SlideShare a Scribd company logo
1 of 23
DRAGNN: A Transition­based Framework for 
Dynamically Connected Neural Networks
Lingpeng Kong et al. [2017]
Review editted by PyunginPaek@modulabs
Table of Contents
● Introduction of DRAGNN
● Transition system & TBRU
● Examples of TBRU
● Combinations of TBRUs
● Experiments and Discussions
               Intro. of DRAGONDRAGNN*
● General Model for Structured Prediction
– By connecting multiple TBRUs, we can extend to 
architectures such as seq2seq, attention mechanism, 
recursive tree models.. 
DRAGNN 
TBRU
TBRU
TBRU TBRU
* Kong et al. [2017] A Transition­based Framework for Dynamically Connected Neural Networks ,   https://arxiv.org/abs/1703.04474
SyntaxNet­II (from Google)
● Major upgrade to SyntaxNet (Mar/15, 2017)
– Multilingual language understanding
– Joint modeling of multiple levels of linguistic structure
– Neural­network architectures to be created dynamically during processing of a 
sentence or document
● Implementation*
 : combination of below 2.
– Recurrent multi­task parsing model (with DRAGNN)
– Character­based representation (with LSTM) as input to DRAGNN
● DRAGNN as the new core
– Architecture as a series of modular unit(TBRU)s
– Connections between modules are unfolded dynamically.
 → Dynamic Recurrent Acyclic Graphcal Neural Networks
* Alberti et al. [2017] SyntaxNet Models for the CoNLL 2017 Shared Task ,   
   (paper) https://arxiv.org/abs/1703.04929  (code) https://github.com/tensorflow/models/tree/master/syntaxnet
(Roughly Speaking) Generalization??
RNN module
LSTM, GRU, ...
Structured Models
Seq2seq, recursive tree, 
bi­LSTM tagger, 
Encoder/Decoder...
Generalize TBRU
Generalize DRAGNN
Table of Contents
● Introduction of DRAGNN
● Transition system & TBRU
● Examples of TBRU
● Combinations of TBRUs
● Experiments and Discussions
Transition System
● Transition System : T 
– T = {S, A, t}
– S(x) : Set of States  ( s+
 ∈ S(x) : Start State  )
– x :  input  (eg. sentence)
– A(s, x) : Set of allowed decision  for any s ∈  S
– transition function t(s, d, x)
● s’ = t(s, d, x)   : new state s’ for any decision d ∈ A(s, x)
● s’ = t(s, d)  for brevity
● Complete structure : sequence of state/decision pairs 
– (s1
, d1
)...(sn
, dn
)
– s1 = s+
 ,  di
 ∈ A(s, x)
– si+1
 = t(si
, di
)
Transition Based Recurrent Unit (TBRU)
● m(s) : Input embedding function
– eg. lookup op  m(s) : S   R→ k
● r(s) : recurrence function (Connection to previous states) 
– mapping : states   set of previous time steps→
– r(si
) : S   P{1, …, i­1}  ( P : power set, so n(P) variable)→
● (RNN) Network Cell : computes new hidden representation
– hs
 = RNN( m(s), {hi
 | i ∈ r(s)} ) 
( Logically : r,m   h   d )→ →
di
   argmax← d   A(si)∈
wd
T
hi
Inference with TBRUs
1. Initialize s1
 = s+
2. For i = 1, …, n:
(a) Update the hidden state:
hi
   RNN(m(s← i
), {hj
 | j   r(s∈ i
)})
(b) Update the transition state:
di
   argmax← d   A(si)∈
wd
T
hi
si+1
   t(s← i
, di
) 
Table of Contents
● Introduction of DRAGNN
● Transition system & TBRU
● Examples of TBRU
● Combinations of TBRUs
● Experiments and Discussions
TBRU Ex.(1) Simple LSTM Tagger
input x sequence of word embedding
x = {x1
, …, xn
}
output d sequence of tags
d1
, …, dn
discrete state s state = previous sequence of tags
si
 = {d1
, …, di­1
}
input embedding 
m(s)
word embedding for the next token
m(si
) = xi
recurrence function 
r(s)
network connected only to the previous state
r(si
) = {i­1}
RNN and hidden 
state h
Simple LSTM cell
hi
 = RNN(xi
, hi­1
)
Transition
t(s, d)
si+1
   t(s← i
, di
) as LSTM cell state
TBRU Ex.(2) SyntaxNet­I* 
input x sequence of character­level multi­scale representation
x = {x1
, …, xn
}
output d     if action = {Shift}:                   d = input word
      
elif action = {Left/Right Arc}: d = partial­built tree
discrete state s State = stack squence of {d1
, …, di­1
}
It contains all words and partial­built trees on the stack
input embedding 
m(s)
Sentence structure projected to 52 feature embedding
recurrence function 
r(s)
Empty
r(si
) = {}
RNN and hidden 
state h
Feed­forward MLP (Not RNN!!) 
di
 = MLP(m(si
))
Transition
t(s, d)
si+1
   t(s← i
, di
) as next stack sequence
* Andor et al. [2016] Globally Normalized Transition­based Neural Networks,    https://arxiv.org/abs/1603.06042
Arc­standard
Transition System
Refer Andor et al. [2016] 
Global Normalization
Arc­Standard Transition System*
Input Word Top Tree node On the Stack
Bob Bob
gave gave
Alice gave
a gave a
pretty gave a pretty
flower gave flower
on gave flower on
Monday gave flower on
. gave
* Nivre [2006] Inductive dependency parsing.  
Table of Contents
● Introduction of DRAGNN
● Transition system & TBRU
● Examples of TBRU
● Combinations of TBRUs
● Experiments and Discussions
Now, let’s combine TBRUs !!
● Multiple TBRUs, having different recurrence r(s).
– So, they act as different transition task.
● Multiple TBRUs, but run one at a time, after another.
– Multiple TBRUs have common global step­counter.
● Then, if a hidden state hA
 of TBRU A is connected to 
recurrence r(sB
) of TBRU B??
● Possibly, might we reach to multi­task, hierarchical learning?
TBRU Combination Examples (1)
Important Note:
A TBRU can be both 
a decoder of it’s own task
and
an encoder of other TBRU’s of another TBRU
at the same time.
Stack Propagation
● Using explicit structure improves encoder­decoder.
–  For newswire text, TBRU exceeds pure encoder­decoder seq2seq.
● Utilizing parse representations improves summarization.
– Multi­task learning setup
● Deep stacked bi­directional parsing
– Bi­directional approach gives significant advantages.
– Final model (5 TBRUs) : 1 L­R pos­tagging   2 shift only   2 arc­parser bi­directionally→ →
TBRU Combination Examples (2)
Table of Contents
● Introduction of DRAGNN
● Transition system & TBRU
● Examples of TBRU
● Combinations of TBRUs
● Experiments and Discussions
How to train a DRAGNN
● Assumption
– there should be at least one TBRU for which example x along with gold decision(d*
)  
sequence. 
● Log­likelihood objective
– For TBRU which has gold 
decision sequence,
–  Θ are combined parameters across all TBRUs.
● Where decision d1:N
 from?
– 1) case when they are part of gold annotations
– 2) prediction from unrolling previous TBRU
● In practice, task sampling is usually preceded by pre­training steps.
Experiment Result (1)
● Using explicit structure improves encoder­decoder.
–  For newswire text, TBRU exceeds pure encoder­decoder seq2seq.
● Utilizing parse representations improves summarization.
– Multi­task learning setup
Experiment Result (2)
● Deep stacked bi­directional parsing
– Bi­directional approach gives significant advantages.
– Final model (5 TBRUs) : 1 L­R pos­tagging   2 shift →
only   2 arc­parser bi­directionally→
Conclusion & Discussions
● Primarily focused on syntactic parsing, but can it 
provide general means of sharing representations 
between tasks?
● Can this approach be globally normalized?
● Multi­task learning possible with a single model?
– By combining translation, summarization, and even 
abstraction and reasoning.
– And beyond language processing

More Related Content

Similar to [Paper Review] DRAGNN

Real time text stream processing - a dynamic and distributed nlp pipeline
Real time text stream  processing - a dynamic and distributed nlp pipelineReal time text stream  processing - a dynamic and distributed nlp pipeline
Real time text stream processing - a dynamic and distributed nlp pipelineConference Papers
 
Querying Mongo Without Programming Using Funql
Querying Mongo Without Programming Using FunqlQuerying Mongo Without Programming Using Funql
Querying Mongo Without Programming Using FunqlMongoDB
 
Xbase implementing specific domain language for java
Xbase  implementing specific domain language for javaXbase  implementing specific domain language for java
Xbase implementing specific domain language for javaYash Patel
 
Open Chemistry, JupyterLab and data: Reproducible quantum chemistry
Open Chemistry, JupyterLab and data: Reproducible quantum chemistryOpen Chemistry, JupyterLab and data: Reproducible quantum chemistry
Open Chemistry, JupyterLab and data: Reproducible quantum chemistryMarcus Hanwell
 
[Nuxeo World 2013] Nuxeo Studio
[Nuxeo World 2013] Nuxeo Studio[Nuxeo World 2013] Nuxeo Studio
[Nuxeo World 2013] Nuxeo StudioNuxeo
 
BISSA: Empowering Web gadget Communication with Tuple Spaces
BISSA: Empowering Web gadget Communication with Tuple SpacesBISSA: Empowering Web gadget Communication with Tuple Spaces
BISSA: Empowering Web gadget Communication with Tuple SpacesSrinath Perera
 
Document Summarization
Document SummarizationDocument Summarization
Document SummarizationPratik Kumar
 
Performance Comparison between Pytorch and Mindspore
Performance Comparison between Pytorch and MindsporePerformance Comparison between Pytorch and Mindspore
Performance Comparison between Pytorch and Mindsporeijdms
 
A template based algorithm for automatic summarization and dialogue managemen...
A template based algorithm for automatic summarization and dialogue managemen...A template based algorithm for automatic summarization and dialogue managemen...
A template based algorithm for automatic summarization and dialogue managemen...eSAT Journals
 
What are the major components of MongoDB and the major tools used in it.docx
What are the major components of MongoDB and the major tools used in it.docxWhat are the major components of MongoDB and the major tools used in it.docx
What are the major components of MongoDB and the major tools used in it.docxTechnogeeks
 
Apache Spark 101 - Demi Ben-Ari
Apache Spark 101 - Demi Ben-AriApache Spark 101 - Demi Ben-Ari
Apache Spark 101 - Demi Ben-AriDemi Ben-Ari
 
International Journal of Engineering Research and Development (IJERD)
International Journal of Engineering Research and Development (IJERD)International Journal of Engineering Research and Development (IJERD)
International Journal of Engineering Research and Development (IJERD)IJERD Editor
 
Sharing of cluster resources among multiple Workflow Applications
Sharing of cluster resources among multiple Workflow ApplicationsSharing of cluster resources among multiple Workflow Applications
Sharing of cluster resources among multiple Workflow Applicationsijcsit
 
Greedy Enough for the Grid?
Greedy Enough for the Grid?Greedy Enough for the Grid?
Greedy Enough for the Grid?Matteo Romanello
 
Analysis of the evolution of advanced transformer-based language models: Expe...
Analysis of the evolution of advanced transformer-based language models: Expe...Analysis of the evolution of advanced transformer-based language models: Expe...
Analysis of the evolution of advanced transformer-based language models: Expe...IAESIJAI
 

Similar to [Paper Review] DRAGNN (20)

Real time text stream processing - a dynamic and distributed nlp pipeline
Real time text stream  processing - a dynamic and distributed nlp pipelineReal time text stream  processing - a dynamic and distributed nlp pipeline
Real time text stream processing - a dynamic and distributed nlp pipeline
 
Querying Mongo Without Programming Using Funql
Querying Mongo Without Programming Using FunqlQuerying Mongo Without Programming Using Funql
Querying Mongo Without Programming Using Funql
 
short_story.pptx
short_story.pptxshort_story.pptx
short_story.pptx
 
Xbase implementing specific domain language for java
Xbase  implementing specific domain language for javaXbase  implementing specific domain language for java
Xbase implementing specific domain language for java
 
Open Chemistry, JupyterLab and data: Reproducible quantum chemistry
Open Chemistry, JupyterLab and data: Reproducible quantum chemistryOpen Chemistry, JupyterLab and data: Reproducible quantum chemistry
Open Chemistry, JupyterLab and data: Reproducible quantum chemistry
 
[Nuxeo World 2013] Nuxeo Studio
[Nuxeo World 2013] Nuxeo Studio[Nuxeo World 2013] Nuxeo Studio
[Nuxeo World 2013] Nuxeo Studio
 
express.pdf
express.pdfexpress.pdf
express.pdf
 
BISSA: Empowering Web gadget Communication with Tuple Spaces
BISSA: Empowering Web gadget Communication with Tuple SpacesBISSA: Empowering Web gadget Communication with Tuple Spaces
BISSA: Empowering Web gadget Communication with Tuple Spaces
 
Document Summarization
Document SummarizationDocument Summarization
Document Summarization
 
Spark
SparkSpark
Spark
 
Performance Comparison between Pytorch and Mindspore
Performance Comparison between Pytorch and MindsporePerformance Comparison between Pytorch and Mindspore
Performance Comparison between Pytorch and Mindspore
 
A template based algorithm for automatic summarization and dialogue managemen...
A template based algorithm for automatic summarization and dialogue managemen...A template based algorithm for automatic summarization and dialogue managemen...
A template based algorithm for automatic summarization and dialogue managemen...
 
What are the major components of MongoDB and the major tools used in it.docx
What are the major components of MongoDB and the major tools used in it.docxWhat are the major components of MongoDB and the major tools used in it.docx
What are the major components of MongoDB and the major tools used in it.docx
 
Apache Spark 101 - Demi Ben-Ari
Apache Spark 101 - Demi Ben-AriApache Spark 101 - Demi Ben-Ari
Apache Spark 101 - Demi Ben-Ari
 
master_thesis_greciano_v2
master_thesis_greciano_v2master_thesis_greciano_v2
master_thesis_greciano_v2
 
International Journal of Engineering Research and Development (IJERD)
International Journal of Engineering Research and Development (IJERD)International Journal of Engineering Research and Development (IJERD)
International Journal of Engineering Research and Development (IJERD)
 
Sharing of cluster resources among multiple Workflow Applications
Sharing of cluster resources among multiple Workflow ApplicationsSharing of cluster resources among multiple Workflow Applications
Sharing of cluster resources among multiple Workflow Applications
 
Distributed Coordination-Based Systems
Distributed Coordination-Based SystemsDistributed Coordination-Based Systems
Distributed Coordination-Based Systems
 
Greedy Enough for the Grid?
Greedy Enough for the Grid?Greedy Enough for the Grid?
Greedy Enough for the Grid?
 
Analysis of the evolution of advanced transformer-based language models: Expe...
Analysis of the evolution of advanced transformer-based language models: Expe...Analysis of the evolution of advanced transformer-based language models: Expe...
Analysis of the evolution of advanced transformer-based language models: Expe...
 

Recently uploaded

AI in Action: Real World Use Cases by Anitaraj
AI in Action: Real World Use Cases by AnitarajAI in Action: Real World Use Cases by Anitaraj
AI in Action: Real World Use Cases by AnitarajAnitaRaj43
 
Microsoft BitLocker Bypass Attack Method.pdf
Microsoft BitLocker Bypass Attack Method.pdfMicrosoft BitLocker Bypass Attack Method.pdf
Microsoft BitLocker Bypass Attack Method.pdfOverkill Security
 
Working together SRE & Platform Engineering
Working together SRE & Platform EngineeringWorking together SRE & Platform Engineering
Working together SRE & Platform EngineeringMarcus Vechiato
 
CORS (Kitworks Team Study 양다윗 발표자료 240510)
CORS (Kitworks Team Study 양다윗 발표자료 240510)CORS (Kitworks Team Study 양다윗 발표자료 240510)
CORS (Kitworks Team Study 양다윗 발표자료 240510)Wonjun Hwang
 
Event-Driven Architecture Masterclass: Challenges in Stream Processing
Event-Driven Architecture Masterclass: Challenges in Stream ProcessingEvent-Driven Architecture Masterclass: Challenges in Stream Processing
Event-Driven Architecture Masterclass: Challenges in Stream ProcessingScyllaDB
 
Tales from a Passkey Provider Progress from Awareness to Implementation.pptx
Tales from a Passkey Provider  Progress from Awareness to Implementation.pptxTales from a Passkey Provider  Progress from Awareness to Implementation.pptx
Tales from a Passkey Provider Progress from Awareness to Implementation.pptxFIDO Alliance
 
Introduction to FIDO Authentication and Passkeys.pptx
Introduction to FIDO Authentication and Passkeys.pptxIntroduction to FIDO Authentication and Passkeys.pptx
Introduction to FIDO Authentication and Passkeys.pptxFIDO Alliance
 
AI+A11Y 11MAY2024 HYDERBAD GAAD 2024 - HelloA11Y (11 May 2024)
AI+A11Y 11MAY2024 HYDERBAD GAAD 2024 - HelloA11Y (11 May 2024)AI+A11Y 11MAY2024 HYDERBAD GAAD 2024 - HelloA11Y (11 May 2024)
AI+A11Y 11MAY2024 HYDERBAD GAAD 2024 - HelloA11Y (11 May 2024)Samir Dash
 
How to Check GPS Location with a Live Tracker in Pakistan
How to Check GPS Location with a Live Tracker in PakistanHow to Check GPS Location with a Live Tracker in Pakistan
How to Check GPS Location with a Live Tracker in Pakistandanishmna97
 
JavaScript Usage Statistics 2024 - The Ultimate Guide
JavaScript Usage Statistics 2024 - The Ultimate GuideJavaScript Usage Statistics 2024 - The Ultimate Guide
JavaScript Usage Statistics 2024 - The Ultimate GuidePixlogix Infotech
 
(Explainable) Data-Centric AI: what are you explaininhg, and to whom?
(Explainable) Data-Centric AI: what are you explaininhg, and to whom?(Explainable) Data-Centric AI: what are you explaininhg, and to whom?
(Explainable) Data-Centric AI: what are you explaininhg, and to whom?Paolo Missier
 
Design Guidelines for Passkeys 2024.pptx
Design Guidelines for Passkeys 2024.pptxDesign Guidelines for Passkeys 2024.pptx
Design Guidelines for Passkeys 2024.pptxFIDO Alliance
 
ERP Contender Series: Acumatica vs. Sage Intacct
ERP Contender Series: Acumatica vs. Sage IntacctERP Contender Series: Acumatica vs. Sage Intacct
ERP Contender Series: Acumatica vs. Sage IntacctBrainSell Technologies
 
الأمن السيبراني - ما لا يسع للمستخدم جهله
الأمن السيبراني - ما لا يسع للمستخدم جهلهالأمن السيبراني - ما لا يسع للمستخدم جهله
الأمن السيبراني - ما لا يسع للمستخدم جهلهMohamed Sweelam
 
Generative AI Use Cases and Applications.pdf
Generative AI Use Cases and Applications.pdfGenerative AI Use Cases and Applications.pdf
Generative AI Use Cases and Applications.pdfalexjohnson7307
 
Top 10 CodeIgniter Development Companies
Top 10 CodeIgniter Development CompaniesTop 10 CodeIgniter Development Companies
Top 10 CodeIgniter Development CompaniesTopCSSGallery
 
Event-Driven Architecture Masterclass: Engineering a Robust, High-performance...
Event-Driven Architecture Masterclass: Engineering a Robust, High-performance...Event-Driven Architecture Masterclass: Engineering a Robust, High-performance...
Event-Driven Architecture Masterclass: Engineering a Robust, High-performance...ScyllaDB
 
UiPath manufacturing technology benefits and AI overview
UiPath manufacturing technology benefits and AI overviewUiPath manufacturing technology benefits and AI overview
UiPath manufacturing technology benefits and AI overviewDianaGray10
 
ADP Passwordless Journey Case Study.pptx
ADP Passwordless Journey Case Study.pptxADP Passwordless Journey Case Study.pptx
ADP Passwordless Journey Case Study.pptxFIDO Alliance
 

Recently uploaded (20)

AI in Action: Real World Use Cases by Anitaraj
AI in Action: Real World Use Cases by AnitarajAI in Action: Real World Use Cases by Anitaraj
AI in Action: Real World Use Cases by Anitaraj
 
Microsoft BitLocker Bypass Attack Method.pdf
Microsoft BitLocker Bypass Attack Method.pdfMicrosoft BitLocker Bypass Attack Method.pdf
Microsoft BitLocker Bypass Attack Method.pdf
 
Working together SRE & Platform Engineering
Working together SRE & Platform EngineeringWorking together SRE & Platform Engineering
Working together SRE & Platform Engineering
 
CORS (Kitworks Team Study 양다윗 발표자료 240510)
CORS (Kitworks Team Study 양다윗 발표자료 240510)CORS (Kitworks Team Study 양다윗 발표자료 240510)
CORS (Kitworks Team Study 양다윗 발표자료 240510)
 
Event-Driven Architecture Masterclass: Challenges in Stream Processing
Event-Driven Architecture Masterclass: Challenges in Stream ProcessingEvent-Driven Architecture Masterclass: Challenges in Stream Processing
Event-Driven Architecture Masterclass: Challenges in Stream Processing
 
Overview of Hyperledger Foundation
Overview of Hyperledger FoundationOverview of Hyperledger Foundation
Overview of Hyperledger Foundation
 
Tales from a Passkey Provider Progress from Awareness to Implementation.pptx
Tales from a Passkey Provider  Progress from Awareness to Implementation.pptxTales from a Passkey Provider  Progress from Awareness to Implementation.pptx
Tales from a Passkey Provider Progress from Awareness to Implementation.pptx
 
Introduction to FIDO Authentication and Passkeys.pptx
Introduction to FIDO Authentication and Passkeys.pptxIntroduction to FIDO Authentication and Passkeys.pptx
Introduction to FIDO Authentication and Passkeys.pptx
 
AI+A11Y 11MAY2024 HYDERBAD GAAD 2024 - HelloA11Y (11 May 2024)
AI+A11Y 11MAY2024 HYDERBAD GAAD 2024 - HelloA11Y (11 May 2024)AI+A11Y 11MAY2024 HYDERBAD GAAD 2024 - HelloA11Y (11 May 2024)
AI+A11Y 11MAY2024 HYDERBAD GAAD 2024 - HelloA11Y (11 May 2024)
 
How to Check GPS Location with a Live Tracker in Pakistan
How to Check GPS Location with a Live Tracker in PakistanHow to Check GPS Location with a Live Tracker in Pakistan
How to Check GPS Location with a Live Tracker in Pakistan
 
JavaScript Usage Statistics 2024 - The Ultimate Guide
JavaScript Usage Statistics 2024 - The Ultimate GuideJavaScript Usage Statistics 2024 - The Ultimate Guide
JavaScript Usage Statistics 2024 - The Ultimate Guide
 
(Explainable) Data-Centric AI: what are you explaininhg, and to whom?
(Explainable) Data-Centric AI: what are you explaininhg, and to whom?(Explainable) Data-Centric AI: what are you explaininhg, and to whom?
(Explainable) Data-Centric AI: what are you explaininhg, and to whom?
 
Design Guidelines for Passkeys 2024.pptx
Design Guidelines for Passkeys 2024.pptxDesign Guidelines for Passkeys 2024.pptx
Design Guidelines for Passkeys 2024.pptx
 
ERP Contender Series: Acumatica vs. Sage Intacct
ERP Contender Series: Acumatica vs. Sage IntacctERP Contender Series: Acumatica vs. Sage Intacct
ERP Contender Series: Acumatica vs. Sage Intacct
 
الأمن السيبراني - ما لا يسع للمستخدم جهله
الأمن السيبراني - ما لا يسع للمستخدم جهلهالأمن السيبراني - ما لا يسع للمستخدم جهله
الأمن السيبراني - ما لا يسع للمستخدم جهله
 
Generative AI Use Cases and Applications.pdf
Generative AI Use Cases and Applications.pdfGenerative AI Use Cases and Applications.pdf
Generative AI Use Cases and Applications.pdf
 
Top 10 CodeIgniter Development Companies
Top 10 CodeIgniter Development CompaniesTop 10 CodeIgniter Development Companies
Top 10 CodeIgniter Development Companies
 
Event-Driven Architecture Masterclass: Engineering a Robust, High-performance...
Event-Driven Architecture Masterclass: Engineering a Robust, High-performance...Event-Driven Architecture Masterclass: Engineering a Robust, High-performance...
Event-Driven Architecture Masterclass: Engineering a Robust, High-performance...
 
UiPath manufacturing technology benefits and AI overview
UiPath manufacturing technology benefits and AI overviewUiPath manufacturing technology benefits and AI overview
UiPath manufacturing technology benefits and AI overview
 
ADP Passwordless Journey Case Study.pptx
ADP Passwordless Journey Case Study.pptxADP Passwordless Journey Case Study.pptx
ADP Passwordless Journey Case Study.pptx
 

[Paper Review] DRAGNN