Media & Entertainment
Case Studies
Newsfor Media & Entertainment
20th Century Fox and IBM Research reduce movie marketing production time by using machine learning to select scenes for a movie trailer
20th Century Fox implemented IBM Research to select scenes for inclusion in a movie trailer for a feature-length horror film. This process can take a month manually, but was reduced to 24 hours using machine learning from start to finish.
Linked use case: Automate content generation for video and image based marketing materials
Allen Institute for AI develops drawing-guessing game to study people-software collaboration using machine learning
Allen Institute for AI has developed Iconary, a game that resembles Pictionary where players have to guess a word based on their teammate''s drawing. In this version, a user plays against a bot called AllenAI that was developed by applying machine learning algorithms to records of human played games, drawings and phrases. Whenever a user plays, they contribute to the research project of making computers and humans better coworkers and collaborators.
Artist Ben Snell models sculpture using machine learning
Artist Ben Snell used artificial intelligence to create a sculpture. Machine learning algorithms were fed with a large database of historical artworks, more than 1,000 classical sculptures, and programmed to generate a new piece following the same methodology. The new artwork, called Dio, is currently up for sale at London auction house Phillips.
Linked use case: Create visual art
BBC is developing interactive fictional programs utilising the natural language generation capabilities of smart home assistants such as Alexa and Google Home
BBC plans to develop a natural language processing engine to offer interactive programmes. The initiative is titled ''Talking with Machines'' and will explore content, interaction and software development patterns. The first programme in this series, an interactive audio drama has been already broadcast in November 2017. The programme is now being expanded.
Linked use case: Automate generation of conversational media content
Blackburn Rovers Football Club increasing productivity by identifying staff who had excessive personal computer use
The Blackburn Rovers, a professional soccer club in the UK, started monitoring computer and internet usage of staff using Veriato''s machine learning solution. In addition to this, employee communication is also monitored. The software also helped them stay compliant with the rules and regulation of the worldwide Payment Card Industry (PCI) Data Security Standard.
Linked use case: Monitor and measure employee digital media access to support productivity improvement
Bodybuilding.com improves app engagement through personalized narratives about user’s weekly workouts using NLG and machine learning.
Bodybuilding.com uses natural language generation to generate narratives about weekly workouts and connect with users better. The platform generates more than 100,000 workout recaps weekly from structured data collected from Bodybuilding.com''s app.
Linked use case: Personalise sales pitch, content and proposition based on customer analytics
Brooklyn Dynamics researching an athlete scouting application based on machine learning of historic data
Brooklyn Dynamics has been building data sets of individual performance - their "data CV" - to support player measurement. This builds on the approach to discovering undervalued players popularised in ''Moneyball''.
Linked use case: Track and analyse individual sports players to predict likelihood of success
BuzzFeed attracts more suitable job candidates through recruitment response automation using uses IBM Watson and Uncubed
BuzzFeed is working with recruitment platform vendor Uncubed to provide video answers to questions candidates have regarding the company and roles during the application process, automating the responses via chatbot and allowing candidates to be better self-selecting.
Linked use case: Automate responses to candidate questions during recruitment and enhance candidate engagement
Bytedance ranks web content according to the preference of each user by analysing their behaviour using machine learning
Bytedance offers Toutiao - an app which uses machine learning to discover content from the web for easy consumption. The algorithm presents content ranked according to the preference of each user uncovered by analysing clicks, views, comments etc on the app. The app is extremely popular in China with more than 120m users who spent ~76 minutes on the app daily.
Linked use case: Personalise media content recommendations
Caesar Entertainment drives revenue by optimising floor layouts, pricing and even restaurant menus to better match customer segments
Caesars Entertainment tracks customer behaviour through loyalty programs and records their entire journey in the casino through transaction data, win/loss, videos and emotion recognition. They use this data to better align pricing decisions, floor arrangements, menu with customer segments.
Linked use case: Personalise loyalty programs and promotional offerings to individual customers
Caesars Entertainment improves email open rates by 12% and click-throughs by 24% by personalising content using deep learning
Caesars Entertainment utilised the Persado platform to send emails to customers with personalised emotional language to improve the open and click-through rates by 12% and 24% respectively.
Linked use case: Create tailored emails
Caesars Entertainment improves promotional targeting having increased customer individual spend tracking from 58 to 85 percent with loyalty programs and machine learning
Caesars Entertainment tracks customer behaviour through loyalty programs and records their entire journey in the casino from transactions such as shopping, dining, floor spending. Using this data they segment customers and offer customised rewards such as reduced room rates, airline rates, coupons etc.
Linked use case: Personalise loyalty programs and promotional offerings to individual customers
Caesars Entertainment offers customers on the casino floor real-time incentives to bet more by identifying their behaviour and emotions through machine vision
Caesars Entertainment monitors customer sentiments through win/loss at stations as well as emotion recognition and behavioural analysis from visual feeds. This data is analysed using machine learning to quickly identify offers for customers to encourage them to spend more.
Linked use case: Personalise loyalty programs and promotional offerings to individual customers
Clarivate Analytics aims to improve article peer review efficiency by adding natural language processing to its ScholarOne platform
Clarivate Analytics'' has teamed up with UNSILO to add natural language processing abilities to its journal peer review software, ScholarOne. This allows papers submitted for review to be automatically summarised and checked against other papers for plagiarism.
Linked use case: Detect plagiarism in documents
Columbia Business School researchers determine that unusual language in news stories can forecast future market stress through natural language processing
Columbia Business School researchers applied natural language processing techniques to news stories about corporations and assess the unusualness of the language to determine if investors utilise this information and the market adjust accordingly. They found a positive correlation but time lag between unusualness and story sentiment and market stress.
Linked use case: Predict asset price movements based on greater quantities of data to inform trading strategies
Condé Nast improves average response time to issues during new content and capability rollout to their websites by 70% using machine learning
Condé Nast ensures consumers of digital and video content always get a smooth experience using AI. Rollouts of new content or features are constantly monitored and application performance is analysed using machine learning to detect anomalies or issues enabling them to react promptly.
Linked use case: Monitor application performance for resource management and measure KPIs and to identify issues before they become widepsread
Dadabots creates lo-fi black metal album using deep neural networks
CJ Carr and and Zack Zukowski built Dadabots, a system capable of producing actual waveforms and raw sound, such as even wailing vocals. Using a modified SampleRNN it generates music in modern genres based. It created ''Coditany of Timeness'' by breaking down Krallice''s 2011 album ''Diotima'' into small segments of audio and imitating its style. Dadabods has released three programmatic albums using this technique.
Deep Stack''s poker AI delivered statistically significant wins in Texas Hold ''Em poker games through deep learning.
Deep Stack is one of several AI tools set up to play poker - this is a challenge because of the lack of full information that might be available in a game like chess. Hence the bluff element. Deep Stack was able to deliver a statistically significant levels of wins against professional poker players.
Linked use case: Play limited information games like poker to championship levels
Dictionary.com automates software testing before release, saving days worth of effort using machine learning
Dictionary.com uses Test.ai machine learning based platform to test apps, pages and features before release of upgrades or maintenance. Test.ai''s software keeps a log of baseline versions and compares upgrades quickly to identify anomalies. This has helped Dictionary.com reduce testing time from days to minutes.
Linked use case: Automate testing of web applications and apps as part of continuous development
Disney researchers identify, track and predict movie audience enjoyment using facial recognition
Disney and affiliated university researchers introduce and test a method for detecting audience facial emotion during film watching. The model was able to identify emotions as well as predict the future emotional state of audience members after 10 minutes of observation. The potential benefit is for improving marketing research.
Linked use case: Measure biometric and neurological response in product pre-testing
Elections Canada plans to use AI to identify electoral misinformation campaigns on social media
Elections Canada plans to use AI to monitor conversations on social media with regards to elections. The intent is to identify misinformation campaigns before they spread as a safeguard.
Linked use case: Censor user generated content on social media and other platforms
Electronic Arts is developing an AI agent which can test for quality assurance during video game production by learning to play against humans
EA is developing AI to test video games for quality assurance during game development. The current proof-of-concept research has resulted in an AI that can play a first-person shooter video game against humans with a measure of success.
Linked use case: Automate software coding, testing and production deployment
Evernote doubles click-to-apply on Glassdoor and attracts ten to forty percent more candidates from underrepresented groups by optimising job descriptions using machine learning
With the help of Textio, Evernote started attracting more candidates from diverse backgrounds. Textio edits job descriptions and messages to a more neutral language. Evernote received 40% more applications from female candidates and 10% more from other underepresented groups.
Linked use case: Optimise job posting descriptions to attract better and more diverse candidates
Facebook is trying to combat misinformation in the site’s news feed with machine learning
Facebook is trying to reduce misinformation in the site’s news feed with the use of machine learning. The company''s AI has been trained to evaluate the source of an article, along with other signals like negative comments, and sends the suspicious link to human fact-checkers. If they then rate it as false they reduce future views of these links by 80%.
Linked use case: Automate flagging and/or removal of user generated and uploaded media content
First Book increases repeat sales by 331% and revenue by 35% by targeting customers who are more likely to do repeat business using machine learning
First Book, a non-profit firm which sells books at a discounted rate to low-income families, turned to machine learning to better segment and understand publishers and buyers. They could then identify organisations which are more likely to do repeat business using machine learning and target them effectively achieving 331% increase in repeat sales success rate.
Linked use case: Identify most promising potential customers for campaign to target