Artificial Intelligence: Evolution and its Impact on Media and Marketing
August 2014
What is Big Data?
BIG
DATAa term to classify data sets
that are too large, complex
and dynamic for any
conventional data tools to
capture, store, manage
and analyze
The VOLUME, VARIETY and VELOCITY of the data creates
challenges for data processing systems
Volume Variety
Velocity
500
Million
Tweets
Per Day
204
Million
Emails
Per Minute
3.5
Billion
Google searches
a day
People to People
Social networks, blogs, forums
People to Machine
Digital TV, E-Commerce, Bank
Cards
Machine to Machine
GPS Devices, Bar Code Scanners
The volume of digital data
stored is equivalent to every
person in the world tweeting
3 tweets per minute for 765
years nonstop
Marketers Are Not Prepared To Handle
Big Data…
• Brands have large pools of data from loyalty cards, receipts,
medical records, tax returns, etc. but most are challenged in
harnessing insights from this data
Data is only as good as the insights pulled out and
there’s too much data for a human-directed
manual system to take on
• 71% of CMOs say their organization are not prepared to deal
with the explosion of big data over the next few years
• The richness and nuances of human communication
make it difficult to represent info as data
And not only is it a lot, it’s complicated
…So Artificial Intelligence Systems are
Beginning to Decode Data
• Artificial Intelligence Systems are
computer systems that exhibit
problem-solving and decision making
capabilities that emulate human
intelligence
• AI is helping marketers and
advertisers gain insights from
consumer data collected by social
networks, purchase habits, mobile
footprints, and more
Key Terms in AI
Machine Learning. Systems that can learn
without being explicitly programmed by detecting
patterns and adjusting actions accordingly
• Example: Facebook's News Feed content
algorithm changes according to the user's
personal interactions with other users
Deep Learning. A subcategory of machine
learning that deals with the use of neural networks
to improve speech recognition, image recognition,
and natural language processing
• Example: Google Now learns a users’ search
and travel patterns to proactively promote
content / functionality on a consumer’s
mobile device
Key Terms in AI and Applications
Back-Propagation. Training
a computer to match images with
labels or tags that were pre-
defined (if enough pictures of
cats are uploaded and tagged the
system will start to identify new
photos of cats)
• Example: Facebook
suggests people to tag in
photos, without you doing
it manually
Text Mining. Analyzing
natural language in documents,
e-mail messages and other free-
form text.
• Example: Google scans
incoming and outgoing
emails for information to be
used for ads
Sentiment Analysis.
Aims to determine the
attitude of a speaker or
writer.
• Example: Airlines monitor
Twitter in real-time to
assess customer feedback;
AI is used to corral and
queue those responses for
customer service reps
Image Recognition.
Technology that strives to
identify objects, people,
buildings, places, logos, etc.
• Pinterest uses to
enhance search and
power the “Explore
Interest” page based
on historical pins.
• Music Xray can predict if a song will turn into a
pop hit and discover artist with similar sounds.
– Algorithm successfully predicted success of Norah Jones
and Maroon 5
– Has given 5,000 artists opportunities in the music industry
– May eventually rule the world of pop music
• Iamus- a computer cluster powered by
Melomics’ technology- is capable of composing
contemporary classical music without human
aid in 8 minutes
– Has been taught general technics and plausible actions
– Starts with a small initial input and then uses a complex
algorithm create a full composition
– Can adapt to best fit the aesthetic criteria provided
Real World AI: Music Industry
• Medical data repositories apply AI to
electronic records to enhance diagnosis and
treatments based on past health history and
treatments.
• AI is being used to predict which patients
were likely to need certain medications in
the future.
– An Indiana University study showed
that using AI for patient care shows
outcomes that were 50% better and
more cost effective.
• AI is also being used to optimize staffing
schedules based on patient schedules and
historical needs.
Real World AI: Healthcare
Click here or here to learn more
about AI in Healthcare
AI is Revolutionizing Everything!
Click to Explore
Journalism Marketing Cinema
Culinary SportsConsulting Video Games
Insurance
Media and Publishing: The Robot Journalist
– AP uses for corporate earnings stories for
business news report in conjunction with
data from Zacks Investment Research
– Used by Edmunds to generate car
descriptions
– Yahoo uses to write Fantasy Football recaps
– Creates real-estate listings
– Internal sales reports for Fortune 500
companies
– Forbes uses to publish earning previews
Click here or here to learn more about
“The Robot Journalist”
Click to Learn More
Media and Publishing: The Robot Analysts
AI Analytics systems take big data sets &
produce comprehensible information
and insights
• AI Analytics benefits:
– Increase efficiency and productivity of
analysis and reports
– Greater consistency in logic, structure
and narratives
– Experts freed from laborious, repetitive
data analysis
• Zenith currently uses AI for competitive
analysis using data from around 400-500
unique ads with 40K-50K airings (including
spot TV, local cable, and local radio) to create
weekly summary reports for large markets.
Click to Learn More
Media and Publishing: The Robot Media
Buyer
Click the Robot to download a
programmatic 101
Media and Publishing: The Robot Shrink
Emotional tone, voice intonations and facial
expressions can be extremely telling of a consumer
attitude towards a brand, event or person
• Utilize deep learning/ image recognition software
• Uses for sentiment analysis
– Customer Service
• Many brands use automated bots on Twitter,
sometimes with unintentionally awkward results.
Click here for more info
– Market Analysis and Trading Desks
• Monitor social buzz and important news that could
move stock by analyzing sentiment and emotion in
human language. Click here or here to learn more.
Media and Publishing: The Robot Autuer
Reasons For Image Recognition
– Facilitate Mobile Commerce
– Customer Engagement
– Form insights on consumer
attitudes/behaviors based on logo detection
Companies Using Image Recognition
Technology:
– Pinterest: enhancing search and develop
relevant ad product
– Google: “Reverse Image Search”
– Amazon: Firefly technology
– Target: Scan magazine ads to add to online
shopping cart
Click here or on brand names for more info
Media and Publishing: The Robot Chef
Chef Watson
– Discovers and invents new dishes
based on what it’s learned from
recipes plus an understanding of
what chemical flavors go together
– Input fundamental ingredients, the
kind of dish desired and a style to
create a full list of ingredients and
suggested preparation instructions
Click here, here or here to learn more
about “The Robot Chef”
• In terms of direct, applicable use TODAY for AI in marketing and advertising, the most valuable
applications can be found in programmatic, data-driven digital campaigns
• Contact your buyer for more information on this topic
• On the horizon, image recognition AI is extremely promising for advertising on visual social
platforms such as Instagram, Pinterest, and even Snapchat
• AI can identify logos, retail outlets, products, etc. in photo and video content, produce
structured data around them, and evolve to the point where media buyers will be able to
buy the “visual graph” in addition to the social graph, the search graph, etc..
• Long term, AI shows promise in dynamic delivery based on mood, sentiment, and intent on
traditional platforms such as video through addressable advertising.
Implications for Marketing And Advertising
• AI is extremely versatile in its use and applications for business beyond
marketing, and extends to streamlining business processes, customer
service, CRM, publishing, and marketing automation
• Companies who successfully are able to harness AI in being able to process
big data, will have a huge advantage over their competitors as it evolves
AI Is Not Just About Advertising, It’s About Your
Business
Hot Companies In AI: Click To Explore
Want More?
K-Yun Steele
K-Yun.Steele@zenithmedia-na.com
212-858-6155
Jennifer Loeb
Jennifer.Loeb@zenithmedia-na.com
212-859-1842
Sources Used• http://blogs.datadirect.com/2014/04/2014-data-connectivity-outlook.html
• http://www.fastcolabs.com/3026423/why-google-is-investing-in-deep-learning
• http://realcomm.com/advisory/advisoryPrint.asp?AdvisoryID=544
• http://whatis.techtarget.com/definition/machine-learning
• http://wikibon.org/blog/big-data-statistics/
• http://www.emc.com/about/news/press/2011/20110628-01.htm
• http://www.huffingtonpost.com/2013/08/26/image-recognition-apps_n_3815554.html
• http://www.theguardian.com/technology/2014/apr/15/gmail-scans-all-emails-new-google-terms-clarify
• http://io9.com/5973551/this-classical-music-was-created-by-a-supercomputer-in-less-than-a-second
• http://techland.time.com/2013/01/04/finally-a-computer-that-writes-contemporary-music-without-human-help/
• http://www.wfs.org/futurist/2013-issues-futurist/may-june-2013-vol-47-no-3/pop-goes-algorithm
• http://www.smh.com.au/it-pro/cloud/social-media-analytics-predicting-box-office-success-20140228-hvfg6.html
• http://www.marketplace.org/topics/business/whats-behind-future-hit-movies-algorithm
• http://www.ox.ac.uk/media/news_stories/2013/130822_1.html
• https://www.youtube.com/watch?v=1hkMhng4q74#t=132
• http://mashable.com/2014/04/23/data-online-every-minute/
• http://www.internetlivestats.com/google-search-statistics/
• http://nymag.com/daily/intelligencer/2014/07/why-robot-journalism-is-great-for-journalists.html
• http://www.slate.com/articles/technology/technology/2014/07/automated_insights_to_write_ap_earnings_reports_why_robots_can_t_take_journalists.html
• http://www.fastcodesign.com/3032501/ibms-watson-is-now-a-cooking-app-with-infinite-recipes#1
• http://www.slate.com/articles/technology/future_tense/2014/06/chef_watson_with_bon_appetit_ibm_s_supercomputer_enters_the_culinary_world.html
• http://www.bonappetit.com/test-kitchen/inside-our-kitchen/article/chef-watson-in-the-ba-test-kitchen
• http://www.npr.org/blogs/health/2013/05/20/182593855/if-your-shrink-is-a-bot-how-do-you-respond
• http://www.nbcnews.com/business/consumer/missingthepoint-companies-alienate-customers-twitter-bots-scripted-responses-f6C10760870
• http://techcrunch.com/2014/02/03/twitter-raises-its-enterprise-cred-with-thomson-reuters-sentiment-analysis-deal/
• http://www.mobilemarketer.com/cms/news/software-technology/18380.html
• ARRIA. “Articulate Intelligence” http://www.arria.com/
• BI Intelligence. “Social Media’s New Big Data Frontier- Artificial Intelligence, Deep Learning, and Predictive Marketing. February 2014
Artificial Intelligence: Evolution and its Impact on Marketing

Artificial Intelligence: Evolution and its Impact on Marketing

  • 1.
    Artificial Intelligence: Evolutionand its Impact on Media and Marketing August 2014
  • 2.
    What is BigData? BIG DATAa term to classify data sets that are too large, complex and dynamic for any conventional data tools to capture, store, manage and analyze The VOLUME, VARIETY and VELOCITY of the data creates challenges for data processing systems Volume Variety Velocity 500 Million Tweets Per Day 204 Million Emails Per Minute 3.5 Billion Google searches a day People to People Social networks, blogs, forums People to Machine Digital TV, E-Commerce, Bank Cards Machine to Machine GPS Devices, Bar Code Scanners The volume of digital data stored is equivalent to every person in the world tweeting 3 tweets per minute for 765 years nonstop
  • 3.
    Marketers Are NotPrepared To Handle Big Data… • Brands have large pools of data from loyalty cards, receipts, medical records, tax returns, etc. but most are challenged in harnessing insights from this data Data is only as good as the insights pulled out and there’s too much data for a human-directed manual system to take on • 71% of CMOs say their organization are not prepared to deal with the explosion of big data over the next few years • The richness and nuances of human communication make it difficult to represent info as data And not only is it a lot, it’s complicated
  • 4.
    …So Artificial IntelligenceSystems are Beginning to Decode Data • Artificial Intelligence Systems are computer systems that exhibit problem-solving and decision making capabilities that emulate human intelligence • AI is helping marketers and advertisers gain insights from consumer data collected by social networks, purchase habits, mobile footprints, and more
  • 5.
    Key Terms inAI Machine Learning. Systems that can learn without being explicitly programmed by detecting patterns and adjusting actions accordingly • Example: Facebook's News Feed content algorithm changes according to the user's personal interactions with other users Deep Learning. A subcategory of machine learning that deals with the use of neural networks to improve speech recognition, image recognition, and natural language processing • Example: Google Now learns a users’ search and travel patterns to proactively promote content / functionality on a consumer’s mobile device
  • 6.
    Key Terms inAI and Applications Back-Propagation. Training a computer to match images with labels or tags that were pre- defined (if enough pictures of cats are uploaded and tagged the system will start to identify new photos of cats) • Example: Facebook suggests people to tag in photos, without you doing it manually Text Mining. Analyzing natural language in documents, e-mail messages and other free- form text. • Example: Google scans incoming and outgoing emails for information to be used for ads Sentiment Analysis. Aims to determine the attitude of a speaker or writer. • Example: Airlines monitor Twitter in real-time to assess customer feedback; AI is used to corral and queue those responses for customer service reps Image Recognition. Technology that strives to identify objects, people, buildings, places, logos, etc. • Pinterest uses to enhance search and power the “Explore Interest” page based on historical pins.
  • 7.
    • Music Xraycan predict if a song will turn into a pop hit and discover artist with similar sounds. – Algorithm successfully predicted success of Norah Jones and Maroon 5 – Has given 5,000 artists opportunities in the music industry – May eventually rule the world of pop music • Iamus- a computer cluster powered by Melomics’ technology- is capable of composing contemporary classical music without human aid in 8 minutes – Has been taught general technics and plausible actions – Starts with a small initial input and then uses a complex algorithm create a full composition – Can adapt to best fit the aesthetic criteria provided Real World AI: Music Industry
  • 8.
    • Medical datarepositories apply AI to electronic records to enhance diagnosis and treatments based on past health history and treatments. • AI is being used to predict which patients were likely to need certain medications in the future. – An Indiana University study showed that using AI for patient care shows outcomes that were 50% better and more cost effective. • AI is also being used to optimize staffing schedules based on patient schedules and historical needs. Real World AI: Healthcare Click here or here to learn more about AI in Healthcare
  • 9.
    AI is RevolutionizingEverything! Click to Explore Journalism Marketing Cinema Culinary SportsConsulting Video Games Insurance
  • 10.
    Media and Publishing:The Robot Journalist – AP uses for corporate earnings stories for business news report in conjunction with data from Zacks Investment Research – Used by Edmunds to generate car descriptions – Yahoo uses to write Fantasy Football recaps – Creates real-estate listings – Internal sales reports for Fortune 500 companies – Forbes uses to publish earning previews Click here or here to learn more about “The Robot Journalist” Click to Learn More
  • 11.
    Media and Publishing:The Robot Analysts AI Analytics systems take big data sets & produce comprehensible information and insights • AI Analytics benefits: – Increase efficiency and productivity of analysis and reports – Greater consistency in logic, structure and narratives – Experts freed from laborious, repetitive data analysis • Zenith currently uses AI for competitive analysis using data from around 400-500 unique ads with 40K-50K airings (including spot TV, local cable, and local radio) to create weekly summary reports for large markets. Click to Learn More
  • 12.
    Media and Publishing:The Robot Media Buyer Click the Robot to download a programmatic 101
  • 13.
    Media and Publishing:The Robot Shrink Emotional tone, voice intonations and facial expressions can be extremely telling of a consumer attitude towards a brand, event or person • Utilize deep learning/ image recognition software • Uses for sentiment analysis – Customer Service • Many brands use automated bots on Twitter, sometimes with unintentionally awkward results. Click here for more info – Market Analysis and Trading Desks • Monitor social buzz and important news that could move stock by analyzing sentiment and emotion in human language. Click here or here to learn more.
  • 14.
    Media and Publishing:The Robot Autuer Reasons For Image Recognition – Facilitate Mobile Commerce – Customer Engagement – Form insights on consumer attitudes/behaviors based on logo detection Companies Using Image Recognition Technology: – Pinterest: enhancing search and develop relevant ad product – Google: “Reverse Image Search” – Amazon: Firefly technology – Target: Scan magazine ads to add to online shopping cart Click here or on brand names for more info
  • 15.
    Media and Publishing:The Robot Chef Chef Watson – Discovers and invents new dishes based on what it’s learned from recipes plus an understanding of what chemical flavors go together – Input fundamental ingredients, the kind of dish desired and a style to create a full list of ingredients and suggested preparation instructions Click here, here or here to learn more about “The Robot Chef”
  • 16.
    • In termsof direct, applicable use TODAY for AI in marketing and advertising, the most valuable applications can be found in programmatic, data-driven digital campaigns • Contact your buyer for more information on this topic • On the horizon, image recognition AI is extremely promising for advertising on visual social platforms such as Instagram, Pinterest, and even Snapchat • AI can identify logos, retail outlets, products, etc. in photo and video content, produce structured data around them, and evolve to the point where media buyers will be able to buy the “visual graph” in addition to the social graph, the search graph, etc.. • Long term, AI shows promise in dynamic delivery based on mood, sentiment, and intent on traditional platforms such as video through addressable advertising. Implications for Marketing And Advertising
  • 17.
    • AI isextremely versatile in its use and applications for business beyond marketing, and extends to streamlining business processes, customer service, CRM, publishing, and marketing automation • Companies who successfully are able to harness AI in being able to process big data, will have a huge advantage over their competitors as it evolves AI Is Not Just About Advertising, It’s About Your Business
  • 18.
    Hot Companies InAI: Click To Explore
  • 19.
  • 20.
    Sources Used• http://blogs.datadirect.com/2014/04/2014-data-connectivity-outlook.html •http://www.fastcolabs.com/3026423/why-google-is-investing-in-deep-learning • http://realcomm.com/advisory/advisoryPrint.asp?AdvisoryID=544 • http://whatis.techtarget.com/definition/machine-learning • http://wikibon.org/blog/big-data-statistics/ • http://www.emc.com/about/news/press/2011/20110628-01.htm • http://www.huffingtonpost.com/2013/08/26/image-recognition-apps_n_3815554.html • http://www.theguardian.com/technology/2014/apr/15/gmail-scans-all-emails-new-google-terms-clarify • http://io9.com/5973551/this-classical-music-was-created-by-a-supercomputer-in-less-than-a-second • http://techland.time.com/2013/01/04/finally-a-computer-that-writes-contemporary-music-without-human-help/ • http://www.wfs.org/futurist/2013-issues-futurist/may-june-2013-vol-47-no-3/pop-goes-algorithm • http://www.smh.com.au/it-pro/cloud/social-media-analytics-predicting-box-office-success-20140228-hvfg6.html • http://www.marketplace.org/topics/business/whats-behind-future-hit-movies-algorithm • http://www.ox.ac.uk/media/news_stories/2013/130822_1.html • https://www.youtube.com/watch?v=1hkMhng4q74#t=132 • http://mashable.com/2014/04/23/data-online-every-minute/ • http://www.internetlivestats.com/google-search-statistics/ • http://nymag.com/daily/intelligencer/2014/07/why-robot-journalism-is-great-for-journalists.html • http://www.slate.com/articles/technology/technology/2014/07/automated_insights_to_write_ap_earnings_reports_why_robots_can_t_take_journalists.html • http://www.fastcodesign.com/3032501/ibms-watson-is-now-a-cooking-app-with-infinite-recipes#1 • http://www.slate.com/articles/technology/future_tense/2014/06/chef_watson_with_bon_appetit_ibm_s_supercomputer_enters_the_culinary_world.html • http://www.bonappetit.com/test-kitchen/inside-our-kitchen/article/chef-watson-in-the-ba-test-kitchen • http://www.npr.org/blogs/health/2013/05/20/182593855/if-your-shrink-is-a-bot-how-do-you-respond • http://www.nbcnews.com/business/consumer/missingthepoint-companies-alienate-customers-twitter-bots-scripted-responses-f6C10760870 • http://techcrunch.com/2014/02/03/twitter-raises-its-enterprise-cred-with-thomson-reuters-sentiment-analysis-deal/ • http://www.mobilemarketer.com/cms/news/software-technology/18380.html • ARRIA. “Articulate Intelligence” http://www.arria.com/ • BI Intelligence. “Social Media’s New Big Data Frontier- Artificial Intelligence, Deep Learning, and Predictive Marketing. February 2014