

– Fill in missing data – Closure

slide-18
SLIDE 18

Contrast: machine, human

Classical AI

  • Syntactic methods in pattern recognition
  • Statistic methods in pattern recognition
  • Self-Organizing systems
  • Image processing
  • Feature extraction
  • Symbol manipulation / LISP / List Processing
  • Pattern matching
  • Games / Decision Trees / Searches

– pruning – combinatorix

  • Chess / Music / Mathematics
  • Data mining
  • Dualism / Pumps
  • Natural languages / Translation

– Eliza

  • Semantic nets / associative nets
  • Neural nets
  • Self-modifying code / Genetic programming
  • Models / Metaphors / Analogies / Parallels
  • Distances / Models / Methods / Contexts
  • Probabilities

– Bayes theorem

Classical Human

  • Limited by time, money,

energy, patience

  • Persistence
  • Comparison
  • Parallels, metaphors,

models, analogies

  • Negotiation

– concession ladder

  • Tool collectors
  • Common sense
  • Expectation

– foresight

  • Belief
slide-19
SLIDE 19

New Taxonomy within AI

  • Handling of Missing / Occulted data
  • Concentration / Distribution of Features
  • Graphical and symbolic processing

– Blurring the borderline

  • Parallelism / Metaphors
  • Limited Reductivism
  • Holographic

leads to

  • Corridor Methods
slide-20
SLIDE 20

Conclusions

  • Artificial Intelligence

– niche applications – no generalized solutions

  • Unique human “fill-in” ability

– deal with hidden / occulted data – reach closure

  • Corridor Methods