#!GFKNT 2.0 # This is an automatically generated file. Do not edit. #/ #? #$0 #C22-04-2004 22:47:17 #^011000000000000000000000 %+ NN=Praveen's Notes on Everything ID=1 II=0 DC=22-04-2004 22:47:17 TI=0 TS=4 CX=422 CY=0 BG=clWindow CH=0 FC=clWindowText FN=Century FS=12 LN=1033 ST= FL=101110000000110000000000 SN=8 TW=244 EN=New node TB=clWindow TH=1 TC=clWindowText TN=MS Sans Serif TZ=8 TY= %- LV=0 ND=Root DI=15 NF=000000100000000000000000 BC=clWindow IX=0 SS=67 %: {\rtf1\ansi\ansicpg1252\deff0{\fonttbl{\f0\fnil\fcharset0 Century;}} \viewkind4\uc1\pard\lang1033\f0\fs24 The contents of these notes are personal. Do NOT read any further. \par } %- LV=1 ND=ToDo DI=5 NF=000000000000000000000000 BC=clWindow IX=0 SS=15 %: {\rtf1\ansi\ansicpg1252\deff0{\fonttbl{\f0\fnil\fcharset0 Century;}} {\colortbl ;\red255\green0\blue0;} \viewkind4\uc1\pard\lang1033\f0\fs24 25 April 2004 \par \par Go to school and print out PS papers and go sit somewhere and work on the draft. Need the ability to print, too. Maybe library might be fine. Or even PLAB might not be bad. \par \par \pard{\pntext\f0 1)\tab}{\*\pn\pnlvlbody\pnf0\pnindent0\pnstart1\pnqr\pndec{\pntxta)}} \fi-440\li440 Call Sprint\par {\pntext\f0 2)\tab}Deposit money in bank -- remember to get cash before leaving. \par {\pntext\f0 3)\tab}Talk to pratim and then David Rabjohns. \par \cf1{\pntext\f0 4)\tab}Finish paper draft. \par {\pntext\f0 5)\tab}Get a few more examples done\par {\pntext\f0 6)\tab}Try to write an analysis of whats going on in the examples.\par \pard\cf0\par \par } %- LV=1 ND=Research DI=2 NF=000000100000000000000000 BC=clWindow IX=0 SS=9 %: {\rtf1\ansi\ansicpg1252\deff0{\fonttbl{\f0\fnil\fcharset0 Century;}} \viewkind4\uc1\pard\lang1033\b\f0\fs28 Research\par \fs24\par \b0 Everything that is directly or indirectly related to my thesis research. There might be stuff here that just might not make the cut for the thesis, but will be considered. \par } %- LV=2 ND=Knowledge Representation DI=3 NF=000000100000000000000000 BC=clWindow IX=0 SS=34 %: {\rtf1\ansi\ansicpg1252\deff0{\fonttbl{\f0\fnil\fcharset0 Century;}} \viewkind4\uc1\pard\lang1033\b\f0\fs24 Knowledge Representation. \par \b0\par This is where all my KR ideas/thought/struggles will go.\par } %- LV=3 ND=Healthcare DI=4 NF=000000000000000000000000 BC=clWindow IX=0 SS=1301 %: {\rtf1\ansi\ansicpg1252\deff0{\fonttbl{\f0\fnil\fcharset0 Century;}{\f1\fswiss\fcharset0 Arial;}} {\colortbl ;\red255\green0\blue0;\red0\green0\blue0;} \viewkind4\uc1\pard\lang1033\ul\b\f0\fs28 What is the annual cost of healthcare in the US?\ulnone\fs24\par \b0\par A solution: \par \cf1 Lets say everyone was insured. \par Total cost of healthcare has to be less than insurance premiums, for\par the insurance companies to stay in business. \par Cost of healthcare = average insurance premium * population\par = 3,000 * 300 million ~ 1 trillion\par [1.6 trillion last year. How does this work?]\cf2\f1\par \cf0\f0\par Cyc already has InsuranceOrganization --> InsuranceProvider ---> InsuranceCompany\par \par New constants needed: \par \par Insurance -- this is the concept of insurance. Maybe we mean InsuranceService, because that is what the insurance company provides you. In return of that service, you pay the InsurancePremium. \par \par HealthInsuranceService\par \par HealthInsurancePremium: base this on life expectancy\par (lifeExpectancyForGroupInRegion France MaleHuman (YearsDuration 74.5))\par (healthInsurancePremiumForGroupInRegion UnitedStatesOfAmerica Person (DollarsPerYear 3000))\par \par \b Paying for a product.\b0 \par We also need a general theory of how things are bought, sold and paid for. Lets start with paying for. We can -- \par \pard{\pntext\f0 1)\tab}{\*\pn\pnlvlbody\pnf0\pnindent0\pnstart1\pnqr\pndec{\pntxta)}} \fi-440\li440 buy one unit, pay for that\par {\pntext\f0 2)\tab}buy a subscription, and pay for all of it at once, and get products over a period of time. \par {\pntext\f0 3)\tab}buy something, or purchase a plan, in which one has to pay over a period. Examples -- rent, tuition, etc. \par \pard We already have\par (cost OBJ COST)\par \par \par \par (defparameter *health-query* \par '(annualCost HealthCare UnitedStatesOfAmerica (YearFn 2003) ?money))\par \par (defparameter *health-suggestions*\par '(;; This one gets reused.\par (defSuggestion PerBuyerStrategy\par (annualSales ?obj ?place ?time ?money)\par :subgoals ((annualExpensePerBuyer ?obj ?place ?time ?money-per-buyer)\par (numberOfBuyers ?obj ?place ?time ?number))\par :result-step (evaluate ?money (TimesFn ?money-per-buyer ?number)))\par \par ;; annual expense per buyer on a specific object. If we have discrete\par ;; units of that object that are bought separately, then it is the product\par ;; of number of units by cost per unit. on the other hand if it is something\par ;; that is paid for monthly or in installments, we want to add that all up\par ;; and find the annualExpense. For example, rent. In case of insurance, \par ;; its more sensible to assume we know the annual insurance premium \par ;; so we dont have to do much here. This example demonstrates the \par ;; issues in generality of suggestions. Suggestions have unique names, \par ;; so this one has to be named differently from UnitaryStrategyForCost\par \par (defSuggestion ExpenseOnHealthStrategy\par (annualExpensePerBuyer HealthCare ?place ?time ?money-per-buyer)\par :subgoals ((healthInsurancePremiumForGroupInRegion ?place Person (DollarsPerYear ?money-per-buyer)))\par \par ;;; (defSuggestion FrequencyStrategy\par ;;; (annualUnitsBoughtPerBuyer ?obj ?place ?time ?units)\par ;;; :subgoals ((\par (defSuggestion FractionOfTotalStrategy\par (numberOfBuyers ?obj ?place ?time ?number)\par :subgoals ((populationDuring ?place ?any-time ?total) ;; This is bad, it should be more clever than this\par (percentOfBuyers ?obj ?place ?percent))\par :result-step (evaluate ?number (QuotientFn (TimesFn ?total ?percent) 100)))\par \par (defparameter *health-facts*\par '((healthInsurancePremiumForGroupInRegion UnitedStatesOfAmerica Person (DollarsPerYear 3000))\par (percentOfBuyers HealthCare UnitedStatesOfAmerica 100)))\par \par \par \par \par } %- LV=3 ND=QR-Bote-Examples DI=10 NF=000000000000000000000000 BC=clWindow IX=0 %: {\rtf1\ansi\ansicpg1252\deff0{\fonttbl{\f0\fnil\fcharset0 Courier New;}{\f1\fnil\fcharset0 Courier;}} \viewkind4\uc1\pard\lang1033\f0\fs20 ;;;; -*- Mode: LISP; Syntax: Common-Lisp; Base: 10 -*-\par ;;;; ---------------------------------------------------------------------------\par ;;;; File name: qr-bote-examples.lsp\par ;;;; System: FIRE/BotE\par ;;;; Author: Praveen Paritosh\par ;;;; Created: April 16, 2004 13:31:39\par ;;;; Purpose: Examples for BotE via solve\par ;;;; ---------------------------------------------------------------------------\par ;;;; Modified: Friday, April 30, 2004 at 22:21:49 by paritosh\par ;;;; ---------------------------------------------------------------------------\par \par (in-package :common-lisp-user)\par \par ;; How many popcorns would fill this room?\par ;; To Run the example\par ;; (fire::setup-solve-test-case *popcorn-suggestions* *popcorn-facts* :KB)\par ;; (multiple-value-setq (ans tree) (fire::solve *popcorn-query* :response :bindings))\par ;; For more solutions\par ;; (fire::get-solution tree)\par \par ;; Query \par (defparameter *popcorn-query* '(CountContained CS381ClassRoom Popcorn ?number))\par \par (defparameter *popcorn-suggestions*\par '((defSuggestion VolumeStrategyForCount\par (CountContained ?container ?contained ?count)\par :subgoals ((volumeOfObject ?container ?v1)\par (volumeOfObject ?contained ?v2))\par :result-step (evaluate ?count (QuotientFn ?v1 ?v2))) \par (defSuggestion VolumeFormulaForSphere\par (volumeOfObject ?obj ?vol)\par :test (shapeOfObject ?obj SphereShape)\par :subgoals ((radius ?obj ?radius))\par :result-step (evaluate ?vol\par (TimesFn 4.187 ;;4/3*Pi \par (ExponentFn ?radius 3))))\par (defSuggestion VolumeFormulaForRectangular3DShape\par (volumeOfObject ?obj ?vol)\par :test (shapeOfObject ?obj Rectangular3DShape)\par :subgoals ((heightOfObject ?obj ?h) \par (lengthOfObject ?obj ?l) \par (widthOfObject ?obj ?w)) \par :result-step (evaluate ?vol (TimesFn ?h ?l ?w)))))\par \par ;; Knowledge that must be around \par (defparameter *popcorn-facts* \par '((isa CS381ClassRoom ClassRoom)\par (shapeOfObject CS381ClassRoom Rectangular3DShape)\par (shapeOfObject Popcorn SphereShape)\par (heightOfObject CS381ClassRoom 3)\par (lengthOfObject CS381ClassRoom 15)\par (widthOfObject CS381ClassRoom 10)\par (radius Popcorn 0.02)))\par \par ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;\par ;; How much money is spent on Newspapers in the US?\par ;; (annualSales NewspaperCopy UnitedStatesOfAmerica (YearFn 2003) ?money)\par \par ;; To Run the example\par ;; (fire::setup-solve-test-case *news-suggestions* *news-facts* :KB)\par ;; (multiple-value-setq (ans tree) (fire::solve *news-query* :response :bindings))\par ;; For more solutions\par ;; (fire::get-solution tree)\par \par (defparameter *news-query*\par '(annualSales NewspaperCopy UnitedStatesOfAmerica (YearFn 2003) ?money))\par \par (defparameter *news-suggestions*\par '((defSuggestion PerBuyerStrategy\par (annualSales ?obj ?place ?time ?money)\par :subgoals ((annualExpensePerBuyer ?obj ?place ?time ?money-per-buyer)\par (numberOfBuyers ?obj ?place ?time ?number))\par :result-step (evaluate ?money (TimesFn ?money-per-buyer ?number)))\par (defSuggestion UnitaryStrategyForCost\par (annualExpensePerBuyer ?obj ?place ?time ?money-per-buyer)\par :subgoals ((annualUnitsBoughtPerBuyer ?obj ?place ?time ?units)\par (cost ?obj ?unit-cost))\par :result-step (evaluate ?money-per-buyer (TimesFn ?units ?unit-cost)))\par ;;; (defSuggestion FrequencyStrategy\par ;;; (annualUnitsBoughtPerBuyer ?obj ?place ?time ?units)\par ;;; :subgoals ((\par (defSuggestion FractionOfTotalStrategy\par (numberOfBuyers ?obj ?place ?time ?number)\par :subgoals ((populationDuring ?place ?any-time ?total) ;; This is bad, it should be more clever than this\par (percentOfBuyers ?obj ?place ?percent))\par :result-step (evaluate ?number (QuotientFn (TimesFn ?total ?percent) 100))) \par ))\par \par (defparameter *news-facts*\par '((cost NewspaperCopy 0.75)\par (percentOfBuyers NewspaperCopy UnitedStatesOfAmerica 30)\par (annualUnitsBoughtPerBuyer NewspaperCopy UnitedStatesOfAmerica (YearFn 2003) 365))) \par \par ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;\par ;; How Many K-8 Teachers are there in the US?\par ;; (cardinality K-8SchoolTeacher ?numteachers)\par \par ;; (fire::setup-solve-test-case *teacher-suggestions* *teacher-facts* :KB)\par ;; (multiple-value-setq (ans tree) (fire::solve *teacher-query* :response :bindings))\par ;; For more solutions\par ;; (fire::get-solution tree)\par \par (defparameter *teacher-query* \par '(cardinality K-8SchoolTeacher ?numteachers))\par \par (defparameter *teacher-suggestions*\par '((defSuggestion StudentsPerTeacherStrategy\par (cardinality K-8SchoolTeacher ?numteachers)\par :subgoals ((cardinality K-8SchoolStudent ?numstudents)\par (studentsPerTeacher K-8School ?perteacher))\par :result-step (evaluate ?numteachers (QuotientFn ?numstudents ?perteacher)))\par (defSuggestion UniformAgeDistributionStrategy\par (cardinality K-8SchoolStudent ?numstudents)\par :subgoals ((populationDuring UnitedStatesOfAmerica (YearFn 1997) ?population)\par (minimumAge K-8SchoolStudent ?min)\par (maximumAge K-8SchoolStudent ?max)\par (lifeExpectancyForGroupInRegion UnitedStatesOfAmerica \par Person (YearsDuration ?life)))\par :result-step (evaluate ?numstudents (TimesFn (QuotientFn (DifferenceFn ?max ?min) ?life) ?population)))\par ))\par \par (defparameter *teacher-facts*\par '((minimumAge K-8SchoolStudent 5)\par (maximumAge K-8SchoolStudent 14)\par (studentsPerTeacher K-8School 30)))\par \par ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;\par ;; What is the annual cost of healthcare in the US?\par ;; (fire::setup-solve-test-case *health-suggestions* *health-facts* :KB)\par ;; (multiple-value-setq (ans tree) (fire::solve *health-query* :response :bindings))\par ;; For more solutions\par ;; (fire::get-solution tree)\par \par (defparameter *health-query* \par '(annualSales HealthCare UnitedStatesOfAmerica (YearFn 2003) ?money))\par \par (defparameter *health-suggestions*\par '(;; This one gets reused. \par (defSuggestion PerBuyerStrategy\par (annualSales ?obj ?place ?time ?money)\par :subgoals ((annualExpensePerBuyer ?obj ?place ?time ?money-per-buyer)\par (numberOfBuyers ?obj ?place ?time ?number))\par :result-step (evaluate ?money (TimesFn ?money-per-buyer ?number)))\par \par ;; annual expense per buyer on a specific object. If we have discrete\par ;; units of that object that are bought separately, then it is the product\par ;; of number of units by cost per unit. on the other hand if it is something\par ;; that is paid for monthly or in installments, we want to add that all up\par ;; and find the annualExpense. For example, rent. In case of insurance, \par ;; its more sensible to assume we know the annual insurance premium \par ;; so we dont have to do much here. This example demonstrates the \par ;; issues in generality of suggestions. Suggestions have unique names, \par ;; so this one has to be named differently from UnitaryStrategyForCost\par \par (defSuggestion ExpenseOnHealthStrategy\par (annualExpensePerBuyer HealthCare ?place ?time ?money-per-buyer)\par :subgoals ((healthInsurancePremiumForGroupInRegion \par ?place Person \par (DollarsPerYear ?money-per-buyer))))\par \par (defSuggestion FractionOfTotalStrategy\par (numberOfBuyers ?obj ?place ?time ?number)\par :subgoals ((populationDuring ?place ?any-time ?total) ;; This is bad, it should be more clever than this\par (percentOfBuyers ?obj ?place ?percent))\par :result-step (evaluate ?number (QuotientFn (TimesFn ?total ?percent) 100)))))\par \par (defparameter *health-facts*\par '((healthInsurancePremiumForGroupInRegion UnitedStatesOfAmerica Person \par (DollarsPerYear 3000))\par (percentOfBuyers HealthCare UnitedStatesOfAmerica 100)))\par \par ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;\par ;; How many cars are bought per year in the US?\par ;; (fire::setup-solve-test-case *cars-suggestions* *cars-facts* :KB)\par ;; (multiple-value-setq (ans tree) (fire::solve *cars-query* :response :bindings))\par ;; For more solutions\par ;; (fire::get-solution tree)\par \par (defparameter *cars-query* \par '(unitsBoughtPerYear Automobile UnitedStatesOfAmerica (YearFn 2003) ?num))\par \par (defparameter *cars-suggestions*\par ;; This strategy works like this. There is an OBJ that people possess, and to find \par ;; out how many of these are being bought per year, assuming mostly repeat \par ;; buyers, we can say they buy it once the old one is gone bad. So we divide the \par ;; total units of product already there, by the average lifetime of the product. \par '((defSuggestion AgeStrategyForUnitsBought\par (unitsBoughtPerYear ?obj ?place ?time ?units-per-year)\par :subgoals ((unitsTotal ?obj ?place ?time ?total-units)\par (age ?obj (YearsDuration ?life)))\par :result-step (evaluate ?units-per-year (QuotientFn ?total-units ?life)))\par ;; This strategy works for products that are owned by households usually. So\par ;; if we can find the total number of households, and multiply that by the \par ;; average number of units owned by each household, we get the total units. \par ;; Weirdness in cyc knowledge: \par ;; A notable specialization of CohabitationUnit is #$FamilyCohabitationUnit. \par ;; This is closest the idea of households but not yet defined in the KB. \par (defSuggestion HouseholdStrategyForCountingUnits\par (unitsTotal ?obj ?place ?time ?total-units)\par :subgoals ((numberOfHouseholds ?place ?time ?num-households)\par (unitsPerHoushold ?obj ?units-per-household))\par :result-step (evaluate ?total-units \par (TimesFn ?num-households ?units-per-household)))\par (defSuggestion HouseholdViaPopulationStrategy\tab\par (numberOfHouseholds ?place ?time ?num-households)\par :subgoals ((populationDuring UnitedStatesOfAmerica (YearFn 1997) ?population)\par (sizeOf CohabitationUnit UnitedStatesOfAmerica ?size))\par :result-step (evaluate ?num-households (QuotientFn ?population ?size)))))\par \par (defparameter *cars-facts* \par '((sizeOf CohabitationUnit UnitedStatesOfAmerica 3)\par ;; Another example where we need feel for numbers -- in the KB it says \par ;; (YearsDuration 5 50). Now whats a typical value. We dont want to do \par ;; interval math with these things. \par (age Automobile (YearsDuration 10))))\par \par ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;\par ;; What is the weight of solid garbage thrown away by American \par ;; families each year?\par ;; (fire::setup-solve-test-case *garbage-suggestions* *garbage-facts* :KB)\par ;; (multiple-value-setq (ans tree) (fire::solve *garbage-query* :response :bindings))\par ;; For more solutions\par ;; (fire::get-solution tree)\par \par (defparameter *garbage-query* \par '(annualProduction Garbage-Generic UnitedStatesOfAmerica ?garbage-mass))\par \par (defparameter *garbage-suggestions* \par '((defSuggestion HouseholdGarbageStrategy \par \tab (annualProduction Garbage-Generic ?region ?total-garbage-mass)\par :subgoals ((numberOfHouseholds ?region ?num-households)\par (amountProducedInEvent GarbageDisposalEvent\par Garbage-Generic\par (Pound-UnitOfMass ?garbage-per-house)))\par :result-step (evaluate ?total-garbage-mass \par (TimesFn ?num-households ?garbage-per-house)))\par ;; Reuse \par (defSuggestion HouseholdViaPopulationStrategy\tab\par (numberOfHouseholds ?place ?time ?num-households)\par :subgoals ((populationDuring UnitedStatesOfAmerica (YearFn 1997) ?population)\par (sizeOf CohabitationUnit UnitedStatesOfAmerica ?size))\par :result-step (evaluate ?num-households (QuotientFn ?population ?size)))))\par \par (defparameter *garbage-facts* \par '((amountProducedInEvent GarbageDisposalEvent \par Garbage-Generic\par (Pound-UnitOfMass 5))\par ;; reuse\par (sizeOf CohabitationUnit UnitedStatesOfAmerica 3)))\par \par ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;\par ;; How many hotdogs are sold in a baseball season in Wrigley Field? \par (defparameter *hotdog-query* \par \tab '(unitsSold HotDogSandwich WrigleyField BaseballSeason ?num-dogs))\par \par (defparameter *hotdog-suggestions* \par '(;; Number of hotdogs sold per game multiplied by number of games\par (defSuggestion PerEventStrategy \par (unitsSold ?obj ?place ?event ?total-units)\par :test (subEvents ?event ?subevent)\par \tab :subgoals ((numberOfOccurancesOf ?subevent ?event ?num-subevents)\par \tab\tab (unitsSold ?obj ?place ?subevent ?units-per-subevent))\par \tab :result-step (evaluate ?total-units (TimesFn ?num-subevents ?units-per-subevents)))\par \par ;; This one is very similar to PerBuyerStrategy. Can we recast that into unitsSold? \par ;; The earlier one had annualSales and a whole annual thing about it which really \par ;; isnt necessary at all. \par (defSuggestion PerBuyerStrategy2\par \tab (unitsSold ?obj ?place ?time ?total-units-sold)\par \tab :subgoals ((numberOfBuyers ?obj ?place ?time ?num-buyers)\par \tab\tab (unitsBoughtPerBuyer ?obh ?place ?time ?units-per-buyer))\par \tab :result-step (evaluate ?total-units-sold (TimesFn ?num-buyers ?units-per-buyer)))\par ;; Reuse \par (defSuggestion FractionOfTotalStrategy\par (numberOfBuyers ?obj ?place ?time ?number)\par :subgoals ((populationDuring ?place ?time ?total) \par (percentOfBuyers ?obj ?place ?percent))\par :result-step (evaluate ?number (QuotientFn (TimesFn ?total ?percent) 100))) \par ))\par \par (defparameter *hotdog-facts* \par '((populationDuring BaseballGame WrigleyField 40000)\par (subEvents BaseballSeason BaseballHomeGame)\par (numberOfOccurancesIn BaseballSeason BaseballHomeGame 80)\par (percentOfBuyers HotDogSandwich WrigleyField 50)))\par \par \par \par \par ;; How much time would be saved nationwide by increasing the speed \par ;; limit from 55 to 65 mph?\par \par ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;\par ;;; End of Code\par \f1\par } %- LV=3 ND=Quantities in Cyc DI=23 NF=000000100000000000000000 BC=clWindow IX=0 %: {\rtf1\ansi\ansicpg1252\deff0{\fonttbl{\f0\fnil\fcharset0 Century;}} \viewkind4\uc1\pard\lang1033\f0\fs24\par } %- LV=4 ND=All instances of QuantitySlot DI=24 NF=000000000000000000000000 BC=clWindow IX=0 SS=696 %: {\rtf1\ansi\ansicpg1252\deff0{\fonttbl{\f0\fnil\fcharset0 Century;}} \viewkind4\uc1\pard\lang1033\f0\fs24 All instances of QuantitySlot (408):\par \par * accountBalance: The predicate accountBalance is used to indicate the balance of a particular account. (accountBalance ACCT BAL) means that the FinancialAccount ACCT has the balance BAL; BAL is the amount of MonetaryValue either owed by, or available to, the accountHolder (depending upon the type of account).\par \par * accountsPayable: This predicate is used to state the total amount the IntelligentAgent is due to pay. To specify the TimeInterval in which this is true use holdsIn. So (accountsPayable AGENT MONEY) means that the AGENT has due bills of total amount MONEY. Borrowed moneys are not part of accountsPayable unless they are currently due.\par \par * accountsReceivable\par * acquisitionCost\par * actionFormality\par * advanceFee\par * age\par * alertnessLevel\par * altitudeAboveGround\par * altitudeAboveSeaLevel\par * altitudeOfHighestPointIs\par * altitudeOfLowestPointIs\par * ambientPressure\par * ambientRelativeHumidity\par * ambientTemperature\par * ambientVisibility\par * amountActionPerformed\par * amountOfDamageDone\par * amountOfEnergyUsed\par * amountOfPrecipitation\par * amplitudeOfSignal\par * annualInflationRate\par * annualPropertyTax\par * applicationFee\par * appraisedValue\par * approximatePay\par * areaCrimeRate\par * areaOfObject\par * areaOfRegion\par * asCompletedValue\par * atomicCapacityOfRegion-RCC\par * attorneyFee\par * averageAmbientHumidity\par * averageDepth\par * averageDischarge\par * averagePrecipitation\par * averageProfit\par * averageRainFall\par * averageSnowFall\par * averageVelocity\par * averageYearlyIncome\par * bandwidthOfInfoPath\par * bandwidthOfInfoPathType\par * barometricPressure\par * baseRateForOngoingService\par * basicPrice\par * basicPrice-Event\par * basicPriceForType\par * basisPoint\par * birthRate\par * bloodAlcoholLevel\par * bodyHairLevel\par * bodyTemperature\par * boilingPoint\par * bookValue\par * borrowingCost\par * breakEvenPoint\par * breathingRate\par * bridgeLoadCapacity\par * businessCapital\par * canopyClosureOfRegion\par * capitalGain\par * cardinality\par * cashAssetsOfAgent\par * cashFlow\par * cashReserves\par * cashValueOfPolicy\par * chemicalPurityOfObject\par * cholesterolLevel\par * circumferenceOfObject\par * cityFuelConsumption\par * closingCost\par * clothingFormality\par * cloudCeiling\par * cohesivenessOfObject\par * commissionPercentage\par * compassValueOfDirection\par * componentCapacityOfRegion-RCC\par * compoundingPeriod\par * compressibilityOfObject\par * computationalResourcesAvailable\par * computationalResourcesSpent\par * concentrationPotentialOfForce\par * concentrationPotentialOfSet\par * concentrationPotentialOfUnit\par * conflictLevel\par * conformingRate\par * contactToxicity\par * containerCapacity\par * contractChangeFee\par * cooperationLevel\par * cost\par * costRate\par * creditAvailable\par * creditLimit\par * creditPremium\par * currentLevel\par * dailyHighTemperature\par * dailyLowTemperature\par * dealerCost\par * deathRate\par * debt\par * defenseBudgetExpenditures\par * degreeOfMasculinity\par * deliveryFee\par * densityOfObject\par * depthOfObject\par * derivedMultiplicationFactor\par * derivedProbability\par * derivedProbability-Range\par * deviationOfNormalDistribution\par * deviceReliability\par * diameterOfObject\par * diastolicBloodPressure\par * dirtinessOfObject\par * discount\par * discountPoint\par * diskCapacity\par * distanceAboveSeaLevel\par * distanceOfRotation\par * distanceTranslated\par * distributedFlow\par * downPayment\par * drainageArea\par * ductilityOfObject\par * dues\par * duration\par * earnestMoney\par * earningAsset\par * economicStability\par * effectiveRate\par * elasticityOfObject\par * entityCardinality\par * equity-Financial\par * equityInProperty\par * eventTemporalLocality\par * eventTypeTemporalLocality\par * expectedValue\par * exportValue\par * faceValue\par * fanout\par * fee\par * fertilityRate\par * fiftyTwoWeekHigh\par * fiftyTwoWeekLow\par * financeCharge\par * financialAssetsOfAgent\par * fitnessLevel\par * fixedAssetsOfAgent\par * fixedRate\par * flammabilityOfObject\par * flashPoint\par * flowCapacity\par * fluidLevel\par * fluidPressure\par * forceCapacity\par * freezingPoint\par * frequencyOfServiceToSameRecipient\par * frequencyOfWaveType\par * fuelConsumption\par * fuelTankCapacity\par * fullnessOfContainer\par * fullyIndexedRate\par * fundingFee\par * grossDomesticProduct-Slot\par * grossIncome\par * grossNationalProduct-Old\par * groupCardinality\par * groupCohesion\par * gunCaliber\par * hardnessOfObject\par * hasGlamor\par * hasPhysicalAttractiveness\par * hasSocialPower\par * headHairLength\par * heightOfObject\par * heightToReduce\par * highwayFuelConsumption\par * homeValue\par * housePrice\par * illuminationLevel\par * impactAbsorptionOfObject\par * importValue\par * income\par * incomeLimit\par * incomeQualification\par * indexedInterestRate\par * indexValue\par * infantMortalityRate\par * ingestingToxicity\par * inhalingToxicity\par * initialFee\par * insulatingAbility\par * insurancePremium\par * intangibleAssetsOfAgent\par * interestIncome\par * interestPremium\par * interestRate\par * interestRate-Blended\par * interestRate-Compound\par * interestRate-Initial\par * interestRate-Periodic\par * interestRate-Simple\par * interestRateCeiling\par * interestRateGap\par * interestRateOfNote\par * interiorFloorSpace\par * internationalMonetaryReserves\par * investmentReturn\par * iQ\par * landAreaOfRegion\par * lateFee\par * lengthOfObject\par * lengthOfObjectPerimeter\par * lengthOfTopographicFeature\par * lifetimeCap\par * likelihood\par * literacyRate\par * loadCapacity-Conveyance\par * loadCapacity-TakeOff\par * loanAmount\par * loanOriginationFee\par * longTermDebt\par * maintenanceCost\par * malleabilityOfObject\par * marginOfVictory\par * marginOnLoan\par * marketabilityPremium\par * marketPrice\par * marketRate\par * massOfObject\par * maxDepthOfObject\par * maximalDimensionOfObject\par * maximumDepth\par * maximumEntitlement\par * meanOfDistribution\par * meanOfNormalDistribution\par * measure\par * mentalActivityLevel\par * militaryAlertStatus\par * militaryManpower-Active\par * minDepthOfObject\par * minimalCrossSection\par * minimumPayment\par * monetaryValueOfThing\par * moneyTransferred\par * monthlyIncome\par * monthlyPayment\par * mortgageCredit\par * mortgageDebt\par * mortgageInsurancePremium\par * nationalBudget\par * nationalBudgetExpenditures\par * nationalBudgetRevenues\par * netFlow\par * netIncome\par * netInterestYield\par * netMagnitudeOfVector\par * netWorth\par * numberOfChromosomeSets\par * numberOfSpeakersCurrently\par * objectBasicPrice\par * objectPowerDissipation\par * objectStability\par * oilExport\par * oilPriceCharged\par * oilProduction\par * oilRevenue\par * oilSolubility\par * oilSupply\par * operatingCost\par * operatingIncome\par * owedOnAccount\par * ownsWealth\par * passThroughRate\par * pathGrade\par * paymentCap\par * pensionIncome\par * percentNaturalPopulationIncrease\par * percentPopulationInCities\par * percentUrbanPopulation\par * percentWaterByWeight\par * personalWealthOwned\par * physicalActivityLevel\par * physicalMemorySize\par * physicalStructuralIntegrity\par * pITI\par * pointsOnLoan\par * politicalStability\par * populationDensity\par * porosityOfObject\par * power-Geopolitical\par * power-Military\par * powerRating\par * pressureApplied\par * priceRange\par * primeRate-US\par * principal\par * priorProbability\par * priorProbability-Range\par * probabilityOfSentence\par * processPowerConsumption\par * profitMargin\par * projectileCaliber\par * propertyTaxRate\par * propositionLifespan\par * pulseRate\par * purchasePrice\par * qualityOfProduct\par * quantityRequested\par * radius\par * raininessOfRegion\par * rAMSize\par * rateOfFireByAgent\par * rateOfFireInEvent\par * rateOfFireOnTarget\par * rateOfInflation\par * rateOfNote\par * referralFee\par * regularExpense\par * rentalIncome\par * requiresMemory\par * resaleValue\par * residualIncome\par * resistanceOfObject\par * resistanceToBioDeterioration\par * resistanceToCorrosion\par * resistanceToRadiation\par * retailCost\par * returnOnEquity\par * revenue\par * rigidityOfObject\par * roadSpeed\par * rOMSize\par * roofCoverageOfRegion\par * salaryIncome\par * salesCommission\par * salesTax\par * seatingCapacity\par * secrecyOfOperations\par * serviceBasicPrice\par * serviceCharge\par * servicingFee\par * shearStrengthOfObject\par * shelfLife\par * sizeOfObject\par * sizeParameterOfObject\par * slopeOfRegion\par * socialClass\par * spanOfUsability\par * spatialExtent\par * speedOfCPU\par * speedOfFlow\par * speedOfObject-Translation\par * subEscrowFee\par * surfaceAreaOfTopOfObject\par * surfaceAreaOfWholeObject\par * survivalRate\par * systolicBloodPressure\par * taxExemptIncome\par * teaserInterestRate\par * temperatureOfObject\par * tensileStrengthOfSubstance\par * termOfLoan\par * textureOfObject\par * thermalConductivityOfObject\par * thermalEnergy\par * thicknessOfObject\par * thirstLevel\par * totalCharge\par * totalCharge-Event\par * totalChromosomeNumber\par * totalDisplacement\par * totalExportValue-Rate\par * totalImportValue-Rate\par * totalValueOfAssets\par * totalValueOfNonCurrentAssets\par * toughnessOfObject\par * transferCharge\par * transferTax\par * transparencyOfObject\par * tuition\par * underwritingFee\par * unitCasualties-Generic\par * unitOrientation\par * valueOfAccountsPayable\par * valueOfAccountsReceivable\par * valueOfLiabilities\par * valueOfNotesPayable\par * valueOfOwnersEquity\par * vandalismRate\par * vehicleCargoCapacity-Volume\par * vehicleLoadClass\par * ventureCapital\par * virtualMemorySize\par * viscosityOfSubstance\par * volatilityOfSubstance\par * volumeContained\par * volumeOfObject\par * waterAreaOfRegion\par * waterSolubility\par * wetnessOfObject\par * widthOfObject\par * wingSpan\par * yield-FinancialPercentage\par * yieldToMaturity\par } %- LV=4 ND=Strategies based on quantity types DI=25 NF=000000000000000000000000 BC=clWindow IX=0 SS=422 %: {\rtf1\ansi\ansicpg1252\deff0{\fonttbl{\f0\fnil\fcharset0 Century;}} \viewkind4\uc1\pard\lang1033\b\f0\fs24 Quantity Type Strategies\par \b0\par The idea here is this -- are there strategies that can be completely described with the information about the type of quantity and no domain specific knowledge? For example, a rate multiplied by time gives me the amount during that time. \par \par Types of quantities -- \par Averages\par Rates\par Annual Rates\par Flows?\par Relative values\par \par Primitive units: mass, length, time. \par Composite units: volume, density, etc. \par \par } %- LV=2 ND=QR04 Bote Paper DI=14 NF=000000100000000000000000 BC=clWindow IX=0 SS=1753 %: {\rtf1\ansi\ansicpg1252\deff0{\fonttbl{\f0\fnil\fcharset0 Times New Roman;}} \viewkind4\uc1\pard\lang1033\b\f0\fs28 Back of the Envelope Reasoning: How far does problem solving go?\par \b0\fs24 25 April 2004 17:22:46 \par \par \b Introduction \par \b0\par QR has long focused on general purpose sound and complete techniques for qualitative reasoning. Here we take a completely different approach. Back of the envelope reasoning is a kind of qualitative reasoning, albeit weak, and tries to find numeric answers. That brings us to the question -- what is a qualitative reasoning? \par \par "Qualitative reasoning is the area of AI which creates representations for continuous aspects of the world, such as space, time, and quantity, which support reasoning with very little information." [Forbus, 1996, CRC survey]\par \par There have been a bunch of representations, and reasoning techniques applied to this. Most qualitative reasoning techniques are general purpose methods, which tell us what possible inferences can be made starting from a qualitatively represented scenario. Lets consider the following reasoning -- \par \par << Pick a BotE example >>\par \par Get some cites from problem solving, heuristic problem solving, etc. Cite Pisan's Thermodynamics problem solver. His thesis has problem solving citations. How are we different from that? Is this just more domain general? Cite Jin's hybrid simulator as yet another kind of qualitative reasoning. Cite Ken's Simulations or Memories paper. Now, this paper introduces one more general problem solving technique. This wasnt really mentioned in the Simulations/Memories paper. Go back to the idea of common sense qualitative reasoning, and the constraints laid out in my QR-01 paper. Say that the goal is to add to the suite of qualitative reasoning tools. \par \par Now define the BotE domain, and why it is interesting in its own right. \par \par \b A model of Back of the Envelope Reasoning\par \b0\par A preliminary version of this was presented in [Cite QR-01, CogSci-03]. The two components to BotE -- \par \pard{\pntext\f0 1)\tab}{\*\pn\pnlvlbody\pnf0\pnindent0\pnstart1\pnqr\pndec{\pntxta)}} \fi-440\li440 Problem solving: The process that gets us from the problem to the solution. The strategy. The plan. We might know the answer, or we might have some ideas of different strategies that just might work, or we might have no clue of what to do on this problem, but know a similar problem, and then figuring out if we can borrow the solution, or how to tweak it to get to the answer for the current problem. Given a problem, figuring out what is relevant, and how to get to the solution. Might involve decomposing the problem into bits that might be known and recombining them. The job of problem solving is also to maintain a workspace where the various things that are being tried are kept track of along with the dependencies. \par {\pntext\f0 2)\tab}Feel for numbers (need better name, this is too casual): Since the final answer is numeric, at some point, the problem solving bottoms out with making numeric estimates that are easier to make, or known, as compared to the original one. Most qualitative representations are symbolic, and principled/natural/cognitively plausible mapping between the symbolic representations and numeric values have not been made. \par \pard\par \b Suggestions architecture as a framework for problem solving\par \b0\par Problem solving is a large area in AI. <> Two key components of problem solving are: \par \pard{\pntext\f0 1)\tab}{\*\pn\pnlvlbody\pnf0\pnindent0\pnstart1\pnqr\pndec{\pntxta)}} \fi-440\li440 Providing the distinction between proposing something and doing it. e.g., planning and executing. \par {\pntext\f0 2)\tab}Keeping track of what to do next, and when we are done. \par \pard\par We chose the simplest problem solving framework. Suggestions [cite] represent possible decompositions. For example, a suggestion in our implementation might look like: \par \par <>\par \par \b Literature\par \b0\par Go back and forth in citations from: \par Bhaskar, R., & Simon, H. A. (1977). Problem solving in semantically rich domains: An example from engineering thermodynamics. Cognitive Science, 1, 193-215.\par \par Broadbent, D. E. (1977). Levels, hierarchies, and the locus of control. Quarterly Journal of Experimental Psychology, 29, 181-201.\par \par Newell, A., & Simon, H. A. (1972). Human problem solving. Englewood Cliffs, NJ: Prentice-Hall.\par \par More problem solving reusable representations of strategies papers -- \par \par \b FERMI: A flexible expert reasoner with multi-domain inferencing*1\par \b0 Jill H. LarkinFrederick ReifJaime CarbonellAngela Gugliotta\par Cognitive Science\par Volume 12, Issue 1 , January-March 1988, Pages 101-138\par Expert reasoning combines voluminous domain-specific knowledge with more general factual and strategic knowledge. Whereas expert system builders have recognized the need for specificity and problem-solving researchers the need for generality, few attempts have been made to develop expert reasoning engines combining different kinds of knowledge at different levels of generality. This paper reports on the FERMI project, a computer-implemented expert reasoner in the natural sciences that encodes factual and strategic knowledge in separate semantic hierarchies. The principled decomposition of knowledge according to type and level of specificity yields both power and cross-doman generality, as demonstrated in FERMI's ability to apply the same principles of invariance and decomposition to solve problems in fluid statics, DC-circuits, and centroid location. Hierarchical knowledge representation and problem-solving principles are discussed, and illustrative problem-solving traces are presented. \b\par } %- LV=3 ND=Solve Pseudocode DI=16 NF=000000000000000000000000 BC=clWindow IX=0 SS=433 %: {\rtf1\ansi\ansicpg1252\deff0{\fonttbl{\f0\fmodern\fprq1\fcharset0 Courier New;}} \viewkind4\uc1\pard\lang1033\f0\fs24 get-solution(original-goal) \par if in-play?(original-goal) \par get-next-solution(original-goal)\par else if the agenda is empty then quit \par process-agenda \par \par get-next-solution(ao-node)\par if there are cached solutions at this node\par ;; cur-bmarker points to the current solution \par increment cur-bmarker(ao-node)\par return cached solution\par if goal-node?(ao-node)\par if ao-node has a child suggestion that is in-play\par get-next-solution(in-play-suggestion(ao-node))\par ;; No in-play suggestion, if this goal has a younger\par ;; sibling that can give us more new bindings, re-instantiate\par ;; this node with those bindings. \par if younger-siblings(ao-node)\par bindings = get-next-solution(younger-sibling(ao-node))\par if bindings found \par enqueue-on-agenda(re-instantiate-node(ao-node,bindings))\par if suggestion-node?(ao-node)\par ;; seek downward\par solution = get-next-solution(eldest-child,ao-node)\par add-bindings(solution,ao-node)\par increment cur-bmarker(ao-node)\par return solution \par \par propagate-bindings(goal-node)\par if elder-sibling?(goal-node)\par increment cur-bmarker(goal-node)\par next-node = instantiate-node(elder-sibling(goal-node),current-bindings(goal-node))\par add-to-tree(next-node)\par enqueue-on-agenda(next-node)\par return\par ;; No elder sibling, so we can now see if the parent node got solved\par if parent(goal-node) exists\par bindings = get-next-solution(goal-node)\par while we have bindings for goal-node\par result-bindings = do-result-step(parent(goal-node))\par if result-bindings found\par add-bindings(result-bindings,parent(goal-node)\par propagate-bindings(parent(parent(goal-node)))\par bindings = get-next-solution(goal-node)\par \par process-agenda(ao-node)\par if goal-node?(ao-node)\par solutions = ask(goal-node) ;; primitive problem solving step\par if solutions found \par propagate-bindings(goal-node)\par else suggestions = gather-suggestions(goal-node) \par if suggestions found\par enqueue-on-agenda(suggestions)\par else ;; no suggestions found, see if we can re-instiate this node\par if younger-siblings(goal-node)\par bindings = get-next-solution(younger-sibling(goal-node)\par if bindings found\par enqueue-on-agenda(re-instantiate-node(ao-node,bindings))\par else update-failed (goal-node)\par if suggestion-node?(ao-node)\par ;; since the subgoals are ordered, add the first one to agenda\par enqueue-on-agenda(youngest-child(ao-node))\par \par \par } %- LV=2 ND=Thesis DI=20 NF=000000000000000000000000 BC=clWindow IX=0 SS=152 %: {\rtf1\ansi\ansicpg1252\deff0{\fonttbl{\f0\fnil\fcharset0 Century;}{\f1\fnil\fcharset2 Symbol;}} \viewkind4\uc1\pard\lang1033\b\f0\fs28 Thesis plan\par \b0\fs24\par There's three threads here -- \par \pard{\pntext\f0 1)\tab}{\*\pn\pnlvlbody\pnf0\pnindent0\pnstart1\pnqr\pndec{\pntxta)}} \fi-440\li440 SOLVE: Building a library of strategies to solve BotE problems. Represent about fifty different problems, and try to build a reusable strategy library. Do an analysis of this domain [the kinds of knowledge involved in this domain] and problem solving strategies. Ken says "a really deep understanding of the functional import of equations in common sense reasoning" -- what does that exactly mean? More -- "I\rquote ll bet right now you don\rquote t have enough experience with these problems to be confident of an analysis of the general structure of the kinds of laws that there are (think about the functional analysis Yusuf did for equations for engineering problem solving; there\rquote s probably something analogous one can say there that will make a nice chapter in your thesis)." One of the implications might be extending the suggestions representation and SOLVE to handle more abstract strategies. An interesting thing to explore might be to look at using at statistics compiled from problem solving episodes to focus the problem solving, e.g., in estimating the difficulty for a strategy/goal. Will a study like this make for a journal paper by itself? \par {\pntext\f0 2)\tab}CARVE: Starting from what we said in CogSci 2004 paper, implement CARVE. Domain -- CIA factbook. Issues: a) Computational implementation, b) adding structural information about quantities (some of which could be obtained from mining the data, like corr+ and corr-). CARVE, by itself is an interesting system, and the claim here will be that it is a cognitively-plausible model of building qualitative representations. Evidence will come from existing literature and maybe maybe some experimental data. Will this make a submission for Cognitive Science? Maybe we need one more domain there? \par {\pntext\f0 3)\tab}Using the representations built by CARVE, build the Analogical Estimator. Show better SME/MAC-FAC results with these representations. Plug the analogical estimator as a primitive estimation strategy in SOLVE and show better problem solving results (one or more of -- more problems answered, better answers, or faster answers). This will be the thesis. \par \pard\par Timeline: \par \pard{\pntext\f1\'B7\tab}{\*\pn\pnlvlblt\pnf1\pnindent0{\pntxtb\'B7}}\fi-200\li200 First SOLVE, then CARVE and then the combining. I can see this happening by this year end. \par {\pntext\f1\'B7\tab}Awfully far-fetched, but are there any conferences before the next AAAI/CogSci season? \par {\pntext\f1\'B7\tab}Minor distractions which I shouldnt really do: results of the corpus analysis for dimensional adjectives. Mining the AQUAINT corpus for numbers. \par \pard\par Right now: 14 May 2004 \par \pard{\pntext\f1\'B7\tab}{\*\pn\pnlvlblt\pnf1\pnindent0{\pntxtb\'B7}}\fi-200\li200 CogSci Paper revise and submit. \par {\pntext\f1\'B7\tab}Keep working on SOLVE. \par {\pntext\f1\'B7\tab}Should I write a proposal? \par \pard\par \par \par } %- LV=1 ND=Non Research Ideas DI=6 NF=000000100000000000000000 BC=clWindow IX=0 SS=25 %: {\rtf1\ansi\ansicpg1252\deff0{\fonttbl{\f0\fnil\fcharset0 Century;}} \viewkind4\uc1\pard\lang1033\b\f0\fs28 Non Research Ideas\par \b0\fs24\par This is a place for all sorts of ideas, which are not towards my research/thesis. Some of these might be so bad that I'll pursue after I have tenure.\par \par } %- LV=2 ND=Motivequest: Mining usenet DI=8 NF=000000000000000000000000 BC=clWindow IX=0 SS=1368 %: {\rtf1\ansi\ansicpg1252\deff0{\fonttbl{\f0\fnil\fcharset0 Century;}} \viewkind4\uc1\pard\lang1033\f0\fs24 Did a google for mining usenet postings. \par \par This is something like very much like that that this guy did for programming language -- \par http://merd.sourceforge.net/pixel/language-study/usenet-traffic-ranking/\par http://www.kittyjoyce.com/eric/log/archive/000170.html\par \par This seems interesting: \par Mining Newsgroups Using Networks Arising From Social Behavior\par http://www2003.org/cdrom/papers/refereed/p688/688-agrawal/\par \par Brain Trust: Mining the Community Mind\par http://www.econtentmag.com/Articles/ArticlePrint.aspx?ArticleID=1049&CategoryID=15\par \par Data Mining and Information Extraction papers by Ray Mooney:\par http://www.cs.utexas.edu/users/ml/papers/discotex-aaai-00.pdf\par \par Style Mining of Newsgroups: \par http://www.iit.edu/~steiste1/paper2.pdf\par \par Mining the peanut gallery: Opinion Extraction and Semantic Classification of Product Reviews. \par http://www.kushaldave.com/p451-dave.pdf\par \par Babble and Loops at IBM:\par http://www.research.ibm.com/SocialComputing/babble.htm\par http://www.pliant.org/personal/Tom_Erickson/loops.chi99.paper.html\par Webfountain http://www.almaden.ibm.com/webfountain/\par Article on how it works \par http://www.spectrum.ieee.org/WEBONLY/publicfeature/jan04/0104comp1.html\par \par Warren Sack: Conversational Mapping\par http://www.sims.berkeley.edu/~sack/cm/index.html\par Position paper http://www.sims.berkeley.edu/~sack/Pubs/sack-cscw00.html\par \par Visual Text Mining: \par http://www.txtkit.sw.ofcd.com/\par } %- LV=3 ND=MIT Technology review article DI=9 NF=000000000000000000000000 BC=clWindow IX=0 SS=905 %: {\rtf1\ansi\ansicpg1252\deff0{\fonttbl{\f0\fnil\fcharset0 Century;}} \viewkind4\uc1\pard\lang1033\f0\fs24 Mining for Meaning\par Software\par \par \par By Claire Tristram\par Innovation\par July/August 2001\par \par \tab\tab\tab\tab\par \par Online newsgroups are popular gathering spots; over the years they\rquote ve logged millions of opinions on topics ranging from politics to appliances. The largest newsgroup network, Usenet, boasts 500 million messages posted since 1995; unlike postings in chat rooms and online forums, such messages tend to be uncensored-and preserved.\par \par \par All these postings add up to a trove of public opinion that sociologists, linguists and market researchers would love to analyze; and software projects at IBM and the University of California at Berkeley are beginning to develop the analytical tools they\rquote ll need. Unlike Web search engines, which try to find the best matches for any one query, these efforts focus on understanding how communities of individuals interact online, and how their opinions evolve.\par \par To begin taking on this difficult task, IBM\rquote s Babble software depicts conversations as dynamic circular graphs in which icons representing frequent talkers cluster at the center, and less chatty participants move toward the circumference. \ldblquote People do in fact cluster together when talking, then drift apart,\rdblquote says Thomas Erickson, research analyst at IBM.\par \par But that\rquote s only a first step. Beyond charting the chatters lies the task of examining what they\rquote re saying. At the University of California, Berkeley, computational linguist Warren Sack\rquote s software maps how often words or phrases appear, and how close they are to one another. \ldblquote In effect you\rquote re building a thesaurus of terms that relate directly to the conversation being studied,\rdblquote says Sack. \ldblquote You can see constellations of conversations, and see which topics are being discussed more than others.\rdblquote One test of this Conversation Map tool helped pinpoint when online participants began thinking of Gulf War syndrome as a \ldblquote disease\rdblquote rather than a cluster of symptoms.\par \par Sack and others say they\rquote re still years away from a commercial product. When the software is available, though, market researchers just might be the customers: with the right tools, they could turn newsgroups containing millions of opinions into the ultimate focus group.\par } %- LV=3 ND=Data Mining open source software DI=13 NF=000000000000000000000000 BC=clWindow IX=0 SS=292 %: {\rtf1\ansi\ansicpg1252\deff0{\fonttbl{\f0\fnil\fcharset0 Century;}} \viewkind4\uc1\pard\lang1033\f0\fs24 R-Project: http://www.r-project.org/\par Weka: http://www.cs.waikato.ac.nz/ml/weka/\par \par These folks scaled up R to do something more interesting. At least they claim to. \par http://www.stats.ox.ac.uk/pub/bdr/RipleyChen.pdf\par \par Looks like an interesting article on Data mining and tools. \par http://www.biomedcentral.com/openaccess/archive/?page=features&issue=4\par \par \par \par } %- LV=3 ND=Sentiment analysis DI=18 NF=000000000000000000000000 BC=clWindow IX=0 SS=488 %: {\rtf1\ansi\ansicpg1252\deff0{\fonttbl{\f0\fnil\fcharset0 Century;}} \viewkind4\uc1\pard\lang1033\f0\fs24 8 May 2004 10:52:38\par Identifying Collocations for Recognizing Opinions (2001)\par Janyce Wiebe, Theresa Wilson, Matthew Bell\par \par These people try to find subjectivity, whether or not a document refers to opinions (subjective) or just plain facts (objective). Probably a first step to analyzing the sentiment.\par \par Mining the Peanut Gallery: Opinion Extraction and Semantic Classification of Product Reviews (Make Corrections) Kushal Dave, Steve Lawrence, David M. Pennock\par http://citeseer.ist.psu.edu/569726.html\par This one looks real good.\par } %- LV=3 ND=Hours DI=19 NF=000000000000000000000000 BC=clWindow IX=0 SS=119 %: {\rtf1\ansi\ansicpg1252\deff0{\fonttbl{\f0\fnil\fcharset0 Century;}} \viewkind4\uc1\pard\lang1033\f0\fs24 Hours\par \par Thu -- 1 (Meeting) 5/6\par Sat -- 2\par Sun -- 3\par Mon -- 3\par Tue -- 2\par Wed -- 1 + 1 (Meeting) 5/12\par \par Next meeting Monday 5/17\par \par } %- LV=2 ND=How to do literature search? DI=7 NF=000000000000000000000000 BC=clWindow IX=0 SS=490 %: {\rtf1\ansi\ansicpg1252\deff0{\fonttbl{\f0\fnil\fcharset0 Century;}{\f1\fnil\fcharset2 Symbol;}} \viewkind4\uc1\pard\lang1033\f0\fs24 22 April 2004 \par \par Write a document about how to do a literature search as quickly as possible. Specific audience: CS. Finding conferences which are right. Which journals to follow. \par \par \pard{\pntext\f1\'B7\tab}{\*\pn\pnlvlblt\pnf1\pnindent0{\pntxtb\'B7}}\fi-200\li200 Finding relevant papers from an initial idea. \par {\pntext\f1\'B7\tab}Keeping track, following up on the research over time -- whats going on in the field?\par \pard\par Write a prelim version -- use an example from AI. and generalize, then send email to faculty/grads soliciting other information sources, etc. and then solidify this document. \par } %- LV=2 ND=Eleanor Marx DI=17 NF=000000000000000000000000 BC=clWindow IX=0 SS=26 %: {\rtf1\ansi\ansicpg1252\deff0{\fonttbl{\f0\fnil\fcharset0 Century;}} \viewkind4\uc1\pard\lang1033\b\f0\fs28 References on Eleanor Marx\par \b0\fs24\par Author, etc.: \tab Kapp, Yvonne, 1903-\par Title: \tab Eleanor Marx, by Yvonne Kapp.\par Publisher: \tab London, Lawrence & Wishart,\par Date: \tab 1972-76.\par Notes: \tab Bibliography: p. [308]-313. (v. 1)\par Incomplete contents: \tab v. 1. Family life (1855-1883)--v. 2. The crowded years (1884-1898)\par Library of Congress subject headings: \tab Aveling, Eleanor (Marx), 1855-1898.\par ISBN: \tab 0853152489\par Location: \tab MAIN Library\par Call number: \tab 335.40942 A948Zk\par Library has: \tab v.1-2\par \par Author, etc.: \tab Florence, Ronald.\par Title: \tab Marx's daughters : Eleanor Marx, Rosa Luxemburg, Angelica Balabanoff / Ronald Florence.\par Publisher: \tab New York : Dial Press,\par Date: \tab 1975.\par ISBN: \tab 0803754329\par Location: \tab MAIN Core\par Call number: \tab 335.4 F632m\par Location: \tab MAIN Library\par Call number: \tab 335.4 F632m\par \par Author, etc.: \tab Tsuzuki, Ch\'afushichi.\par Title: \tab The life of Eleanor Marx, 1855-1898: a socialist tragedy.\par Publisher: \tab Oxford, Clarendon P.,\par Date: \tab 1967.\par Library of Congress subject headings: \tab Aveling, Eleanor Marx, 1855-1898.\par Location: \tab MAIN Library\par Call number: \tab 335.40942 A948Yt\par \par \par The daughters of Karl Marx : family correspondence, 1866-1898 / commentary...\par Author, etc.: \tab Longuet, Jenny Marx, 1844-1883.\par Title: \tab The daughters of Karl Marx : family correspondence, 1866-1898 / commentary and notes by Olga Meier ; translated and adapted by Faith Evans ; introduction by Sheila Rowbotham.\par Edition: \tab 1st ed.\par Publisher: \tab New York : Harcourt Brace Jovanovich,\par Date: \tab c1982.\par Type of material: \tab Book\par Description: \tab xl, 342 p., [8] p. of plates : ill. ; 24 cm.\par ISBN: \tab 0151239711 :\par Location: \tab MAIN Library\par Call number: \tab 335.40922 L858fX\par \par Shelley's socialism / by Edward Aveling and Eleanor Marx Aveling.\par Author, etc.: \tab Aveling, Edward Bibbins, 1851-1898.\par Title: \tab Shelley's socialism / by Edward Aveling and Eleanor Marx Aveling.\par Publisher: \tab London : Journeyman Press,\par Date: \tab 1975.\par ISBN: \tab 0904526046\par Location: \tab MAIN Library\par Call number: \tab 821.7 S54Zav\par \par Author, etc.: \tab Hastings, Michael, 1938 Sept. 2-\par Title: \tab Tussy is me: a romance.\par Publisher: \tab London, Weidenfeld & Nicolson,\par Date: \tab 1970.\par Type of material: \tab Book\par Description: \tab 430 p. 23 cm.\par ISBN: \tab 0297177745\par Location: \tab MAIN Library\par Call number: \tab 823.914 H358t\par \par The woman question / by Edward and Eleanor Marx Aveling.\par Author, etc.: \tab Aveling, Edward Bibbins, 1851-1898.\par Title: \tab The woman question / by Edward and Eleanor Marx Aveling.\par Edition: \tab 4th thousand.\par Publisher: \tab London : S. Sonnenschein : Lowrey & Co.,\par Date: \tab 1887.\par Type of material: \tab Book\par Description: \tab 16 p. ; 17 cm.\par Location: \tab MAIN Library\par Call number: \tab 335 A948w\par Copy: \tab 2\par \par \par } %- LV=2 ND=Sociology of AI research DI=21 NF=000000000000000000000000 BC=clWindow IX=0 %: {\rtf1\ansi\ansicpg1252\deff0{\fonttbl{\f0\fnil\fcharset0 Century;}} \viewkind4\uc1\pard\lang1033\f0\fs24\par } %- LV=2 ND=Tversky and Kahneman DI=22 NF=000000000000000000000000 BC=clWindow IX=0 SS=742 %: {\rtf1\ansi\ansicpg1252\deff0{\fonttbl{\f0\fnil\fcharset0 Century;}} \viewkind4\uc1\pard\lang1033\b\f0\fs24 Larry's critique of TK\par \b0\par First, what Larry believes and says. Then I will write my own interpretations and additions to what he is saying and where exactly I disagree. \par \par There are two main points that Larry makes - \par \pard{\pntext\f0 1)\tab}{\*\pn\pnlvlbody\pnf0\pnindent0\pnstart1\pnqr\pndec{\pntxta)}} \fi-440\li440 Decontextualized experiments: The experiments of TK ask people to make decisions, which seem like simpler versions of real-world decision making tasks on the surface. Real world decisions are made in context, and people use a lot of contextual information in making those decisions. But TK take away all the context, and all the contextual information used is implicit. They can then manipulate these implicit contexts to get people to perform any way they want them to. \par \pard\par \par Intuition is prone to systematic errors and biases. \par \par } %- LV=1 ND=WWW DI=11 NF=000000000000000000000000 BC=clWindow IX=0 SS=51 %: {\rtf1\ansi\ansicpg1252\deff0{\fonttbl{\f0\fnil\fcharset0 Century;}} \viewkind4\uc1\pard\lang1033\b\f0\fs28 WWW\b0\fs24\par \par Stuff I find on web that I like. My annotated bookmarks file. \par } %- LV=2 ND=Stuff to buy DI=12 NF=000000000000000000000000 BC=clWindow IX=0 SS=109 %: {\rtf1\ansi\ansicpg1252\deff0{\fonttbl{\f0\fnil\fcharset0 Century;}} \viewkind4\uc1\pard\lang1033\f0\fs24 Shoes -- Dr. Martens 2B04 Series - Retro Athletic looks real good. \par http://www.zappos.com/n/p/dp/1388574/c/18289/g/men/s/7.html\par http://www.zappos.com/n/p/dp/1519663/c/13019/g/men/s/7.html\par \par Other ideas: \par http://www.zappos.com/n/p/dp/787449/c/6673/g/women/s/4.html\par http://www.hsn.com/cnt/prod/default.aspx?pfid=808073&club_id=808073&sz=20&sf=AS0423&rdr=1\par } %%