Transcription

10/24/2017Watson and IBM i :Building a cognitive businessJesse R. [email protected] is upon us.The biggest taxi companyowns no cars.The largest accommodation companyowns no real estate.The largest retailercarries no inventory.The biggest media companyowns no content.1

10/24/2017This disruption is fueled by three forces.The proliferation of differenttypes of data.The ability to buildbusiness in code withthe API econom y.The powerful capabilities andoutcomes brought on bycognitive computing.More devices are creatingmore es of code ina smartphonelines of code ina pacemakerlines of code ina new carlines of code insmart appliance2

10/24/2017Three capabilities differentiate cognitive systems fromtraditional programmed computing systems LearningReasoningUnderstandingCognitive systems understandlike humans do.They reason. They understandunderlying ideas and concepts. Theyform hypothesis. They infer andextract concepts.They never stop learning gettingmore valuable with time. Advancingwith each new piece of information,interaction, and outcome. Theydevelop “expertise”. . allowing them to interact with humans.Cognitive SystemsMACHINE LEARNING / DEEP LEARNINGWatsonNarrative 2017 International Business Machines Corporation3

10/24/2017Cognitive SystemsMACHINE LEARNING – SIMPLE EXAMPLEWatsonNarrative 2017 International Business Machines CorporationCognitive SystemsMACHINE LEARNING – SIMPLE EXAMPLEWatsonNarrative 2017 International Business Machines Corporation4

10/24/2017Cognitive systems forge a new partnershipbetween man and machine.C o g n it ive Syst em sexcel at:Humans excel at:Locating KnowledgeCommon SensePattern IdentificationMoralsNatural LanguageImaginationMachine LearningCompassionAbstractionEliminate BiasEndless CapacityDilemmasDreamingGeneralizationThe priceof not knowing. 2016 IBM Corporation105

10/24/2017Cognitive systems rely on collections ofdata and information:Data, information, and expertise createthe foundation.Examples include:Analyst reportstweetsWire tap transcriptsBattlefield docsE-mailsTextsForensic reportsNewspapersBlogsWikiCourt rulingsInternational crime databaseStolen vehicle dataMissing persons data11 and then leverage Watson APIsto apply cognitive capabilities.Retrieve and Rank50 underlying technologiesEntity ExtractionSentiment AnalysisEmotion Analysis (Beta)Keyword ExtractionConcept TaggingTaxonomy ClassificationAuthor ExtractionLanguage DetectionText ExtractionMicroformats ParsingFeed DetectionLinked Data SupportConcept ExpansionConcept InsightsDialogDocument ConversionLanguage TranslationNatural Language ClassifierPersonality insightsRelationship ExtractionRetrieve and RankTone AnalyzerEmotive Speech to TextText to SpeechFace DetectionImage Link ExtractionImage TaggingText DetectionVisual InsightsVisual RecognitionAlchemyData NewsTradeoff AnalyticsNatural LanguageClassifierTone Analyzer126

10/24/2017The market is validating the benefits ofcognitive.“No doubt, Watson has the means toradically change the industry. ““IBM Crafts a Role for ArtificialIntelligence in Medicine.”“[Watson] is specifically designed to supportthe development of a broad range ofenterprise solutions.”“IBM Watson represents a boldtechnological and visionary step”“What is distinctive about IBM is thebreadth of its effort to create Watson tools for a wide range of developers.”“ it’s not just AI algorithms themselves thathave improved, but the ability to deliver them”“IBM’s [Watson] can help banks withcomplex financial operations and attackimportant health care problems.”‘You can't do this without Watson. -Former SunCEO Scott McNealy. His startup, Wayin, usesWatson to trawl and drag photos.“The worldwide cognitive software platformsmarket will grow to 30 billion by 2018, at aCAGR”IDC: Worldwide Cognitive Software Platforms Forecast, 2015-2019: TheEmergence of a New Market (#258781, September 2015, DavidSchubmehl)IDC: IBM’s Go-to-Market Transformation – Deeper, Wider, Newer(#AP257527, April 2015, Chris Zhang, Sabharinath Balasubramanian, MayurSahni)137

10/24/2017WatsonNarrative8

10/24/2017WatsonNarrativeWatson and IBM iWatsonNarrative189

10/24/2017Connecting IBM i to WatsonServices & APIs 2016 International Business Machines CorporationIBM i and Watson Data Connect for IBM i– DB2 Web Query for i Asking questions using SQL Connecting from open source languages, RPG, etc. Node.JS and Python toolkits for Bluemix Integrated Web Services (IWS) 2016 International Business Machines Corporation10

10/24/2017Connecting IBM i to WatsonServices & APIsDataConnect 2016 International Business Machines CorporationBluemix’s Data ConnectAddedApril, 201722 2016 International Business Machines Corporation11

10/24/2017Moving data from DB2 for i to Watson Analytics23 2016 International Business Machines CorporationWatson Analytics24 2016 International Business Machines Corporation12

10/24/2017Learning how to ask a good questionWatson Analytics & DB2 for i data25 2016 International Business Machines CorporationControlled visualizationWatson Analytics & DB2 for i data26 2016 International Business Machines Corporation13

10/24/2017Moving data directly from DB2 for i into WA27 2016 International Business Machines CorporationAnalysis and Discovery28 2016 International Business Machines Corporation14

10/24/2017Many options for sharing insights29 2016 International Business Machines CorporationIncorporating Social Media content 2016 International Business Machines Corporation15

10/24/2017#IBMi Tweets 2016 International Business Machines Corporation#IBMi Tweets 2016 International Business Machines Corporation16

10/24/2017#IBMi Tweets 2016 International Business Machines Corporation#IBMi Tweets 2016 International Business Machines Corporation17

10/24/2017#IBMi Tweets 2016 International Business Machines CorporationWhat does it cost? (watson.analytics.ibmcloud.com)Visit this site for complete pricing ics/purchase36 2016 International Business Machines Corporation18

10/24/2017Data security in Watson AnalyticsDetails about the data security and other topics are answered requently-asked-questions-2/37 2016 International Business Machines CorporationIncorporating Social Media content38 2016 International Business Machines Corporation19

10/24/2017Incorporating Social Media content39 2016 International Business Machines Corporation#IBMi Tweets YTD40 2016 International Business Machines Corporation20

10/24/2017Watson and Data PreparationData needs to be in a certain state for input into Watson Services. And if this is an ongoing part of youranalytics strategy, automating the Data prep will be desired and may consist of: Data Consolidation‒ Data spread across multiple application databases, multiple servers/LPARs – even some that is not inDB2 for i Data Reorganization‒ Watson services may expect data to be in certain formats‒ Creating single table/file or publishing XML or JSONUntangling Data‒ Only the RPG programmer understands how the data is stored:‒ “If field COMPANY 001, join to File B, else join to File C” logic;‒ Dates stored in non date data types‒ Multiple data elements stored in a single field41 2016 International Business Machines CorporationThe Data Warehouse Data Warehousing or Data Mart Automation‒ The Data Warehouse is still a key system of insight tosupport advanced analytics‒ Untangled, cleansed data is automatically added to thewarehouse nightly‒ The data warehouse becomes the feed for many analytics– advanced and classic Business IntelligenceAutomates processes to ingest or publish unstructured databetween DB2 and Watson And automates all the data prep A meta data driven solution also documents the rules theRPG programmer built into the applicationConventional Data PlatformSystems ofRecordRDBMS OLTPACIDStructuredDataSystems ofInsightData Warehouse& MartsStructuredDataTransactions and Queries42 2016 International Business Machines Corporation21

10/24/2017DB2 Web Query DataMigrator ETL Extension Meta Data Driven Data Prep‒ Automate consolidation, organization, “untangling” and optionally, the build ofa data warehouse‒‒‒‒Consolidate data from many different data sourcesBuild transforms through SQL, RPG, use of built in or DB2 functions, etc.Can create flat files, relational, incorporate blobs (e.g., XML, JSON)Run data flows off of IBM i job scheduler ETL (Extract, TRANSFORM, and Load)‒ All components run in IBM i‒ Multiple load types can be defined INTEGRATED with DB2 Web Query‒ Can leverage meta data created with DB2 Web Query or vice versa‒ Shared services and administration43 2016 International Business Machines CorporationPreparing Data For Watson; Visualizing Insights with DB2 Web QueryIBM Watson AnalyticsGain insights fromWatson AnalyticsVisualize/deploy insights withDB2 Web QueryBluemix Data ConnectDB2 Web Queryfor i withDataMigratorIBM BluemixDB2 for iDB2 for LUWDB2for iOracleMS SQL ConsolidateOn-premiseDataSecure GatewayPostgreSQLPower Systems(IBM i) 2016 International Business Machines Corporation22

10/24/2017Data ConnectServices & APIsDataConnect 2016 International Business Machines CorporationData Connect 2016 International Business Machines Corporation23

10/24/2017For the Developer Services & APIsSQL 2016 International Business Machines CorporationFor the Developer Services & APIsSQL 2016 International Business Machines Corporation24

10/24/2017For the Developer Services & APIsIBM i languages:Python, Node.JS, RPG 2016 International Business Machines CorporationWatson API Explorer Use Watson API Explorer to test APIs and view live responses from the server.– Watson API et/“Language Translator” of Watson API Explorer is used in our sample program written by ILE RPG. 2016 International Business Machines Corporation25

10/24/2017Watson API : Watson Developer Cloud Enable cognitive computing capabilities in your app using IBM Watson‘s Language,Vision, Speech and Data APIs.LanguageVisionSpeechData InsightsNatural Language ClassifierClassify text sentencesRetrieve and RankReturn answer candidates for natural language questionsConversationAutomate interaction with end users byadding natural language interface toapplicationPersonality InsightsEstimate an individual's characteristics fromtextTone Analyzer (Unsupported Japanese)Analyze text emotion, sociability and styleDocument ConversionConvert a document to a new formatNatural Language Understanding (Unsupported Japanese)understand the language of short texts and make predictionsabout how to handle them.Language Translator (Partially Unsupported Japanese) ※1Translate text from one language to anotherVisual RecognitionDetect meaning included in image contentsSpeech to TextConvert speech to textText to SpeechConvert text to speechDiscovery (Unsupported Japanese)Add cognitive search and content analysis engines to applications to identify patterns, trends, and actionableinsights that help to make better decisionsTradeoff Analytics (Unsupported Japanese)Support tp make better choices when faced with services-catalog.html※1 For the Language Translator, only in the news domain Japanese is supported. 2016 International Business Machines CorporationConnecting RPG to Watson Details on how to the following link.– Paul Tuohy ”RPG TALKS TO WATSON” Copyright 2017 IT 6-story01/ 2016 International Business Machines Corporation26

10/24/2017Language Translator- Request URL Clicked on the link for Language translation Under Translate/Get, Click on the option to"Translates the input text from the sourcelanguage to the target language."– Input the following and clickmodel id : ja-entext : This is a test. “Request URL” is displayed– ranslator/api/v2/translate?model id enja&text This%20is%20a%20test.– The variable parts being the from and twolanguages (ja-en) and the encoded text(%20 is the encoding for a space). 2016 International Business Machines CorporationRPG Sample Program using Watson API This ILE RPG sample application uses the Watson API "Language Translator"– Translate the original sentence with “Language Translator” and output the resulton the 5250 screen※ Not recommend that your programshould be green screen, don't wantanyone to get the impression that there isany special web configuration, oranything like that, involved in thisprocess. 2016 International Business Machines Corporation27

10/24/2017ILE RPG Sample Program using Watson API- Cooperative image of ILE RPG and Watson API Use DB2 for i HTTP functions available on IBM i 7.1 and later– Provide REST HTTP method sample SQL procedure (function) in "SYSTOOLS" schema– REST call to Watson API using HTTPGETBLOB PGLE 原文の言語SYSTOOLS.HTTPGETBLOB言語HTTP GET RequestLanguage net/languagetranslator/api/v2/translate?model id enja&text This%20is%20a%20test. ,’’)CLOBGet resource 訳文DB2for iIBM iPower Systems 2016 International Business Machines CorporationRPG Sample Program using Watson API- Display File Sample source : Display File “TOWATSOND.DSPF”– Input fieldsFROMLANGTOLANGFROMTEXT– Output fieldsTOTEXTSQLCODEO– F3 : exit the programAAAAAAAAAAAAAAAAAAAAAAAINDARADSPSIZ(24 80 *DS3)CF03(03)R QLCODEO50O1 27'Watson翻訳'COLOR(BLU)3 1'言語 (1 英語, 2 スペイン語,'3 35'3 フランス語, 4 イタリア語,'3 67' 5 日本語)'4 1'原文の言語:'4 15VALUES(1 2 3 4 5)EDTCDE(X)4 20'訳文の言語:'4 34VALUES(1 2 3 4 5)EDTCDE(X)6 1'原文:'8 1CHECK(LC)13 1'訳文'14 119 1'SQLCode:'19 1024 2'F3 Exit'COLOR(BLU) 20