SlideShare a Scribd company logo
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Business Process Automation
Using Crowdsourcing
Mark Chien
GM of Requester
Amazon Mechanical Turk
A I M 3 5 2
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Agenda
Introduction to MTurk for business process
automation
Business process automation customers on MTurk
Guest speakers
Using MTurk for business process automation
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
What is crowdsourcing?
Harnessing the collective power of many individuals
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
What is Amazon Mechanical Turk?
Amazon Mechanical Turk (MTurk) operates an online crowdsourcing
marketplace for work that requires human intelligence
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Microwork: Typically small, repeatable tasks that give workers
variety and flexibility
Each item
can be sent
to a worker
Or multiple
workers to increase
accuracy
Or through
multiple steps with
different workers
What kind of work is on MTurk?
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Amazon Mechanical Turk customers
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
What is business process automation?
“Business process automation (BPA) is the technology-
enabled automation of complex business processes”
- Wikipedia
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Choosing the right business processes with MTurk
• Processes that can be broken down into smaller, simpler tasks
• Tasks require minimal context and can handle high turnover
• Handle non-confidential/critical data
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Common business process automation use cases on
MTurk
• Data collection/gathering
• Content moderation
• Image/Text/Audio/Video transcription
• Categorization
• Data validation
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Advantages of MTurk for business process automation
Cost savings
• Avoid the overhead and fixed costs associated with hiring and managing a
temp workforce
• Free up time and resources from your team
Elastic scale
• Use as much or as little as you need based on your demand
• There are no contracts or minimum required spend
• 24/7 global workforce to complete your work whenever needed
Diversity
• Access to 500,000 workers across 190 countries, across various demographics
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Finding opportunities to use MTurk
• Large teams of temporary or contract workers
• Processes that have already been outsourced
• Teams doing repetitive, low-context tasks
• Teams challenged with handling large deviations in volume of work
• Need for 24x7 manual processing
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Content moderation and classification
Company: Online travel marketplace
Problem: Need help to moderate and categorize hotel
images uploaded by users
Approach: User-generated images are sent to MTurk in
real-time for review
MTurk task: Ensure images meet guidelines and
categorize them
Result: Effectively moderate and categorize these
photos before they are posted publicly, so that the most
relevant photos are surfaced on the hotel listings
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Validating sales accounts
Company: Online job search portal
Problem: Keeping information about customers
up to date in Salesforce
Approach: Extract customer websites from
Salesforce records and submit to MTurk
MTurk task: Confirm that the address is still valid,
find out if the company has been acquired, collect
updated data points about the company
Result: Accurate sales data and more-efficient sales
teams
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Validating catalog information
Company: Consumer travel information service
Problem: Need to confirm that recommended
restaurants are still open
Approach: Have MTurk workers visit restaurant
websites to confirm hours of operation
MTurk task: Visit the website and confirm that
it is still open
Result: Users get most up-to-date information
about restaurants on their travels
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Assessing damage after disasters
Company: Disaster assessment
Problem: Need to get building damage
assessments completed quickly to take action
Approach: Have MTurk workers review photos of
the damage
MTurk task: Assess and categorize the damage
Result: Faster response times and more-accurate
decisions
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Food Genius / US Foods Data Science
David Falck
Executive Director
dave@getfoodgenius.com
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
US Foods + Food Genius
• Chicago-based start-up, acquired by US Foods
in 2016 (2nd largest food distribution
company in the US)
• Support sellers’ conversations with customers
in the field with data
• Provide valuable, actionable
recommendations based on accurate and
fresh restaurant data
• Goal: Make positive impact in our sellers’ day-
to-day routines (and their wallets)
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Background
• Small, independent team that needed to scale quickly
• Familiarity with crowdsourcing – including code development and data
science projects
• Experimented with MTurk for research
• Early wins spurred continued investment in MTurk
• Shared case studies and knowledge with rest of company
• Expanding MTurk use cases to other US Foods business units
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Why Mechanical Turk?
• We use Mechanical Turk to find up-to-date, real data to help our customers
• Once we have enough external data, coupling it with our internal data makes it MUCH
more valuable…
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Mechanical Turk workflows
Example 1: Find the “real” website and menus of a restaurant
(not delivery, Facebook menus)
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Mechanical Turk workflows
Example 2: Multi-step information extraction from images
OCR
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Initial process
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Key Learning #1
Feedback from Workers:
• We didn’t take into consideration fair payment
• We didn’t make our HITs’ instructions clear enough
Solution:
• Remember, even though you can access their services via APIs,
Workers are real people!
• Make the instructions super simple, even fun, if you can.
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Key Learning #2
Feedback from Workers:
• Manual verification took longer than expected, QA of 4,500 URLs took 6 days
• Coverage was low too – only about 30%
Solution:
• Build a cadre of Qualified Workers and use them!
• Use Workers to help with QA work
• Provide examples, better list of prohibited domains, empty answers will be
rejected
• QA the work closely and reject work that doesn’t meet our criteria
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Results
• Our manual work went from 100% to 50% to 25% as we
made refinements to our process.
• Coverage went from 30%-63% for homepage URLs
• Workers are happier – many reach out and inquire when
we’re doing a batch of work next
• Modifying the original process lets us pivot to other asks
quickly and efficiently, such as collecting reviews, social
media data, etc. about restaurants
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Summary
Make crowdsourcing work for your business – You can make the
process efficient, cost-effective and scalable by putting yourself in
the shoes of a Worker and continuously improving your processes for
them
Remember:
• Listen to the Workers
• Develop effective tasks
• Measure and enhance your processes
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Indeed
Andy Hilal
Product Manager
Indeed Job Spotter
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Agenda
Job Spotter: Getting local jobs onto Indeed
Using MTurk for content moderation
Results and impact
Making your MTurk program work
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
OCR and photo analysis
“CioSO
eli
cioso Super Rico Pollo
PERUVIAN STYLE
CHARCOAL
CHICKEN
GRAND
ESPECIAL Pollo entero
PENING$99
con
SOPA DE
THUR
CARNE ASADA
PUPUSAS
LOMO SALTADO
TACOS
POLLO FRITO
N
CHARCOAL CHICKEN”
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Machine learning (with humans teaching)
What is the business being shown?
Is it a hiring sign?
• Indeed keeps high standards for job content
• No business opportunities
• No training programs
• No volunteer positions
• No elected offices
• No sexist language
• … and many more
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
User-generated content risks
Low quality
• Low lighting, blurry
Objectionable content
• Faces of passersby on the street
• Worse
Fraud
• Photos of newspaper want ads
• Falsified or duplicated job sign submissions
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
From local vendor to MTurk
Nearshore Vendor Amazon MTurk
Speed of Completion 2841 minutes 34 minutes
Quality 9.73% variance 9.93% variance
Cost 50 cents / task 7 cents / task
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Our biggest wins
Assume goodwill
Build clear, fast tools
Provide contextual help content
Low HIT expiry time
Golden questions and clear warnings
Price HITs correctly
Qualify a closed set of Workers
Talk to your Workers :)
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
MTurk concepts primer
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Steps for using MTurk for business process automation
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Define workflow
• Break complex tasks down to simple, discrete tasks
• Determine how you want the task flow to work
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Design tasks
Use HTML/JS/CSS to design your task
Provide clear instructions and assume no context
Complete the task yourself:
• To validate instructions
• Determine average completion time to set appropriate reward
• To confirm that the task UI works
Use the sandbox to test
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Set workforce
Use Qualifications to determine who gets to do your tasks
MTurk system Qualifications
• Experience-based (# of approved HITs, approval rate)
• Locale (US, states)
• Demographics (adult, gender, degree)
Define custom Qualifications
• Inclusion Qualifications (such as whitelist)
• Exclusion Qualifications (such as blacklist)
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Load tasks
Use the API to load dataset programmaticallyUse the UI to load dataset via CSV
Or
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Review results
Use the API to download results and SNS
notifications when tasks are completed
Use the UI to view progress and
download results via CSV
Or
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Refine process
• Analyze results to assess for quality
• Review feedback provided by workers
• Update instructions and/or task design
• Update qualification membership or requirements
• Update reward for desired quality and throughput
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Get started at www.mturk.com
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Related Sessions
Wednesday, Nov 28
Session | AIM351 – Harness the Power of Crowdsourcing with Amazon
Mechanical Turk
12:15 – 1:15 p.m. | Aria East, Plaza Level, Orovada 2
Wednesday, Nov 28
Workshop | AIM362 – Crowdsourcing Data Collection with
Amazon Mechanical Turk
3:15 – 5:30 p.m. | Venetian, Level 4, Lando 4205
Wednesday, Nov 28
Chalk Talk | AIM358 - Human-in-the-Loop for Machine Learning
7 – 8 p.m. | Aria West, Level 3, Starvine 7
Thank you!
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Mark Chien
mchien@amazon.com
© 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.

More Related Content

What's hot

[NEW LAUNCH!] Introducing Amazon Personalize: Real-time Personalization and R...
[NEW LAUNCH!] Introducing Amazon Personalize: Real-time Personalization and R...[NEW LAUNCH!] Introducing Amazon Personalize: Real-time Personalization and R...
[NEW LAUNCH!] Introducing Amazon Personalize: Real-time Personalization and R...
Amazon Web Services
 
Machine Learning at the Edge (AIM302) - AWS re:Invent 2018
Machine Learning at the Edge (AIM302) - AWS re:Invent 2018Machine Learning at the Edge (AIM302) - AWS re:Invent 2018
Machine Learning at the Edge (AIM302) - AWS re:Invent 2018
Amazon Web Services
 
Unlock the Full Potential of Your Media Assets, ft. Fox Entertainment Group (...
Unlock the Full Potential of Your Media Assets, ft. Fox Entertainment Group (...Unlock the Full Potential of Your Media Assets, ft. Fox Entertainment Group (...
Unlock the Full Potential of Your Media Assets, ft. Fox Entertainment Group (...
Amazon Web Services
 
Debugging Gluon and Apache MXNet (AIM423) - AWS re:Invent 2018
Debugging Gluon and Apache MXNet (AIM423) - AWS re:Invent 2018Debugging Gluon and Apache MXNet (AIM423) - AWS re:Invent 2018
Debugging Gluon and Apache MXNet (AIM423) - AWS re:Invent 2018
Amazon Web Services
 
Amazon, awsreinvent2018, Artificial Intelligence & Machine Learning, AIM422, ...
Amazon, awsreinvent2018, Artificial Intelligence & Machine Learning, AIM422, ...Amazon, awsreinvent2018, Artificial Intelligence & Machine Learning, AIM422, ...
Amazon, awsreinvent2018, Artificial Intelligence & Machine Learning, AIM422, ...
Amazon Web Services
 
Transform the Modern Contact Center Using Machine Learning and Analytics (AIM...
Transform the Modern Contact Center Using Machine Learning and Analytics (AIM...Transform the Modern Contact Center Using Machine Learning and Analytics (AIM...
Transform the Modern Contact Center Using Machine Learning and Analytics (AIM...
Amazon Web Services
 
Serverless Stream Processing Tips & Tricks (ANT358) - AWS re:Invent 2018
Serverless Stream Processing Tips & Tricks (ANT358) - AWS re:Invent 2018Serverless Stream Processing Tips & Tricks (ANT358) - AWS re:Invent 2018
Serverless Stream Processing Tips & Tricks (ANT358) - AWS re:Invent 2018
Amazon Web Services
 
We Power Tech: Addressing Intersectionality in Tech (WPT203-S) - AWS re:Inven...
We Power Tech: Addressing Intersectionality in Tech (WPT203-S) - AWS re:Inven...We Power Tech: Addressing Intersectionality in Tech (WPT203-S) - AWS re:Inven...
We Power Tech: Addressing Intersectionality in Tech (WPT203-S) - AWS re:Inven...
Amazon Web Services
 
Releasing Mission-Critical Software at Amazon (DEV209-R1) - AWS re:Invent 2018
Releasing Mission-Critical Software at Amazon (DEV209-R1) - AWS re:Invent 2018Releasing Mission-Critical Software at Amazon (DEV209-R1) - AWS re:Invent 2018
Releasing Mission-Critical Software at Amazon (DEV209-R1) - AWS re:Invent 2018
Amazon Web Services
 
Build an Intelligent Multi-Modal User Agent with Voice and NLU (AIM340) - AWS...
Build an Intelligent Multi-Modal User Agent with Voice and NLU (AIM340) - AWS...Build an Intelligent Multi-Modal User Agent with Voice and NLU (AIM340) - AWS...
Build an Intelligent Multi-Modal User Agent with Voice and NLU (AIM340) - AWS...
Amazon Web Services
 
Building, Training, and Deploying fast.ai Models Using Amazon SageMaker (AIM4...
Building, Training, and Deploying fast.ai Models Using Amazon SageMaker (AIM4...Building, Training, and Deploying fast.ai Models Using Amazon SageMaker (AIM4...
Building, Training, and Deploying fast.ai Models Using Amazon SageMaker (AIM4...
Amazon Web Services
 
Build AWS Skills Through Community-Led User Groups (DVC202) - AWS reInvent 20...
Build AWS Skills Through Community-Led User Groups (DVC202) - AWS reInvent 20...Build AWS Skills Through Community-Led User Groups (DVC202) - AWS reInvent 20...
Build AWS Skills Through Community-Led User Groups (DVC202) - AWS reInvent 20...
Amazon Web Services
 
Human-in-the-Loop for Machine Learning (AIM358-R1) - AWS re:Invent 2018
Human-in-the-Loop for Machine Learning (AIM358-R1) - AWS re:Invent 2018Human-in-the-Loop for Machine Learning (AIM358-R1) - AWS re:Invent 2018
Human-in-the-Loop for Machine Learning (AIM358-R1) - AWS re:Invent 2018
Amazon Web Services
 
Market Prediction Using ML: Experiment with Amazon SageMaker and the Deutsche...
Market Prediction Using ML: Experiment with Amazon SageMaker and the Deutsche...Market Prediction Using ML: Experiment with Amazon SageMaker and the Deutsche...
Market Prediction Using ML: Experiment with Amazon SageMaker and the Deutsche...
Amazon Web Services
 
Use Amazon Rekognition to Power Video Creative Asset Production (ADT202) - AW...
Use Amazon Rekognition to Power Video Creative Asset Production (ADT202) - AW...Use Amazon Rekognition to Power Video Creative Asset Production (ADT202) - AW...
Use Amazon Rekognition to Power Video Creative Asset Production (ADT202) - AW...
Amazon Web Services
 
Introducing Amazon SageMaker - AWS Online Tech Talks
Introducing Amazon SageMaker - AWS Online Tech TalksIntroducing Amazon SageMaker - AWS Online Tech Talks
Introducing Amazon SageMaker - AWS Online Tech Talks
Amazon Web Services
 
[NEW LAUNCH!] How to build and deploy Windows file system in AWS using Amazon...
[NEW LAUNCH!] How to build and deploy Windows file system in AWS using Amazon...[NEW LAUNCH!] How to build and deploy Windows file system in AWS using Amazon...
[NEW LAUNCH!] How to build and deploy Windows file system in AWS using Amazon...
Amazon Web Services
 
Create a Serverless Searchable Media Library (AIM342-R1) - AWS re:Invent 2018
Create a Serverless Searchable Media Library (AIM342-R1) - AWS re:Invent 2018Create a Serverless Searchable Media Library (AIM342-R1) - AWS re:Invent 2018
Create a Serverless Searchable Media Library (AIM342-R1) - AWS re:Invent 2018
Amazon Web Services
 
Serverless Architectural Patterns and Best Practices (ARC305-R2) - AWS re:Inv...
Serverless Architectural Patterns and Best Practices (ARC305-R2) - AWS re:Inv...Serverless Architectural Patterns and Best Practices (ARC305-R2) - AWS re:Inv...
Serverless Architectural Patterns and Best Practices (ARC305-R2) - AWS re:Inv...
Amazon Web Services
 
How Fannie Mae Processes over a Quarter Million Loans per Day with Amazon S3 ...
How Fannie Mae Processes over a Quarter Million Loans per Day with Amazon S3 ...How Fannie Mae Processes over a Quarter Million Loans per Day with Amazon S3 ...
How Fannie Mae Processes over a Quarter Million Loans per Day with Amazon S3 ...
Amazon Web Services
 

What's hot (20)

[NEW LAUNCH!] Introducing Amazon Personalize: Real-time Personalization and R...
[NEW LAUNCH!] Introducing Amazon Personalize: Real-time Personalization and R...[NEW LAUNCH!] Introducing Amazon Personalize: Real-time Personalization and R...
[NEW LAUNCH!] Introducing Amazon Personalize: Real-time Personalization and R...
 
Machine Learning at the Edge (AIM302) - AWS re:Invent 2018
Machine Learning at the Edge (AIM302) - AWS re:Invent 2018Machine Learning at the Edge (AIM302) - AWS re:Invent 2018
Machine Learning at the Edge (AIM302) - AWS re:Invent 2018
 
Unlock the Full Potential of Your Media Assets, ft. Fox Entertainment Group (...
Unlock the Full Potential of Your Media Assets, ft. Fox Entertainment Group (...Unlock the Full Potential of Your Media Assets, ft. Fox Entertainment Group (...
Unlock the Full Potential of Your Media Assets, ft. Fox Entertainment Group (...
 
Debugging Gluon and Apache MXNet (AIM423) - AWS re:Invent 2018
Debugging Gluon and Apache MXNet (AIM423) - AWS re:Invent 2018Debugging Gluon and Apache MXNet (AIM423) - AWS re:Invent 2018
Debugging Gluon and Apache MXNet (AIM423) - AWS re:Invent 2018
 
Amazon, awsreinvent2018, Artificial Intelligence & Machine Learning, AIM422, ...
Amazon, awsreinvent2018, Artificial Intelligence & Machine Learning, AIM422, ...Amazon, awsreinvent2018, Artificial Intelligence & Machine Learning, AIM422, ...
Amazon, awsreinvent2018, Artificial Intelligence & Machine Learning, AIM422, ...
 
Transform the Modern Contact Center Using Machine Learning and Analytics (AIM...
Transform the Modern Contact Center Using Machine Learning and Analytics (AIM...Transform the Modern Contact Center Using Machine Learning and Analytics (AIM...
Transform the Modern Contact Center Using Machine Learning and Analytics (AIM...
 
Serverless Stream Processing Tips & Tricks (ANT358) - AWS re:Invent 2018
Serverless Stream Processing Tips & Tricks (ANT358) - AWS re:Invent 2018Serverless Stream Processing Tips & Tricks (ANT358) - AWS re:Invent 2018
Serverless Stream Processing Tips & Tricks (ANT358) - AWS re:Invent 2018
 
We Power Tech: Addressing Intersectionality in Tech (WPT203-S) - AWS re:Inven...
We Power Tech: Addressing Intersectionality in Tech (WPT203-S) - AWS re:Inven...We Power Tech: Addressing Intersectionality in Tech (WPT203-S) - AWS re:Inven...
We Power Tech: Addressing Intersectionality in Tech (WPT203-S) - AWS re:Inven...
 
Releasing Mission-Critical Software at Amazon (DEV209-R1) - AWS re:Invent 2018
Releasing Mission-Critical Software at Amazon (DEV209-R1) - AWS re:Invent 2018Releasing Mission-Critical Software at Amazon (DEV209-R1) - AWS re:Invent 2018
Releasing Mission-Critical Software at Amazon (DEV209-R1) - AWS re:Invent 2018
 
Build an Intelligent Multi-Modal User Agent with Voice and NLU (AIM340) - AWS...
Build an Intelligent Multi-Modal User Agent with Voice and NLU (AIM340) - AWS...Build an Intelligent Multi-Modal User Agent with Voice and NLU (AIM340) - AWS...
Build an Intelligent Multi-Modal User Agent with Voice and NLU (AIM340) - AWS...
 
Building, Training, and Deploying fast.ai Models Using Amazon SageMaker (AIM4...
Building, Training, and Deploying fast.ai Models Using Amazon SageMaker (AIM4...Building, Training, and Deploying fast.ai Models Using Amazon SageMaker (AIM4...
Building, Training, and Deploying fast.ai Models Using Amazon SageMaker (AIM4...
 
Build AWS Skills Through Community-Led User Groups (DVC202) - AWS reInvent 20...
Build AWS Skills Through Community-Led User Groups (DVC202) - AWS reInvent 20...Build AWS Skills Through Community-Led User Groups (DVC202) - AWS reInvent 20...
Build AWS Skills Through Community-Led User Groups (DVC202) - AWS reInvent 20...
 
Human-in-the-Loop for Machine Learning (AIM358-R1) - AWS re:Invent 2018
Human-in-the-Loop for Machine Learning (AIM358-R1) - AWS re:Invent 2018Human-in-the-Loop for Machine Learning (AIM358-R1) - AWS re:Invent 2018
Human-in-the-Loop for Machine Learning (AIM358-R1) - AWS re:Invent 2018
 
Market Prediction Using ML: Experiment with Amazon SageMaker and the Deutsche...
Market Prediction Using ML: Experiment with Amazon SageMaker and the Deutsche...Market Prediction Using ML: Experiment with Amazon SageMaker and the Deutsche...
Market Prediction Using ML: Experiment with Amazon SageMaker and the Deutsche...
 
Use Amazon Rekognition to Power Video Creative Asset Production (ADT202) - AW...
Use Amazon Rekognition to Power Video Creative Asset Production (ADT202) - AW...Use Amazon Rekognition to Power Video Creative Asset Production (ADT202) - AW...
Use Amazon Rekognition to Power Video Creative Asset Production (ADT202) - AW...
 
Introducing Amazon SageMaker - AWS Online Tech Talks
Introducing Amazon SageMaker - AWS Online Tech TalksIntroducing Amazon SageMaker - AWS Online Tech Talks
Introducing Amazon SageMaker - AWS Online Tech Talks
 
[NEW LAUNCH!] How to build and deploy Windows file system in AWS using Amazon...
[NEW LAUNCH!] How to build and deploy Windows file system in AWS using Amazon...[NEW LAUNCH!] How to build and deploy Windows file system in AWS using Amazon...
[NEW LAUNCH!] How to build and deploy Windows file system in AWS using Amazon...
 
Create a Serverless Searchable Media Library (AIM342-R1) - AWS re:Invent 2018
Create a Serverless Searchable Media Library (AIM342-R1) - AWS re:Invent 2018Create a Serverless Searchable Media Library (AIM342-R1) - AWS re:Invent 2018
Create a Serverless Searchable Media Library (AIM342-R1) - AWS re:Invent 2018
 
Serverless Architectural Patterns and Best Practices (ARC305-R2) - AWS re:Inv...
Serverless Architectural Patterns and Best Practices (ARC305-R2) - AWS re:Inv...Serverless Architectural Patterns and Best Practices (ARC305-R2) - AWS re:Inv...
Serverless Architectural Patterns and Best Practices (ARC305-R2) - AWS re:Inv...
 
How Fannie Mae Processes over a Quarter Million Loans per Day with Amazon S3 ...
How Fannie Mae Processes over a Quarter Million Loans per Day with Amazon S3 ...How Fannie Mae Processes over a Quarter Million Loans per Day with Amazon S3 ...
How Fannie Mae Processes over a Quarter Million Loans per Day with Amazon S3 ...
 

Similar to Business Process Automation Using Crowdsourcing (AIM352) - AWS re:Invent 2018

How Trupanion Became an AI-driven Company for Pets
How Trupanion Became an AI-driven Company for PetsHow Trupanion Became an AI-driven Company for Pets
How Trupanion Became an AI-driven Company for Pets
Amazon Web Services
 
An Agile Approach to Cloud Adoption
An Agile Approach to Cloud AdoptionAn Agile Approach to Cloud Adoption
An Agile Approach to Cloud Adoption
Amazon Web Services
 
Practical Human-in-the-Loop Machine Learning
 Practical Human-in-the-Loop Machine Learning Practical Human-in-the-Loop Machine Learning
Practical Human-in-the-Loop Machine Learning
Amazon Web Services
 
Using Amazon Mechanical Turk to Crowdsource Data Collection (AIM359) - AWS re...
Using Amazon Mechanical Turk to Crowdsource Data Collection (AIM359) - AWS re...Using Amazon Mechanical Turk to Crowdsource Data Collection (AIM359) - AWS re...
Using Amazon Mechanical Turk to Crowdsource Data Collection (AIM359) - AWS re...
Amazon Web Services
 
Innovation at Amazon
Innovation at AmazonInnovation at Amazon
Innovation at Amazon
Amazon Web Services
 
AWS STARTUP DAY 2018 I Innovation @ Amazon
AWS STARTUP DAY 2018 I Innovation @ AmazonAWS STARTUP DAY 2018 I Innovation @ Amazon
AWS STARTUP DAY 2018 I Innovation @ Amazon
AWS Germany
 
AWS Think Big Workshop: Experimenting with Data
AWS Think Big Workshop: Experimenting with DataAWS Think Big Workshop: Experimenting with Data
AWS Think Big Workshop: Experimenting with Data
Amazon Web Services
 
Success Story of migrating entire infrastructure from AWS Singapore to AWS Mu...
Success Story of migrating entire infrastructure from AWS Singapore to AWS Mu...Success Story of migrating entire infrastructure from AWS Singapore to AWS Mu...
Success Story of migrating entire infrastructure from AWS Singapore to AWS Mu...
Pranesh Vittal
 
Success story of migrating entire infrastructure from AWS Singapore to AWS Mu...
Success story of migrating entire infrastructure from AWS Singapore to AWS Mu...Success story of migrating entire infrastructure from AWS Singapore to AWS Mu...
Success story of migrating entire infrastructure from AWS Singapore to AWS Mu...
AWS User Group Bengaluru
 
Leadership Session: Overview of Amazon Digital User Engagement Solutions (DIG...
Leadership Session: Overview of Amazon Digital User Engagement Solutions (DIG...Leadership Session: Overview of Amazon Digital User Engagement Solutions (DIG...
Leadership Session: Overview of Amazon Digital User Engagement Solutions (DIG...
Amazon Web Services
 
Startup Day Kyiv: How we think about Innovation at Amazon
Startup Day Kyiv: How we think about Innovation at Amazon Startup Day Kyiv: How we think about Innovation at Amazon
Startup Day Kyiv: How we think about Innovation at Amazon
Amazon Web Services
 
Tendências na Transformação Digital
Tendências na Transformação DigitalTendências na Transformação Digital
Tendências na Transformação Digital
Amazon Web Services LATAM
 
AWS Initiate - Tendências da Transformação Digital
AWS Initiate - Tendências da Transformação DigitalAWS Initiate - Tendências da Transformação Digital
AWS Initiate - Tendências da Transformação Digital
Amazon Web Services LATAM
 
Continuously Delivering Your Software on AWS - Adrian White - AWS TechShift A...
Continuously Delivering Your Software on AWS - Adrian White - AWS TechShift A...Continuously Delivering Your Software on AWS - Adrian White - AWS TechShift A...
Continuously Delivering Your Software on AWS - Adrian White - AWS TechShift A...
Amazon Web Services
 
[REPEAT] Better Analytics Through Natural Language Processing (AIM405-R) - AW...
[REPEAT] Better Analytics Through Natural Language Processing (AIM405-R) - AW...[REPEAT] Better Analytics Through Natural Language Processing (AIM405-R) - AW...
[REPEAT] Better Analytics Through Natural Language Processing (AIM405-R) - AW...
Amazon Web Services
 
The Future of API Management Is Serverless
The Future of API Management Is ServerlessThe Future of API Management Is Serverless
The Future of API Management Is Serverless
Chris Munns
 
Enterprise Cloud Adoption
Enterprise Cloud Adoption Enterprise Cloud Adoption
Enterprise Cloud Adoption
Tom Laszewski
 
How to Wrangle Data for Machine Learning on AWS
 How to Wrangle Data for Machine Learning on AWS How to Wrangle Data for Machine Learning on AWS
How to Wrangle Data for Machine Learning on AWS
Amazon Web Services
 
Industrial Transformation
Industrial TransformationIndustrial Transformation
Industrial Transformation
Amazon Web Services
 
Digital Transformation Playbook in Five Steps (ARC322) - AWS re:Invent 2018
Digital Transformation Playbook in Five Steps (ARC322) - AWS re:Invent 2018Digital Transformation Playbook in Five Steps (ARC322) - AWS re:Invent 2018
Digital Transformation Playbook in Five Steps (ARC322) - AWS re:Invent 2018
Amazon Web Services
 

Similar to Business Process Automation Using Crowdsourcing (AIM352) - AWS re:Invent 2018 (20)

How Trupanion Became an AI-driven Company for Pets
How Trupanion Became an AI-driven Company for PetsHow Trupanion Became an AI-driven Company for Pets
How Trupanion Became an AI-driven Company for Pets
 
An Agile Approach to Cloud Adoption
An Agile Approach to Cloud AdoptionAn Agile Approach to Cloud Adoption
An Agile Approach to Cloud Adoption
 
Practical Human-in-the-Loop Machine Learning
 Practical Human-in-the-Loop Machine Learning Practical Human-in-the-Loop Machine Learning
Practical Human-in-the-Loop Machine Learning
 
Using Amazon Mechanical Turk to Crowdsource Data Collection (AIM359) - AWS re...
Using Amazon Mechanical Turk to Crowdsource Data Collection (AIM359) - AWS re...Using Amazon Mechanical Turk to Crowdsource Data Collection (AIM359) - AWS re...
Using Amazon Mechanical Turk to Crowdsource Data Collection (AIM359) - AWS re...
 
Innovation at Amazon
Innovation at AmazonInnovation at Amazon
Innovation at Amazon
 
AWS STARTUP DAY 2018 I Innovation @ Amazon
AWS STARTUP DAY 2018 I Innovation @ AmazonAWS STARTUP DAY 2018 I Innovation @ Amazon
AWS STARTUP DAY 2018 I Innovation @ Amazon
 
AWS Think Big Workshop: Experimenting with Data
AWS Think Big Workshop: Experimenting with DataAWS Think Big Workshop: Experimenting with Data
AWS Think Big Workshop: Experimenting with Data
 
Success Story of migrating entire infrastructure from AWS Singapore to AWS Mu...
Success Story of migrating entire infrastructure from AWS Singapore to AWS Mu...Success Story of migrating entire infrastructure from AWS Singapore to AWS Mu...
Success Story of migrating entire infrastructure from AWS Singapore to AWS Mu...
 
Success story of migrating entire infrastructure from AWS Singapore to AWS Mu...
Success story of migrating entire infrastructure from AWS Singapore to AWS Mu...Success story of migrating entire infrastructure from AWS Singapore to AWS Mu...
Success story of migrating entire infrastructure from AWS Singapore to AWS Mu...
 
Leadership Session: Overview of Amazon Digital User Engagement Solutions (DIG...
Leadership Session: Overview of Amazon Digital User Engagement Solutions (DIG...Leadership Session: Overview of Amazon Digital User Engagement Solutions (DIG...
Leadership Session: Overview of Amazon Digital User Engagement Solutions (DIG...
 
Startup Day Kyiv: How we think about Innovation at Amazon
Startup Day Kyiv: How we think about Innovation at Amazon Startup Day Kyiv: How we think about Innovation at Amazon
Startup Day Kyiv: How we think about Innovation at Amazon
 
Tendências na Transformação Digital
Tendências na Transformação DigitalTendências na Transformação Digital
Tendências na Transformação Digital
 
AWS Initiate - Tendências da Transformação Digital
AWS Initiate - Tendências da Transformação DigitalAWS Initiate - Tendências da Transformação Digital
AWS Initiate - Tendências da Transformação Digital
 
Continuously Delivering Your Software on AWS - Adrian White - AWS TechShift A...
Continuously Delivering Your Software on AWS - Adrian White - AWS TechShift A...Continuously Delivering Your Software on AWS - Adrian White - AWS TechShift A...
Continuously Delivering Your Software on AWS - Adrian White - AWS TechShift A...
 
[REPEAT] Better Analytics Through Natural Language Processing (AIM405-R) - AW...
[REPEAT] Better Analytics Through Natural Language Processing (AIM405-R) - AW...[REPEAT] Better Analytics Through Natural Language Processing (AIM405-R) - AW...
[REPEAT] Better Analytics Through Natural Language Processing (AIM405-R) - AW...
 
The Future of API Management Is Serverless
The Future of API Management Is ServerlessThe Future of API Management Is Serverless
The Future of API Management Is Serverless
 
Enterprise Cloud Adoption
Enterprise Cloud Adoption Enterprise Cloud Adoption
Enterprise Cloud Adoption
 
How to Wrangle Data for Machine Learning on AWS
 How to Wrangle Data for Machine Learning on AWS How to Wrangle Data for Machine Learning on AWS
How to Wrangle Data for Machine Learning on AWS
 
Industrial Transformation
Industrial TransformationIndustrial Transformation
Industrial Transformation
 
Digital Transformation Playbook in Five Steps (ARC322) - AWS re:Invent 2018
Digital Transformation Playbook in Five Steps (ARC322) - AWS re:Invent 2018Digital Transformation Playbook in Five Steps (ARC322) - AWS re:Invent 2018
Digital Transformation Playbook in Five Steps (ARC322) - AWS re:Invent 2018
 

More from Amazon Web Services

Come costruire servizi di Forecasting sfruttando algoritmi di ML e deep learn...
Come costruire servizi di Forecasting sfruttando algoritmi di ML e deep learn...Come costruire servizi di Forecasting sfruttando algoritmi di ML e deep learn...
Come costruire servizi di Forecasting sfruttando algoritmi di ML e deep learn...
Amazon Web Services
 
Big Data per le Startup: come creare applicazioni Big Data in modalità Server...
Big Data per le Startup: come creare applicazioni Big Data in modalità Server...Big Data per le Startup: come creare applicazioni Big Data in modalità Server...
Big Data per le Startup: come creare applicazioni Big Data in modalità Server...
Amazon Web Services
 
Esegui pod serverless con Amazon EKS e AWS Fargate
Esegui pod serverless con Amazon EKS e AWS FargateEsegui pod serverless con Amazon EKS e AWS Fargate
Esegui pod serverless con Amazon EKS e AWS Fargate
Amazon Web Services
 
Costruire Applicazioni Moderne con AWS
Costruire Applicazioni Moderne con AWSCostruire Applicazioni Moderne con AWS
Costruire Applicazioni Moderne con AWS
Amazon Web Services
 
Come spendere fino al 90% in meno con i container e le istanze spot
Come spendere fino al 90% in meno con i container e le istanze spot Come spendere fino al 90% in meno con i container e le istanze spot
Come spendere fino al 90% in meno con i container e le istanze spot
Amazon Web Services
 
Open banking as a service
Open banking as a serviceOpen banking as a service
Open banking as a service
Amazon Web Services
 
Rendi unica l’offerta della tua startup sul mercato con i servizi Machine Lea...
Rendi unica l’offerta della tua startup sul mercato con i servizi Machine Lea...Rendi unica l’offerta della tua startup sul mercato con i servizi Machine Lea...
Rendi unica l’offerta della tua startup sul mercato con i servizi Machine Lea...
Amazon Web Services
 
OpsWorks Configuration Management: automatizza la gestione e i deployment del...
OpsWorks Configuration Management: automatizza la gestione e i deployment del...OpsWorks Configuration Management: automatizza la gestione e i deployment del...
OpsWorks Configuration Management: automatizza la gestione e i deployment del...
Amazon Web Services
 
Microsoft Active Directory su AWS per supportare i tuoi Windows Workloads
Microsoft Active Directory su AWS per supportare i tuoi Windows WorkloadsMicrosoft Active Directory su AWS per supportare i tuoi Windows Workloads
Microsoft Active Directory su AWS per supportare i tuoi Windows Workloads
Amazon Web Services
 
Computer Vision con AWS
Computer Vision con AWSComputer Vision con AWS
Computer Vision con AWS
Amazon Web Services
 
Database Oracle e VMware Cloud on AWS i miti da sfatare
Database Oracle e VMware Cloud on AWS i miti da sfatareDatabase Oracle e VMware Cloud on AWS i miti da sfatare
Database Oracle e VMware Cloud on AWS i miti da sfatare
Amazon Web Services
 
Crea la tua prima serverless ledger-based app con QLDB e NodeJS
Crea la tua prima serverless ledger-based app con QLDB e NodeJSCrea la tua prima serverless ledger-based app con QLDB e NodeJS
Crea la tua prima serverless ledger-based app con QLDB e NodeJS
Amazon Web Services
 
API moderne real-time per applicazioni mobili e web
API moderne real-time per applicazioni mobili e webAPI moderne real-time per applicazioni mobili e web
API moderne real-time per applicazioni mobili e web
Amazon Web Services
 
Database Oracle e VMware Cloud™ on AWS: i miti da sfatare
Database Oracle e VMware Cloud™ on AWS: i miti da sfatareDatabase Oracle e VMware Cloud™ on AWS: i miti da sfatare
Database Oracle e VMware Cloud™ on AWS: i miti da sfatare
Amazon Web Services
 
Tools for building your MVP on AWS
Tools for building your MVP on AWSTools for building your MVP on AWS
Tools for building your MVP on AWSAmazon Web Services
 
How to Build a Winning Pitch Deck
How to Build a Winning Pitch DeckHow to Build a Winning Pitch Deck
How to Build a Winning Pitch DeckAmazon Web Services
 
Building a web application without servers
Building a web application without serversBuilding a web application without servers
Building a web application without serversAmazon Web Services
 
AWS_HK_StartupDay_Building Interactive websites while automating for efficien...
AWS_HK_StartupDay_Building Interactive websites while automating for efficien...AWS_HK_StartupDay_Building Interactive websites while automating for efficien...
AWS_HK_StartupDay_Building Interactive websites while automating for efficien...Amazon Web Services
 
Introduzione a Amazon Elastic Container Service
Introduzione a Amazon Elastic Container ServiceIntroduzione a Amazon Elastic Container Service
Introduzione a Amazon Elastic Container Service
Amazon Web Services
 

More from Amazon Web Services (20)

Come costruire servizi di Forecasting sfruttando algoritmi di ML e deep learn...
Come costruire servizi di Forecasting sfruttando algoritmi di ML e deep learn...Come costruire servizi di Forecasting sfruttando algoritmi di ML e deep learn...
Come costruire servizi di Forecasting sfruttando algoritmi di ML e deep learn...
 
Big Data per le Startup: come creare applicazioni Big Data in modalità Server...
Big Data per le Startup: come creare applicazioni Big Data in modalità Server...Big Data per le Startup: come creare applicazioni Big Data in modalità Server...
Big Data per le Startup: come creare applicazioni Big Data in modalità Server...
 
Esegui pod serverless con Amazon EKS e AWS Fargate
Esegui pod serverless con Amazon EKS e AWS FargateEsegui pod serverless con Amazon EKS e AWS Fargate
Esegui pod serverless con Amazon EKS e AWS Fargate
 
Costruire Applicazioni Moderne con AWS
Costruire Applicazioni Moderne con AWSCostruire Applicazioni Moderne con AWS
Costruire Applicazioni Moderne con AWS
 
Come spendere fino al 90% in meno con i container e le istanze spot
Come spendere fino al 90% in meno con i container e le istanze spot Come spendere fino al 90% in meno con i container e le istanze spot
Come spendere fino al 90% in meno con i container e le istanze spot
 
Open banking as a service
Open banking as a serviceOpen banking as a service
Open banking as a service
 
Rendi unica l’offerta della tua startup sul mercato con i servizi Machine Lea...
Rendi unica l’offerta della tua startup sul mercato con i servizi Machine Lea...Rendi unica l’offerta della tua startup sul mercato con i servizi Machine Lea...
Rendi unica l’offerta della tua startup sul mercato con i servizi Machine Lea...
 
OpsWorks Configuration Management: automatizza la gestione e i deployment del...
OpsWorks Configuration Management: automatizza la gestione e i deployment del...OpsWorks Configuration Management: automatizza la gestione e i deployment del...
OpsWorks Configuration Management: automatizza la gestione e i deployment del...
 
Microsoft Active Directory su AWS per supportare i tuoi Windows Workloads
Microsoft Active Directory su AWS per supportare i tuoi Windows WorkloadsMicrosoft Active Directory su AWS per supportare i tuoi Windows Workloads
Microsoft Active Directory su AWS per supportare i tuoi Windows Workloads
 
Computer Vision con AWS
Computer Vision con AWSComputer Vision con AWS
Computer Vision con AWS
 
Database Oracle e VMware Cloud on AWS i miti da sfatare
Database Oracle e VMware Cloud on AWS i miti da sfatareDatabase Oracle e VMware Cloud on AWS i miti da sfatare
Database Oracle e VMware Cloud on AWS i miti da sfatare
 
Crea la tua prima serverless ledger-based app con QLDB e NodeJS
Crea la tua prima serverless ledger-based app con QLDB e NodeJSCrea la tua prima serverless ledger-based app con QLDB e NodeJS
Crea la tua prima serverless ledger-based app con QLDB e NodeJS
 
API moderne real-time per applicazioni mobili e web
API moderne real-time per applicazioni mobili e webAPI moderne real-time per applicazioni mobili e web
API moderne real-time per applicazioni mobili e web
 
Database Oracle e VMware Cloud™ on AWS: i miti da sfatare
Database Oracle e VMware Cloud™ on AWS: i miti da sfatareDatabase Oracle e VMware Cloud™ on AWS: i miti da sfatare
Database Oracle e VMware Cloud™ on AWS: i miti da sfatare
 
Tools for building your MVP on AWS
Tools for building your MVP on AWSTools for building your MVP on AWS
Tools for building your MVP on AWS
 
How to Build a Winning Pitch Deck
How to Build a Winning Pitch DeckHow to Build a Winning Pitch Deck
How to Build a Winning Pitch Deck
 
Building a web application without servers
Building a web application without serversBuilding a web application without servers
Building a web application without servers
 
Fundraising Essentials
Fundraising EssentialsFundraising Essentials
Fundraising Essentials
 
AWS_HK_StartupDay_Building Interactive websites while automating for efficien...
AWS_HK_StartupDay_Building Interactive websites while automating for efficien...AWS_HK_StartupDay_Building Interactive websites while automating for efficien...
AWS_HK_StartupDay_Building Interactive websites while automating for efficien...
 
Introduzione a Amazon Elastic Container Service
Introduzione a Amazon Elastic Container ServiceIntroduzione a Amazon Elastic Container Service
Introduzione a Amazon Elastic Container Service
 

Business Process Automation Using Crowdsourcing (AIM352) - AWS re:Invent 2018

  • 1.
  • 2. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Business Process Automation Using Crowdsourcing Mark Chien GM of Requester Amazon Mechanical Turk A I M 3 5 2
  • 3. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Agenda Introduction to MTurk for business process automation Business process automation customers on MTurk Guest speakers Using MTurk for business process automation
  • 4. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
  • 5. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. What is crowdsourcing? Harnessing the collective power of many individuals
  • 6. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. What is Amazon Mechanical Turk? Amazon Mechanical Turk (MTurk) operates an online crowdsourcing marketplace for work that requires human intelligence
  • 7. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Microwork: Typically small, repeatable tasks that give workers variety and flexibility Each item can be sent to a worker Or multiple workers to increase accuracy Or through multiple steps with different workers What kind of work is on MTurk?
  • 8. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Mechanical Turk customers
  • 9. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. What is business process automation? “Business process automation (BPA) is the technology- enabled automation of complex business processes” - Wikipedia
  • 10. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Choosing the right business processes with MTurk • Processes that can be broken down into smaller, simpler tasks • Tasks require minimal context and can handle high turnover • Handle non-confidential/critical data
  • 11. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Common business process automation use cases on MTurk • Data collection/gathering • Content moderation • Image/Text/Audio/Video transcription • Categorization • Data validation
  • 12. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Advantages of MTurk for business process automation Cost savings • Avoid the overhead and fixed costs associated with hiring and managing a temp workforce • Free up time and resources from your team Elastic scale • Use as much or as little as you need based on your demand • There are no contracts or minimum required spend • 24/7 global workforce to complete your work whenever needed Diversity • Access to 500,000 workers across 190 countries, across various demographics
  • 13. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Finding opportunities to use MTurk • Large teams of temporary or contract workers • Processes that have already been outsourced • Teams doing repetitive, low-context tasks • Teams challenged with handling large deviations in volume of work • Need for 24x7 manual processing
  • 14. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
  • 15. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Content moderation and classification Company: Online travel marketplace Problem: Need help to moderate and categorize hotel images uploaded by users Approach: User-generated images are sent to MTurk in real-time for review MTurk task: Ensure images meet guidelines and categorize them Result: Effectively moderate and categorize these photos before they are posted publicly, so that the most relevant photos are surfaced on the hotel listings
  • 16. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Validating sales accounts Company: Online job search portal Problem: Keeping information about customers up to date in Salesforce Approach: Extract customer websites from Salesforce records and submit to MTurk MTurk task: Confirm that the address is still valid, find out if the company has been acquired, collect updated data points about the company Result: Accurate sales data and more-efficient sales teams
  • 17. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Validating catalog information Company: Consumer travel information service Problem: Need to confirm that recommended restaurants are still open Approach: Have MTurk workers visit restaurant websites to confirm hours of operation MTurk task: Visit the website and confirm that it is still open Result: Users get most up-to-date information about restaurants on their travels
  • 18. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Assessing damage after disasters Company: Disaster assessment Problem: Need to get building damage assessments completed quickly to take action Approach: Have MTurk workers review photos of the damage MTurk task: Assess and categorize the damage Result: Faster response times and more-accurate decisions
  • 19. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Food Genius / US Foods Data Science David Falck Executive Director dave@getfoodgenius.com
  • 20. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. US Foods + Food Genius • Chicago-based start-up, acquired by US Foods in 2016 (2nd largest food distribution company in the US) • Support sellers’ conversations with customers in the field with data • Provide valuable, actionable recommendations based on accurate and fresh restaurant data • Goal: Make positive impact in our sellers’ day- to-day routines (and their wallets)
  • 21. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Background • Small, independent team that needed to scale quickly • Familiarity with crowdsourcing – including code development and data science projects • Experimented with MTurk for research • Early wins spurred continued investment in MTurk • Shared case studies and knowledge with rest of company • Expanding MTurk use cases to other US Foods business units
  • 22. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Why Mechanical Turk? • We use Mechanical Turk to find up-to-date, real data to help our customers • Once we have enough external data, coupling it with our internal data makes it MUCH more valuable…
  • 23. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Mechanical Turk workflows Example 1: Find the “real” website and menus of a restaurant (not delivery, Facebook menus)
  • 24. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Mechanical Turk workflows Example 2: Multi-step information extraction from images OCR
  • 25. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Initial process
  • 26. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Key Learning #1 Feedback from Workers: • We didn’t take into consideration fair payment • We didn’t make our HITs’ instructions clear enough Solution: • Remember, even though you can access their services via APIs, Workers are real people! • Make the instructions super simple, even fun, if you can.
  • 27. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Key Learning #2 Feedback from Workers: • Manual verification took longer than expected, QA of 4,500 URLs took 6 days • Coverage was low too – only about 30% Solution: • Build a cadre of Qualified Workers and use them! • Use Workers to help with QA work • Provide examples, better list of prohibited domains, empty answers will be rejected • QA the work closely and reject work that doesn’t meet our criteria
  • 28. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Results • Our manual work went from 100% to 50% to 25% as we made refinements to our process. • Coverage went from 30%-63% for homepage URLs • Workers are happier – many reach out and inquire when we’re doing a batch of work next • Modifying the original process lets us pivot to other asks quickly and efficiently, such as collecting reviews, social media data, etc. about restaurants
  • 29. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Summary Make crowdsourcing work for your business – You can make the process efficient, cost-effective and scalable by putting yourself in the shoes of a Worker and continuously improving your processes for them Remember: • Listen to the Workers • Develop effective tasks • Measure and enhance your processes
  • 30. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Indeed Andy Hilal Product Manager Indeed Job Spotter
  • 31. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Agenda Job Spotter: Getting local jobs onto Indeed Using MTurk for content moderation Results and impact Making your MTurk program work
  • 32. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
  • 33. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
  • 34. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. OCR and photo analysis “CioSO eli cioso Super Rico Pollo PERUVIAN STYLE CHARCOAL CHICKEN GRAND ESPECIAL Pollo entero PENING$99 con SOPA DE THUR CARNE ASADA PUPUSAS LOMO SALTADO TACOS POLLO FRITO N CHARCOAL CHICKEN”
  • 35. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Machine learning (with humans teaching) What is the business being shown? Is it a hiring sign? • Indeed keeps high standards for job content • No business opportunities • No training programs • No volunteer positions • No elected offices • No sexist language • … and many more
  • 36. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. User-generated content risks Low quality • Low lighting, blurry Objectionable content • Faces of passersby on the street • Worse Fraud • Photos of newspaper want ads • Falsified or duplicated job sign submissions
  • 37. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. From local vendor to MTurk Nearshore Vendor Amazon MTurk Speed of Completion 2841 minutes 34 minutes Quality 9.73% variance 9.93% variance Cost 50 cents / task 7 cents / task
  • 38. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
  • 39. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Our biggest wins Assume goodwill Build clear, fast tools Provide contextual help content Low HIT expiry time Golden questions and clear warnings Price HITs correctly Qualify a closed set of Workers Talk to your Workers :)
  • 40. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
  • 41. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. MTurk concepts primer
  • 42. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Steps for using MTurk for business process automation
  • 43. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Define workflow • Break complex tasks down to simple, discrete tasks • Determine how you want the task flow to work
  • 44. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Design tasks Use HTML/JS/CSS to design your task Provide clear instructions and assume no context Complete the task yourself: • To validate instructions • Determine average completion time to set appropriate reward • To confirm that the task UI works Use the sandbox to test
  • 45. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Set workforce Use Qualifications to determine who gets to do your tasks MTurk system Qualifications • Experience-based (# of approved HITs, approval rate) • Locale (US, states) • Demographics (adult, gender, degree) Define custom Qualifications • Inclusion Qualifications (such as whitelist) • Exclusion Qualifications (such as blacklist)
  • 46. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Load tasks Use the API to load dataset programmaticallyUse the UI to load dataset via CSV Or
  • 47. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Review results Use the API to download results and SNS notifications when tasks are completed Use the UI to view progress and download results via CSV Or
  • 48. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Refine process • Analyze results to assess for quality • Review feedback provided by workers • Update instructions and/or task design • Update qualification membership or requirements • Update reward for desired quality and throughput
  • 49. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Get started at www.mturk.com
  • 50. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Related Sessions Wednesday, Nov 28 Session | AIM351 – Harness the Power of Crowdsourcing with Amazon Mechanical Turk 12:15 – 1:15 p.m. | Aria East, Plaza Level, Orovada 2 Wednesday, Nov 28 Workshop | AIM362 – Crowdsourcing Data Collection with Amazon Mechanical Turk 3:15 – 5:30 p.m. | Venetian, Level 4, Lando 4205 Wednesday, Nov 28 Chalk Talk | AIM358 - Human-in-the-Loop for Machine Learning 7 – 8 p.m. | Aria West, Level 3, Starvine 7
  • 51. Thank you! © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Mark Chien mchien@amazon.com
  • 52. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.