4 We need better mechanisms for reasoning! Clue: It was this anatomical oddity of US gymnast George Eyser.... Ken Jennings' answer: Missing a hand (wrong) Watson's answer: leg (wrong) Correct answer: Missing a leg
5 Turing’s Dream & Knowledge Challenge - Schubert Natural language is a pretty damn good knowledge representation language Has capabilities that formal KR doesn’t Resist the urge to “simplify so the computer can understand it” Don’t be so afraid of the Ambiguity bogeyman
8 Logical Reasoning: Classic example Birds can fly.Tweety is a bird.Therefore… Tweety can fly.
9 Logical Reasoning: Not-so-classic example Cheap apartments are rare.
10 Logical Reasoning: Not-so-classic example Cheap apartments are rare.Rare things are expensive.
11 Logical Reasoning: Not-so-classic example Cheap apartments are rare.Rare things are expensive.Therefore… Cheap apartments are expensive.So, exactly what was wrong with that??
12 Yeah, what's wrong with that? Logicians say:Not the same sense of "rare", "expensive", etc.I say:Maybe, but punts the problem of translating language/Commonsense to logicLogic is about possible inference; Common Sense is about plausible inference
13 Not so interested in absolute truth as we are in… Plausibility (not necessarily Probability)SimilarityAnalogyRelevanceComputing "intangible" qualities (affect, point of view, connotation, overall "sense")
14 Logical vs. Commonsense knowledge Precise VagueFormal Natural languageExperts General publicExplicit ImplicitConsistent Possibly contradictoryUp-front organization Back-end organization
15 Logical vs. Statistical Reasoning Big debate, much hot air We need to fill in the gap between them Word occurrences are weak evidence Symbolic expressions much stronger But how do you combine lots of them?
17 Open Mind Common Sense “The Wikipedia version of Cyc” since 2000 1 Million English statements, other languagesHow much Commonsense does an average person know?1 human lifetime = 3 billion secondsLess than a billion - Maybe 100 millionHow much domain knowledge does a single expert know?Less than a million - Maybe thousand
24 What AnalogySpace can do It can generalize from sparsely-collected knowledgeIt can identify the most important dimensions in a knowledge spaceIt can classify concepts along those dimensionsIt can create ad-hoc categories (and classify accordingly)It can confirm or question existing knowledge
28 Traditional Logical Inference Inferences goes from True assertion -> True assertion via Inference Rules Good news: Very precise and reliable Bad news: Proof search blows up exponentially Requires precise definitions and assertions GIGO
29 AnalogySpace Inference All possible assertions put in a (big, sparse) box You can rearrange the box along semantic axes Good news: Computationally efficient Tolerant of imprecision, contradiction, disagreement… Stronger than statistical inference Bad news: Can’t be guaranteed to be very precise
30 Not-so-Common SenseUse Common Sense tools & methodology, but knowledge only common to a small groupCollect knowledge from natural language sourcesCollect knowledge from gamesCollect knowledge from existing DBs, Ontologies, .."Blend" with general Commonsense knowledge-> AnalogySpace for specific domain
31 Blending - Havasi Inference combining two AnalogySpaces Specialized and generalized knowledge basesBlending factor
32 CrossBridge - Krishnamurthy AnalogySpace-based technique for Structure Mapping analogy Indexes small networks of concepts & assertions Can do Case-Based Reasoning Electricity flows through Wires -> Water flows through Pipes, or Light flows through Fiber-Optic Cables?
34 Applications in Interface Agents Predictive typing, Speech recognitionStorytelling with Media LibrariesDetection and mitigation of online bullyingOpinion AnalysisGoal-oriented interfaces for Consumer ElectronicsMobile to-do lists, location-aware context-sensitive mapsTranslation, language learning & multi-lingual communicationHelp and customer serviceRecommendation systems, scenario-based recommendationProgramming and code sharing in natural language… and more
35 Example: Earth Sciences Knowledge Collaboration with SchlumbergerCollect Earth Sciences Knowledge for intelligent search & browsing~ 2000 assertions = 300 manual from gameGame = 2 one-hour sessions x 10 people350 concepts, read glossary document
36 Geology sentences Petroleum is a mixture of hydrocarbons. [IsA] Air gun array is used for seismic surveying offshore. [UsedFor]A seismic survey is a measure of seismic-wave travel. [Measures]A wildcat is an exploration well drilled in an unproven area. [IsA]You would drill an exploration well because you want to determine whether hydrocarbons are present. [MotivatedByGoal]
45 ConclusionThere’s been a controversy between logical and statistical reasoning We need to fill in the gap Symbolic representations as source “Do the math” to combine large numbers of them New thinking about Commonsense reasoning