changing,the,compu5ng,landscape,infoapps questions collect and normalize: parse, tag, index,...

30
©Synthexis, LLC Changing the Compu5ng Landscape Sue Feldman, CEO Synthexis [email protected]

Upload: others

Post on 25-Aug-2020

4 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis  

©Synthexis,  LLC    

Changing  the  Compu5ng  Landscape    Sue  Feldman,  CEO  Synthexis  [email protected]  

Page 2: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis   2  

Agenda  •  The  technology  dilemma  today  •  Five  technology  trends  that  will  change  the  compu5ng  landscape  -  Integrated  informa5on  management,  access  and  analysis  -  Dynamic  compu5ng  for  a  changing  informa5on  landscape  �  Probabilis5c  compu5ng  �  Adap5ve  learning  �   Analy5cs  and  big  data  -  Context,  Conversa5on,  InfoApps  •  Integrated  informa5on,  integrated  technologies    

Page 3: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis   3  

The  problem  

•  Too  much  informa5on  •  Changing  quickly  •  ScaLered  among  separate  applica5ons  silos  •  Hard  to  get  a  complete  picture  of  business,  market,  compe55on,  economy  •  Hidden  surprises,  opportuni5es  and  risks  

Page 4: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis   4  

Trends  that  are  Changing  the  Compu5ng  Landscape  •  Integrated  informa5on  plaOorms  •  Dynamic  Compu5ng  -  ShiP  to  probabilis5c  compu5ng  -  Learning  systems  -       Big  data  and  analy5cs  •  Context,  Conversa5on  and  InfoApps  

Source:  The  Answer  Machine  by  Susan  Feldman.    Morgan  &  Claypool  2012.  morganclaypool.com  

Page 5: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis   5  

Page 6: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis   6  

Enterprise  SoPware  Landscape  today  

Databases  

eDiscovery  

Search  

Business  

intelligence  

Reports  

HR    

ERP  

Sales  &  Marke5ng  

VoC  

WCM

 

Archiving  

email  

Page 7: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis  

Collect

Organize

Question

Analyze

Discuss  

Decide  

Tools

Answers

S o u r c e s

In teract

Questions

Integrated  Informa5on  PlaOorm  Processes  

Page 8: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis  

Questions Answers

Prepare Data: •  Normalize •  Extract •  Tag •  Parse •  Index

Analyze/Predict: •  Relationships •  Cause and Effect •  Trends •  Patterns •  Anomalies

Analyze Question •  Disambiguate •  Expand •  Hypothesize

Access and Discuss: Collaboration, Search, BI, Text

Analytics, Databases, Analytics, Workflow, Hadoop/MapReduce, etc.

Visualization

S o u r c e s

Answers

In foApps

Questions

Collect and Normalize: Parse, tag, index, extract, structure

Knowledge bases, rules engines

Role of Text Analytics

Explore,  Discuss,    Iterate,  Collaborate  

Nego;ate  Ques;on  

Page 9: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis  

Trend  2:    Shi@  to  Probabilis;c  Compu;ng  

9  

Dynamic  Compu;ng  

Page 10: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis   10  

Probabilis5c  Compu5ng  

Predicts  the  likelihood  that  an  event,  document  or  diagnosis  matches  a  query,  paLern,  rule  or  model  -  Finds  similar  and  exact  matches  -  Returns  best  matches  with  confidence  scores  based  on  evidence  in  data  -  Uncovers  paLerns  and  surprises  •  Scalable.    Ad  hoc  •  Search-­‐like  architecture  Suitable  for  text,  images,  or  collec3ons  of  varied  informa3on,  including  structured  data  

Page 11: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis  

Sta5c  vs.  Dynamic  Compu5ng  Sta;c   Dynamic  

11  

                                                   Determinis;c                  Probabilis;c  

Pre-­‐defined  parameters   Ad  hoc  parameters  

Exact  matching   Similarity/fuzzy  matching  

Monitor  known  processes  or  events  (sales,     Confidence  scores  

Schemas   Schema-­‐less  

Pre-­‐structured  data   Structured  by  query  or  clustering  

Preset  reports   Interac5ve  and  itera5ve  

Precise   Surprises  

Historical   Current  and  future  

Both  are  needed!  

Page 12: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis  

Trend  3:    Adap;ve  Learning  

12  

Dynamic  Compu;ng  

Page 13: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis  

Adap5ve  Learning  •  Adapt.  Learn  from  new  data  and  from  mistakes  •  May  be  trained  using  a  set  of  examples  for  each  paLern  or  category  

•  Categorize  new  data  to  match  closest  learned  paLern  •  Scalable.  Accommodate  large,  diverse  data  stores  •  More  data  improves  accuracy  •  Detect  changes  and  anomalies  •  Consistent,  without  human  biases  •  Key  ingredient  in  big  data  applica5ons    Examples:    Categoriza3on,  machine  transla3on,  recommenda3on  systems  

13  

Page 14: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis   14  

Dynamic  Compu;ng  

Page 15: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis   15  

Big  Data  and  Analy5cs  

 A  set  of  technologies  that:  -  Solve  complex  informa5on  problems  economically.      

-  Collect,  manage,  mine,  analyze  and  model  large  volumes  or  streams  of  informa5on.  -  Integrate  informa5on  from  varied  sources  for  deeper/broader  understanding  -  Deliver  faster  and  beLer  understanding  of  phenomena:  weather,  diseases,  customers,  products,  risks,  marketplace  trends,  etc.  

-  Enable  high  velocity  data  capture,  discovery  and  analysis  

Page 16: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis  

Big  Data  and  Analy5cs:  The  Informa5on  Advantage  

16  

Micro  segmented  data  for  beLer  healthcare,  more  targeted  sales,  beLer  search  results,  beLer  customer  support.

•  More  data  and  data  types  è    more  clues,  beLer  targe5ng  •  Beyond  canned  reports  è  ad  hoc  discovery  •  Finds  paLerns  in  data  across  sources    

Page 17: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis  

From  Sta5c  to  Dynamic:  A  Spectrum  of  Uses  

17  

Monitor  social    media  Predict  healthcare  

 treatment  outcomes  

Climate  modeling    &  predic5on    

Voice  of  the  customer  Detect/  predict  

 investment  trends  

Fraud  detec5on  &  predic5on    Web  search  

Government  intelligence  

Monitor  and  respond    to  sensor  data  

Past   Future  Present  

eCommerce  recommenda5ons  

Report  

Drug  discovery  Innova5on  support  

Monitor,  Find,  Analyze   Predict  

Churn  preven5on  

Plant  management  

eDiscovery  

Emergency  services    planning  and  support  

Early  product  warning  Traffic  management  

Sales  and  marke5ng  reports  

Decision  support  Financial  reports  

Page 18: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis   18  

Trend  5:  Context,  Conversa;on,  and  InfoApps  

Page 19: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis   19  

Role  of  Context     User  

Trends,  Alerts,  Analyses  decision  support,    

Answers  

Context:  task,  personal  profile,  permissions,      

Adap5ve  learning  systems  

Databases  

Content  analy5cs:  text,  speech,  

image  Business  

intelligence  

Predic5ve  analy5cs  

Search  Apps:  CRM,  VoC,  Web  Analy;cs,  CM  PLM,  ERP,  Finance,  Sales,  Marke;ng  

Rules,  Models  

Knowledge  bases  

Page 20: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis   20  

Context,  Conversa5on,  InfoApps  Create  specialized,  integrated  informa5on  work  environment  specific  to  a  process  or  task  •  Contextually  aware  •  Conversa5onal  and  stateful  •  Integrate  mul5ple  informa5on  sources,  mul5ple  technologies  •  Trained  on  knowledge  bases  for  terminology,  processes  •  Intui5ve  UI  hides  complexity  of  underlying  plaOorm  -  Interac5ve  charts,  graphs,  maps,  visualiza5ons  -  Implicit  queries  -  Natural  user  interac5on:  voice,  typing,  clicking,  gestures  -  Mul5ple  devices  with  appropriate  interfaces  -  Remembers  what  you  were  doing  no  maLer  which  device  you  use  

 

Page 21: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis   21  

Cars  for  Techies  

Page 22: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis   22  

Do  users  have  to  be  mechanics?  

Page 23: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis   23  

Cars  for  Transporta5on  

Page 24: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis  

Examples  

24  

Page 25: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis   25  

Number  of  reported  flu  cases  over  last  5  years  with  type  of  flu,  mortality.   Data  sources:    

•  Social  media  discussions  

•  News  media  Text  extrac3ons  of    loca3on,  3me,  disease,  sen3ment,  social  network  analysis      

Past   Future  Present  

BI/Repor:ng  

Data  sources:  •  Hospital  records  •  Physician  reports  •  Mortality  records  •  News  

Search    &  Pa?ern  Detec:on  

Predic:ve  Modeling  &  Analy:cs  

Emergency  services    planning  and  support  

Discover  emerging  disease  trends  Alert  to  new  reports  

Plan  for  epidemic  management  

Predict  spread  of  disease  

Data  sources:  •  Historical  sources  •  Epidemiology  models  •  Physician  reports  •  Mortality  records  •  Healthcare  guidelines  

Example:    Tracking  the  Flu  

Page 26: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis  

Example:    eCommerce  

26  

Report  on  brand  popularity,  customer  opinion    

Data  sources:    •  Customer  data  •  Social  network  analysis    •  Social  media  discussions  •  News  media  

Past   Future  Present  

BI/Repor:ng  

Data  sources:  •  Transac5ons  •  Web  analy5cs  •  Social  media  

Search    &  Pa?ern  Detec:on  

Predic:ve  Modeling  &  Analy:cs  

Product  planning,    inventory  management  BeLer  search  and  

browsing  based  on  current  buying  trends,  seasonal  informa5on   Recommenda5ons  

Predict  customer  churn  

Data  sources:  •  Customer  data  •  Click  analysis  •  Weather  correlated  

with  purchasing  

Page 27: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis  

Google  Disease  Alert  Map      Presents  data  for  quick  understanding  

Google:    Map  as  Context  

Page 28: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis  

Text  box  is  Calibri  24  

IBM  Watson:    Big    Data,  Conversa5on  and  Context  

Page 29: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis  

Key  Points  •  Gather  everything  •  Filter  by  context:  purpose,  user,  task,  5me,  loca5on  •  Know  what  you  want:    Reports?    Alerts?  Predic5ons?  PaLerns?  •  Avoid  building  in  biases  and  preconcep5ons.    Use  adap5ve  learning  to  reveal  paLerns  in  the  data  •  Discovery  –  of  surprises,  paLerns,  trends,  and  rela5onships  among  people,  products,  and  events-­‐-­‐is  a  key  element  •  Successful  organiza5ons  use  analy5cs  to  out-­‐compete  their  peers  •  Big  Data  and  cogni5ve  compu5ng  techniques  make    organiza5ons  more  nimble  and  flexible  •  More  informa5on  can  be  an  advantage  

29  

Page 30: Changing,the,Compu5ng,Landscape,InfoApps Questions Collect and Normalize: Parse, tag, index, extract, structure Knowledge bases, rules engines Role of Text Analytics Explore,Discuss,

Synthexis  

©Synthexis,  LLC    

Contact  informa5on:  

Sue  Feldman,  CEO  Synthexis,  LLC  [email protected]  TwiLer:  @susanfeldman  

Sue  Feldman,  CEO  Synthexis,  LLC    [email protected]  TwiLer:  @susanfeldman  617-­‐651-­‐0275  

Insert spring picture here