š Welcome to our very first UiPath Community Dubai meetup!
We have picked some of the hottest topics from the AI-powered Automation agenda to make sure you leave our session with worthy knowledge and food for thought:
š Agenda:
Welcome note
Intelligent Document Processing in UiPath
UiPath Document Understanding is a feature that integrates with UiPath's RPA platform to streamline the extraction of information from documents.
1. 'Document Types': supports wide range document
2. 'Intelligent Data Extraction': Utilizes machine learning models for accurate extraction of structured data from unstructured documents.
3. Classifier and Extractor Activities: UiPath provides activities like the Document Classification and Data Extraction activities to train models and extract information.
4. Learning Scope: Allows the system to learn from human feedback, improving accuracy over time.
5. Validation Station: A user-friendly interface for reviewing and correcting extracted data, enhancing the learning process.
6. Integration with Orchestrator: Seamlessly integrates with UiPath Orchestrator for centralized management and monitoring of document processing.
7. Scalability: Scales to handle large volumes of documents, making it suitable for enterprises with diverse document processing needs.
UiPath Document Understanding enhances automation capabilities by making it easier to handle documents with varied formats and extract valuable information accurately.
Shape the future of Document Understanding with Extended Capabilities
Active learning is a next-gen AI-powered experience within UiPath Document Understanding. It is designed to drive 80% faster model training, decreasing the number of samples required when creating customer-specific models.
UiPath's Document Understanding, integrated with Active Learning, enables continuous improvement of machine learning models by leveraging human feedback. Let see the capability:
1. Human-in-the-Loop Learning: Active Learning involves human reviewers validating and correcting model predictions, providing feedback to improve accuracy.
2. Incremental Model Training: As the system receives feedback, it iteratively updates and refines the machine learning models, adapting to new patterns and variations in document structures.
3. Adaptive Data Learning: Active Learning helps the model focus on ambiguous or challenging cases, enhancing its ability to handle diverse document types.
4. Reduced Annotation Effort: By prioritizing uncertain or critical instances for human review, Active Learning minimizes the amount of labeled data needed for training.
5. Enhanced Accuracy Over Time: The system becomes more accurate as it learns from user interactions, resulting in improved document understanding and data extraction capabilities.
UiPath's incorporation of Active Learning in Document Understanding contributes to a more dynamic and adaptable approach to handling document processing tasks within automation workflows.
2. 2
Safe Harborā
This presentation may include forward-looking statements. Forward looking statements include all statements that are not historical facts, and in
some cases, can be identified by terms such as āanticipate,ā ābelieve,ā āestimate,ā āexpect,ā āintend,ā āmay,ā āmight,ā āplan,ā āproject,ā āwill,ā āwould,ā
āshould,ā ācould,ā ācan,ā āpredict,ā āpotential,ā ācontinue,ā or the negative of these terms, and similar expressions that concern our expectations,
strategy, plans or intentions. By their nature, these statements are subject to numerous risks and uncertainties, including factors beyond our
control, that could cause actual results, performance or achievement to differ materially and adversely from those anticipated or implied in the
statements. Although our management believes that the expectations reflected in our statements are reasonable, we cannot guarantee that the
future results, levels of activity, performance or events and circumstances described in the forward-looking statements will be achieved or occur.
Recipients are cautioned not to place undue reliance on these forward-looking statements, which speak only as of the date such statements are
made and should not be construed as statements of fact. ā
This meeting is strictly confidential. By participating in this meeting, you agree to keep any information we provide confidential and not to disclose
any of the information to any other parties without our prior express written permission. Neither the information contained in this presentation, nor
any further information made available by us or any of our affiliates or employees, directors, representatives, officers, agents or advisers in
connection with this presentation will form the basis of or be construed as a contract or any other legal obligation. ā
3. 3
UiPath Dubai Chapter
A UiPath Community Initiative
Unleashing the force of AI-powered intelligent
document processing
Session Organizers
4. 4
Our Speakers
Ashraf El Zarka
UiPath
Vice President and
Managing Director UiPath
Middle East and Africa
Sofiane Sahir
UiPath
Principal Sales Engineer
5. 5
Agenda
ā Intelligent Document Processing in UiPath
UiPath Document Understanding is a feature that integrates with UiPath's RPA platform to streamline the extraction of information from
documents.
ā Active Learning in Document Understanding (Public Preview)
next-gen AI-powered experience within UiPath Document Understanding.
It is designed to drive 80% faster model training, decreasing the number of samples required when creating customer-specific models.
UiPath's Document Understanding, integrated with Active Learning, enables continuous improvement of machine learning models by leveraging
human feedback.
Document
Types
Intelligent
Data
Extraction
Classifier
and
Extractor
Activities
Learning
Scope.
Validation
Station
Integration
with
Orchestrator
Scalability
7. 7
AI-powered automation
Action
UI
API
Human
in the Loop
Open | Flexible | Responsible
AI Solutions
Generative AI
Creative
Broad
Non-deterministic
Specialized AI
Fast
Accurate
Tailored
AI Trust Layer
Context
Documents
Data
Processes
People
Communications
Automation delivers AIās value
8. 8
Get your documents
processed intelligently
Teach your robots to understand documents
using AI-enhanced skills for data extraction
and interpretation.
Drag and drop these capabilities directly into
your automation workflows to embed AI
9. 9
ā¢ Like forms, passports, licenses, time
sheets
ā¢ Fixed in format and can contain
handwriting, signatures, checkboxesā
ā¢ Like invoices, receipts, purchase
orders, medical bills, utility bills
ā¢ Containing fixed and variable parts
like tables
ā¢ Like contracts, agreements, emails,
scripts, drug prescriptions, news
ā¢ No fixed format, free-form
sentences/paragraphs
Which documents can be handled by
Document Understanding?
Structured documents Semi-structured documents Unstructured documents
10. 10
1 2 3
Understand Act
Receive
End-to-end intelligent document
processing (IDP) solution
Extracts relevant data from the
documents
Requests or
communications with
attached documents:
ā¢ Multiple languages
ā¢ Various formats
ā¢ Handwriting
ā¢ Signatures
ā¢ Skewed & low-quality scans
ā¢ Checkboxes
ā¢ Tables
Extracts key intent, sentiment and
context data from messages
Human in the loop
Asking employees to validate the
results if required or in case of
inaccuracies and exceptions.
UiPath Automation
Route the extracted actions and data
to downstream systems for further
processing.
11. 11
Built for: RPA and citizen developers ā no data science skills required
Built-in AI: Specialized AI models and Generative AI
Built as: SaaS, APIs, self-hosted or air-gapped integrated with UiPath Business Automation Platform
1 Receive
3 Act
2 Understand
Documents
ā¢ Multiple languages
ā¢ Various types & formats
ā¢ Handwriting & signatures
ā¢ Tables & checkboxes
ā¢ Skewed & low-quality scans
AI-powered processing Human in the loop UiPath Automation
ā¢ Initiate other automations
ā¢ Input into system of record
ā¢ Drive other actions
Validate the
extracted information
and handle exceptions
Digitize Classify Extract
12. 12
Load taxonomy for
your documents
How Document Understanding works
UiPath Studio
Digitize images using
multiple OCRs
Classify documents Extract named entities
in a taxonomy
Validate and train
supervised models
Export extracted data
2 3 4 5 6
Digitize Classify
Train & Validate
Extract
2 3
5
4
Structured
Semi-structured
Unstructured
Export
RPA
BPM
API
Other systems
6
1
Define taxonomy ā
document types and fields
1
13. 13
Taxonomy manager is
used once at the start to
define the collection of
documents that you
would want to process
as well as business rules.
Additionally, you can
describe what data you
would like to extract.
Load taxonomy
14. 14
Move-For-You Co.
We move so you donāt need to move
PO: NP74006735
1 February 2020
PAYABLE WITHIN 15 DAYS OF RECEIPT
20800 ALMADEN AVE, SUITE 404
SAN JOSE, CA 95120-0520
T: +1 425 555 9876
F: +1 425 555 3456
E: billing@moveforyou-co.com
www.moveforyou-co.com
Bill To:
Tony Tzeng
12345 Mango Lane
Seattle, WA 98108
INVOICE DETAILS
Packing services
Storage fees (1 month)
House move (white-glove service)
Vehicle storage and transport
Sales tax 10%
Total Fee including Tax
FEE
$1,282.00
$1,884.00
$5,320.00
$5,186.00
$1,367.20
$15,039.20
Methods of payment
Personal Check: Move-For-You LLC
Wire Transfer: BigBank Co., Account 123456789-0987ABC
Invoice No: 456200-TZE1
Digitize text in the documents
using OCR
15. 15
Classify and split the documents
Documents scanned into one file
isnāt a problem ā owing to
classifiers, the robot can identify
the document types and split the file
to process the documents
accordingly.
Document Understanding offers
different classification capabilities
ranging from keyword-based to
ML and Gen AI based classification.
16. 16
Validate classification of the
documents
Classification Station is
used to check, correct, and
confirm the results of
document classification and
splitting.
17. 17
You can easily configure
data extraction to choose
most suitable extractor
for each field.
Use a combination of rule-
based and AI-based
approaches to ensure
smooth and accurate
processing of different
documents.
Extract data from the documents
18. 18
Generative
Extractor
Data extraction ā from rules to AI
to a hybrid approach
RegEx-Based
Extractor
Machine Learning
Extractor
Form Extractor
A combination of both ā
rule-based and AI-based approaches
Rule-based
Forms AI
AI-based
Hybrid approach
Note: Forms AI will be soon replaced with the next-gen Document Understanding experience designed for building Specialized AI models for different document types
19. 19
Generative
Extractor
Data extraction ā from rules to AI
to a hybrid approach
RegEx-Based
Extractor
Machine Learning
Extractor
Form Extractor
A combination of both ā
rule-based and AI-based approaches
Rule-based
Forms AI
AI-based
Hybrid approach
Note: Forms AI will be soon replaced with the next-gen Document Understanding experience designed for building Specialized AI models for different document types
20. 20
Generative
Extractor
Mostly unstructured
documents
Data extraction ā from rules to AI
to a hybrid approach
RegEx-Based
Extractor
Structured fields
Machine Learning
Extractor
Mostly semi-structured
documents
Form Extractor
Structured documents
(forms)
A combination of both ā
rule-based and AI-based approaches
Mostly documents combining both structured and less structured formats
Rule-based
Forms AI
Structured documents
(forms)
AI-based
Hybrid approach
Note: Forms AI will be soon replaced with the next-gen Document Understanding experience designed for building Specialized AI models for different document types
21. 21
Rule-based or template-based approach
Relies on rules (like regular
expressions) and templates
(including anchors)
Processes fixed in format
structured data
Ensures high accuracy for
already known documents
22. 22
Specialized AI out-of-the-box models
to accelerate time to value
Human resources, know your
customer (KYC), employee
onboarding, tax
Supply chain, shipping,
manufacturing, retail
Finance, accounts
payable, accounts
receivable, procurement
Lending, banking
Insurance, healthcare, claims
processing
Passports Bill of Lending Invoices Generic Bank Statements
ACORD 25 Certificate of Liability
Insurance
Utility Bills
Certificate of Incorporation/Good
Standing
Invoices Japan
FM 1003 Uniform Residential Loan
Application Form
ACORD 125 Commercial
Insurance Application
W2 Wage and Tax Statement Certificate of Origin Invoices China Financial Statements
ACORD 126 Commercial General
Liability Insurance Section Form
990 Return of Organization Exempt from
Income Tax - Preview
Packing Lists Invoices India Checks
ACORD 131 Umbrella / Excess
Insurance Section Form
W9 Request for Taxpayer Identification
Number and Certification
Children Product Certificate Invoices Australia Vehicle Titles
ACORD 140 Property Insurance
Section Form
1040 US Income Tax Form EU Declaration of Conformity Shipping Invoices
CMS 1500 Health Insurance Claim
Form
ID Cards and Driverās Licenses Shipping Invoices Purchase Orders UB04
4506-T Request for Transcript of Tax
Return
Receipts
I9 Employment Eligibility Verification
Form
Remittance Advice
Generic Document Understanding model to build a model from scratch for any document type, and more Specialized AI models:
The list is not exhaustive. All the models can be trained on your own data to improve their performance for your specific requirements.
23. 23
Gen AI is perfect for
unstructured documents
No need to train custom ML models to
answer questions and summarize content
in free-form unstructured documentsā
use Gen AI instead
Generative extraction
24. 24
Validate the extracted
Information and handle
exceptions using
Validation Station.
Now, ML models can also
be retrained using the data
confirmed or corrected in
the Validation Station.
Validate extraction of the
documents
25. 25
Let the classifiers and
extractors learn from the
data corrected and
validated in the
Classification Station and
Validation Station
respectively.
Learn about sharing data for model retraining here.
Train classifiers and extractors
26. 26
Export the extracted data
End-to-end intelligent
document processing
Start & continue the document
processing workflows with other
automation components. Export the data
for further usage/automation ā for
example, to an Excel spreadsheet, send
as email, SAP, and so on.
27. 27
PROCESS MINING TASK MINING COMMUNICATIONS MINING IDEA CAPTURE & MANAGEMENT
LOW-CODE DEVELOPMENTā UI & API AUTOMATIONā PROCESS ORCHESTRATIONā
INTELLIGENT
DOCUMENT PROCESSINGā
INTEGRATED NLP & AI/ML
Discover
Continuously uncover opportunities for process and task
improvements, āhelping you identify the highest-ROI areas
Automate
Get more done with a digital workforce that seamlessly collaborates with your
people āand automates work via UI and API, powered with native-integrated AIā
Your Applications Your
People
Your
Processes
ANALYTICS CONTINUOUS TESTING
UNIFIED MANAGEMENT
& GOVERNANCEā
FLEXIBLE DEPLOYMENT
Operate
Get an enterprise-grade foundation to run and optimize
a āmission-critical automation program at high scale
29. 29
Generative AI capabilities in Document Understanding
What?
Generative Annotation, Generative
Classification, Generative Extraction
Why?
Generative Annotation (Pre-labeling):
Annotate any document samples with Gen AI,
speeding training from a week to 1-2 days for
complex scenarios, or minutes for simpler forms
Generative Classification: Classifying
documents is fast and easy with Gen AIā
just define the document types; no need to write
rules or train new ML models
Generative Extraction: No need to train
custom ML models to answer questions and
summarize content in free-form unstructured
documentsāuse Gen AI instead
Where and when?
GA in Automation Cloud and activities
in 2023.10
30. 30
Next-generation Document Understanding
with active learning
What?
Active learning is a next-gen AI-
powered experience within UiPath
Document Understanding
Why?
ā¢ Anyone can train AI modelsāno coding
or ML skills required
ā¢ 80% faster model trainingāfrom a
week, down to just a day
ā¢ Guidance on model optimizationā
humans & AI collaborating together
ā¢ Instant model evaluationābuilt-in model
performance analytics
Where and when?
Private Preview via
Automation Cloud in 2023.10
31. 31
Bringing a new unified experience to
training, deploying and monitoring your model
Previously across Document Understanding, Document Manager, AI Center
Build
Load samples, annotate, and
train your model with a guided
experience
Measure
Understand your model
performance
and how to improve it
Run
Deploy and manage
your projects
with ease
Monitor
Monitor and audit
the performance of
your automation
Note: Preview limitations: Currently we can classify and annotate 40 common document types, limited to out-of-the-box Specialized AI models, full list at https://docs.uipath.com/document-understanding/automation-
cloud/latest/user-guide/out-of-the-box-pre-trained-ml-packages; will be addressed and on roadmap for FA in 24.4
32. 32
Next-gen Document Understanding
Experience with Active Learning:
Build
Build
Load samples, annotate, and train your
model with a guided experience
āŖ Upload and manage your documents
āŖ Classify documents automatically
āŖ Use Generative AI to automatically annotate
fields that users validate
āŖ Follow recommendations that guide you
on how to improve your model efficiently
āŖ Model training occurs in the background
to see impact of user actions
Note: Preview limitations: Currently we can classify and annotate 40 common document types, limited to out-of-the-box Specialized AI models, full list at https://docs.uipath.com/document-understanding/automation-
cloud/latest/user-guide/out-of-the-box-pre-trained-ml-packages; will be addressed and on roadmap for FA in 24.4
33. 33
Next-gen Document Understanding
Experience with Active Learning:
Measure
Measure
Understand your model performance
and how to improve it
āŖ Evaluate your project readiness with model
metrics
āŖ Model metrics indicate which parts of the
project need the most attention
āŖ Identified and takes user to the part of the
project with recommended actions to
improve model performance
āŖ See distribution of dataset
34. 34
Next-gen Document Understanding
Experience with Active Learning:
Run
Run
Deploy and manage your projects
with ease
āŖ Evaluate your project readiness with model
Choose and deploy model in just a couple
of clicks
āŖ Easily version your entire project
āŖ Start working in studio, studio web, or via API
35. 35
Next-gen Document Understanding
Experience with Active Learning:
Monitor
Monitor
Monitor and audit the performance
of your automation
āŖ Dashboard integrated to show business
outcomes of the automation
āŖ Easily drill down into status of documents
being processed
āŖ Visibility into human in the loop events āsee
what the validator corrected for full
transparency
āŖ Search through history of documents
processed
37. 37
Thank You!!
Donāt forget to join us for upcoming
sessions in our new virtual home
@UiPath Community Dubai
Join UiPath Community Dubai chapter here:
https://community.uipath.com/dubai/