Zoom Logo

Document Understanding Capabilities + Live Demo - Shared screen with speaker view
Sheetal
32:13
yes we have started it
Mustak Rased
32:44
Thank you
Amin Kamali
46:40
What languages are supported by Document Understanding?
Digvijay Mendgudli
47:25
I guess it's mentioned in the taxonomy json file
SyedPasha_UiPathMVP
50:26
Supported Languages :"SupportedLanguages": [{"Name": "Undetermined Language","Code": "und"},{"Name": "Bulgarian","Code": "bul"},{"Name": "Cantonese","Code": "yue"},{"Name": "Danish","Code": "dan"},{"Name": "Dutch","Code": "nld"},{"Name": "English","Code": "eng"},{"Name": "French","Code": "fra"},{"Name": "Georgian","Code": "kat"},{"Name": "German","Code": "deu"},{"Name": "Greek","Code": "ell"},{"Name": "Hebrew","Code": "heb"},{"Name": "Hindi","Code": "hin"},{"Name": "Italian","Code": "ita"},{"Name": "Japanese","Code": "jpn"},{"Name": "Korean","Code": "kor"},{"Name": "Macedonian","Code": "mkd"},{"Name": "Mandarin","
Mukesh
51:02
is studio pro a mandate for taxonomy ?
SyedPasha_UiPathMVP
51:06
Yes it can be found in Taxanomy.json
SyedPasha_UiPathMVP
52:04
No Mukesh - StudioPro is not mandate - Either Studio or StudioPro will work
Sourav Singh
53:43
You forgot ADD HEADERS in write range activity
Sourav Singh
53:54
Ismith
Mustak Rased
53:59
I have two questions: 1. This can extract data from image file too? 2. What if we have multiple dataset available in raw file ? eg . DATE 01/02/2020 and Date 11/02/2020
Gopal Tewari
55:21
can we train the extracted data by human after the next run?
OnePlus 6
55:24
yes we can extract data file too
OnePlus 6
55:39
from image file
Mukesh
55:56
Can we use any other Ocr apart from omni ?
Ankur Gupta
56:01
What in case the OCR doesn't reads the scanned file well?
Mustak Rased
56:08
Ty
Ankur Gupta
56:15
or the scanned copy is not of good quality?
Rohan @Bangalore
56:16
yes u can use other OCR
kiran, Hyderabad
56:33
is this licensed ?
Amin Kamali
56:35
How to train Regex Extractor?
Sourav Singh
56:39
Can we train our bot by providing sample invoices and we don't require to validate always?
OnePlus 6
56:52
yes Mukesh we can user microsoft ocr google ocr
Digvijay Mendgudli
56:54
that is why we use present validation, so that human could give us the correct value
Parth Doshi
56:55
yes sourav
Navneet.Jain delhi
56:59
what if we want to process multiple pdfs . every time user have to validate it or we can create task in orchestrator for same?
Rohan @Bangalore
56:59
Its not mandatory to validate
Mustak Rased
57:02
For regex build : https://regexr.com/
OnePlus 6
57:21
No Rohan
Shubham @Delhi
57:31
For Regex Build I'll suggest: https://regex101.com/
Sourav Singh
57:53
Rohan If the values are having any error we cannot supply it to the client right?
Sourav Singh
58:40
Hi Parth
Parth Doshi
58:52
hey Sourav
Vivek Dixit@WNS,Gurgaon
59:00
we do the data extraction through Kofax too so how beneficial is UIPath in compare to Kofax
Rohan @Bangalore
59:11
its ur choice to validate or not......u can validate either in stdio or in orchestrator in action center
Sahil Sharma
59:20
how does the bot "feels" like which template to choose/use? Can you please elaborate on the logic?
Amin Kamali
59:21
It looks like we can use multiple extractors in Data Extraction Scope. How does that work?
Sourav Singh
59:22
I have seen Ismith webinar earlier with the same ppt.
Mustak Rased
59:32
lost words..
Parth Doshi
01:00:13
okay Sourav
Jesús Mirón López
01:00:22
Can we train our bot by providing sample invoices and we don't require to validate always?
SyedPasha_UiPathMVP
01:00:22
@Sahil - For Each page we need to provide 5 distinct fields to identify the page and BOT uses this logic
Mustak Rased
01:00:42
Can we train bot to capture data similar to IQ Bot?
Sahil Sharma
01:01:02
yes, but which template it uses.....if there a multiple that's my question. How does the bot decide which template to use out of the 5
Rohan @Bangalore
01:01:03
If your using multiple extractors(ex:regex,form) in same data extraction scope ..it choose best one based on confidence level....we can even select which one to choose
Sahil Sharma
01:01:38
and what if the confidence level is same? then which gets preference?
Parth Doshi
01:01:43
https://youtu.be/H8y65QbsMRw @sahil
SyedPasha_UiPathMVP
01:01:45
The Best Extraction Result and it has accuracy/Confidence score numbers obtained after getting each field data
Parth Doshi
01:01:46
go through this
Sourav Singh
01:01:49
Sahil We have classify document scope activity for the identifying the tempate to be used
Sankar
01:02:33
not able to see the ishmeet scree
Sourav Singh
01:02:59
Parth that's what happens with me
Parth Doshi
01:03:19
Sourav can't get your doubt
Sourav Singh
01:03:21
The same as happening with Ismith
Pavan Kumar
01:03:31
share recording
vishnu
01:03:40
Great session so far. what to do incase if the document contains handwritten text and needs to be extracted ?
Mustak Rased
01:04:06
handwritten documents could be extracted using ocr
Rohan @Bangalore
01:04:11
we can extract hand written document has well...bot extractor is diffrent
Sheetal
01:04:35
is everyone able to see the screen ?
Rohan @Bangalore
01:04:36
intelligent form extractor....wont remember exactly
Sankar
01:04:40
I can't able to see
Rohan @Bangalore
01:04:57
yes..we are able to see screen
Mukesh
01:05:00
visible to me
Venkata Karthik Mylavarapu
01:05:02
yes i am able to see
Sheetal
01:05:06
thanks
Sheetal
01:05:26
@Sankar : Can you please restart and check your settings
OnePlus 6
01:05:36
We tried handwritten documents (signature,place,date) but failed to extract
Rohan @Bangalore
01:06:05
SIgnature it can extract,but it can detect whether signature is present or not in document
Rohan @Bangalore
01:06:16
*signature it cannot extract
Rohan @Bangalore
01:06:48
.it can detect wheather signature is present or not in document
Mustak Rased
01:06:53
Oneplus6: it depends upon ocr
OnePlus 6
01:06:53
Ok any idea on checkboxes and radio buttons
Rohan @Bangalore
01:07:22
we can do for checkbox as well..not sure about radio button
Mustak Rased
01:07:55
you can train your ocr engine.. we use abbyy
Rohan @Bangalore
01:07:57
in machine extractor even option for checkbox....
vinod
01:08:04
Again we tried but failed
Parth Doshi
01:08:34
anyone who needs help with du we can get in touch post session on 7620344502
Amin Kamali
01:10:27
So we can train the classifier. How can we train the extractor?
Rohan @Bangalore
01:10:29
If any1 working on Document Undestanding or ABBYY ..can reach out to me for any help 9900220066
Vibhor @ UiPath
01:13:38
Upcoming Sessions:16th July: UiPath AI Computer Vision: Automation without limitations: https://communityevents.uipath.com/e/mja9vy/18th July: Deep Dive into UiPath RE Framework : https://communityevents.uipath.com/e/m2udyg/25th July: End-to-End Automation of your ITSM (ServiceNow and UiPath Working Together)https://communityevents.uipath.com/e/mvz6zt/
Mukesh
01:24:50
If no one takes an action in orchestrator , will the bot be in idle state ? and till what time
Sourav Singh
01:24:57
Please shown extraction for multiple invoices ..which is the main issue to be solved
Sourav Singh
01:25:05
show
Parth Doshi
01:25:27
https://youtu.be/H8y65QbsMRw @Sourav
Sourav Singh
01:26:18
Parth Is it your youtube channel link
Krishan Gopal @ Tecnics Inc, Hyderabad
01:26:42
Hi Ishmeet, I've two questions1) How about handwritten text (including signatures)Can we achieve with the same approach? (Set of activites) or is there any other OCR/ Package has to be used for that?2) What about multiple invoice templates (Un- Structured), how to achieve that
Parth Doshi
01:27:59
https://youtu.be/H8y65QbsMRw: For multiple invoices of different formats
Sourav Singh
01:28:18
2) What about multiple invoice templates (Un- Structured), how to achieve thatsame question
vishnu
01:28:41
any suggestions for extracting informations from an email body ?
Parth Doshi
01:28:46
you can configure templates
Parth Doshi
01:28:54
can you send a dummy invoice
Parth Doshi
01:29:47
you will have to create taxonomy and template for each type
Sourav Singh
01:30:22
Ismith You haven't show yet invoice data extraction for multiple invoices?
Parth Doshi
01:31:04
https://youtu.be/i6Qv11tbU34: for extraction of multiple invoices @Sourav
Sourav Singh
01:32:14
your link is wrong Parth
Sourav Singh
01:32:24
youtu.be
Amin Kamali
01:32:35
Ishmeet, is it possible to train the extractor activities like the classifier activities?
Hemant Kaushik
01:32:36
using ML how confidently it hass extraxted first date only?
Shubham @Delhi
01:32:37
Will the ML extract help me get unstructured data as well ?
Parth Doshi
01:32:42
https://www.youtube.com/watch?v=i6Qv11tbU34
Sourav Singh
01:33:47
Parth I have seen your video even that way I'm unable to do that
Parth Doshi
01:34:02
what error are you getting
Sourav Singh
01:34:40
Its only extrating data from one nvoice that is provided for Template
vishnu
01:38:40
can we use document understanding to extract information from email ?
GAURAV
01:39:27
Will this document understanding work with Acord forms ?
kiran, Hyderabad
01:39:29
is this licensed versio
Parth Doshi
01:40:15
is the bot attended or unattended when we create a validation station?
Sourav Singh
01:40:28
will it work for multiple invoices ?