SlideShare a Scribd company logo
1 of 19
MULTILAYER FEDERATED
LEARNING
Kalp Pawar(BT18CSE0037)
Rajat Adkine (BT18CSE071)
Apurv Chandel (BT18CSE100)
Under the guidance of:
Dr. Tausif Diwan
Assistant Professor
Indian Institute of Information Technology, Nagpur
Final Year Project
for
Indian Institute of Information Technology, Nagpur
Table of Contents
▪ Introduction
▪ Problem Statement
▪ Proposed Methodology & Algorithm
▪ Result and Conclusion
▪ Thesis Status
▪ Future Scope
INTRODUCTION
⮚ Federated Learning
⮚Cross Device
⮚Cross Silo
⮚ Data and model
⮚ Weights Aggregation and Updates
FEDERATED
LEARNING
Federated learning is a machine learning technique
that trains an algorithm across multiple
decentralized edge devices or servers holding local
data samples, without exchanging them.
Applications
• Google makes extensive use of federated
learning in the Gboard mobile keyboard, as
well as in features on Pixel phones and in
Android Messages.
• Apple is using cross-device FL in iOS 13,
for applications like the QuickType
keyboard and the vocal classifier for “Hey
Siri”.
• doc.ai is developing cross-device FL
solutions for medical research, and Snips
has explored cross-device FL for hotword
detection.
Dataset and Model
We used MNIST dataset in order to test our model where it contains 42,000 images of
handwritten digits which are divided into different folders namely their respective digits.
Dataset and Model
Model:"sequential"
____________________________________________________________________
Layer (type) Output Shape # Param
===================================================================
dense (Dense) (None,200) 157000
___________________________________________________________________
activation (Activation) (None,200) 0
___________________________________________________________________
dense_1 (Dense) (None,200) 40200
___________________________________________________________________
activation_1 (Activation) (None,200) 0
___________________________________________________________________
dense_2 (Dense) (None,10) 2010
___________________________________________________________________
activation_2 (Activation) (None,10) 0
===================================================================
Total params: 199,210
Trainable params: 199,210
Non-trainableparams:0
_______________________________________________________
Federate Architecture
• Global model is shared with
multiple clients.
• Those clients perform training of
machine learning algorithm.
• The local result is aggregated by
using algorithms like FedAvg
,FedProx etc.
CHALLENGES IN FEDERATED LEARNING
• Privacy is first-order concern in FL, even if the experiments are simulations
running on a single machine using public data.
• Efficient aggregation of the local outputs of the models which leads to better
model prediction.
Federated Learning
▪ Cross Device:
▪ This is federated learning between organizations and devices.
▪ Learning is done remotely Update the central model using the appropriate federation technique.
▪ working on a large scale and may require millions of equipment to work for the federation.
▪ The device can go offline.
▪ Cross Silo:
▪ It is a multi-institutional and cross-functional affiliate learning.
▪ Data is partitioned into silos, each has an associated coach.
▪ data is more widely distributed, such as between hospitals, banks, presumably distributed
wearable consumer data aggregated across various fitness apps companies.
Federated Learning
▪ Difference:
▪ Cross-device focuses on mobile and IT devices and cross-silo focuses on devices in
organization but neither of them focuses on both.
▪ Cross-device can support more number of clients but model may take days (1-10
days) to converge.
▪ Cross-silo is relatively more reliable than cross-device federated learning.
▪ Cross-device provides better privacy but may result in decreased efficiency.
PROBLEM STATEMENT
● Models available today are mostly focused on efficiency or functionality, but very few are
focused on privacy.
● One such model that gives enough privacy is cross-device federated learning which is not
very efficient as compared to other models.
● We propose a model which uses the advantages of both cross-device and cross-silo federated
learning for improving privacy and efficiency.
Proposed Methodology
Client selection: The server chooses among the clients fitting the
eligibility criteria. For example, client devices that are plugged in to
the charger and having good connection speed will be eligible for
training our model.
Broadcast: The clients then receive training models and weights
from the server (e.g., TensorFlow graph).
Client computation: Clients then perform local computations and
trains to update the model (e.g., SGD in Federated Averaging).
Aggregation: The server then aggregates the updates by clients.
Various methods of aggregation like FedAvg, FedNova etc. can be
applied for accuracy and encryption and differential privacy can be
applied for increasing privacy.
Model update: The server locally updates the shared model based
on the aggregated update computed from the clients that
participated in the current round.
Algorithm
Server
▪ Result: Creates task for clients and distributes
them to managing nodes
1. for each round t = 0.1.2. do
2. S < - (sample a random set of nodes)
3. end
4. for each node k = S, in managing nodes do
5. Pass data to managing nodes
6. end
Management Node
▪ Result: Returns the accumulated data from its group of
clients
1. S <- (sample a random set of clients)
2. for each client k = S, in Graph (decentralized nodes) do
3. updateWeightUsingFedAvg (ClientTraining (k, W1))
4. end
5. return Wo to server
Algorithm
Client device
▪ Result: Returns the updated weights
1. for each local epoch with i = 1,2,... do
2. for batch B do
3. do processing using broadcasted data
4. end
5. end
6. return W to server
Result and Conclusion
The accuracy of the model is intact, the model is more robust and error tolerant and we have also introduced more privacy.
Further improvements could be made with the help of Fine-Grained training models that uses blockchains for additional
security and monitoring perspective.
Result and Conclusion
▪ More work is required to find out its accuracy as compared with different situations using
different models.
▪ We were able to experiment and confirm that this setting significantly reduces the number of
mishaps occurring due to network issues and uncertain available time of client and reduces
overhead of getting the correct kind of data by selectively filtering the devices on non-PII data.
▪ Moving forward, we need to analyze the performance of this approach in real world datasets as
well as compare various other causal learning approaches.
References
▪ https://paperswithcode.com/paper/flexible-clustered-federated-learning-for.
▪ https://paperswithcode.com/paper/federated-learning-for-mobile-keyboard.
▪ https://paperswithcode.com/paper/advances-and-open-problems-in-federated.
▪ https://paperswithcode.com/paper/differentially-private-federated-deep.
▪ https://paperswithcode.com/paper/model-contrastive-federated-learning
▪ https://www.researchgate.net/publication/339297228_Towards_Blockchain-
Based_Reputation-Aware_Federated_Learning/download
Multi Layer Federated Learning.pptx

More Related Content

Similar to Multi Layer Federated Learning.pptx

Toward Distributed, Global, Deep Learning Using IoT Devices
Toward Distributed, Global, Deep Learning Using IoT DevicesToward Distributed, Global, Deep Learning Using IoT Devices
Toward Distributed, Global, Deep Learning Using IoT DevicesBharath Sudharsan
 
Knowledge Distillation for Federated Learning: a Practical Guide
Knowledge Distillation for Federated Learning: a Practical GuideKnowledge Distillation for Federated Learning: a Practical Guide
Knowledge Distillation for Federated Learning: a Practical GuideXiachongFeng
 
Ieee projects-2014-bulk-ieee-projects-2015-title-list-for-me-be-mphil-final-y...
Ieee projects-2014-bulk-ieee-projects-2015-title-list-for-me-be-mphil-final-y...Ieee projects-2014-bulk-ieee-projects-2015-title-list-for-me-be-mphil-final-y...
Ieee projects-2014-bulk-ieee-projects-2015-title-list-for-me-be-mphil-final-y...birdsking
 
Android based Attendance and examination automation
Android based Attendance and examination automationAndroid based Attendance and examination automation
Android based Attendance and examination automationRitika Mahajan
 
IEEE Projects 2013 For ME Cse Seabirds ( Trichy, Thanjavur, Karur, Perambalur )
IEEE Projects 2013 For ME Cse Seabirds ( Trichy, Thanjavur, Karur, Perambalur )IEEE Projects 2013 For ME Cse Seabirds ( Trichy, Thanjavur, Karur, Perambalur )
IEEE Projects 2013 For ME Cse Seabirds ( Trichy, Thanjavur, Karur, Perambalur )SBGC
 
Bulk Ieee Projects 2013 @ Seabirds ( Chennai, Trichy, Hyderabad, Pune, Mumbai )
Bulk Ieee Projects 2013 @ Seabirds ( Chennai, Trichy, Hyderabad, Pune, Mumbai )Bulk Ieee Projects 2013 @ Seabirds ( Chennai, Trichy, Hyderabad, Pune, Mumbai )
Bulk Ieee Projects 2013 @ Seabirds ( Chennai, Trichy, Hyderabad, Pune, Mumbai )SBGC
 
Ieee projects-2013-2014-title-list-for-me-be-mphil-final-year-students
Ieee projects-2013-2014-title-list-for-me-be-mphil-final-year-studentsIeee projects-2013-2014-title-list-for-me-be-mphil-final-year-students
Ieee projects-2013-2014-title-list-for-me-be-mphil-final-year-studentsPruthivi Rajan
 
Ieee Projects 2013 for Cse @ Seabirds(Trichy, Pudukkottai, Perambalur, Thanja...
Ieee Projects 2013 for Cse @ Seabirds(Trichy, Pudukkottai, Perambalur, Thanja...Ieee Projects 2013 for Cse @ Seabirds(Trichy, Pudukkottai, Perambalur, Thanja...
Ieee Projects 2013 for Cse @ Seabirds(Trichy, Pudukkottai, Perambalur, Thanja...SBGC
 
Machine Learning on mobile devices
Machine Learning on mobile devicesMachine Learning on mobile devices
Machine Learning on mobile devicesSergey Burkov
 
2019-09-05Federated Learning.pdf
2019-09-05Federated Learning.pdf2019-09-05Federated Learning.pdf
2019-09-05Federated Learning.pdfjimjones227147
 
Presentation 7.pptx
Presentation 7.pptxPresentation 7.pptx
Presentation 7.pptxShivam327815
 
IRJET- An Efficient Data Replication in Salesforce Cloud Environment
IRJET-  	  An Efficient Data Replication in Salesforce Cloud EnvironmentIRJET-  	  An Efficient Data Replication in Salesforce Cloud Environment
IRJET- An Efficient Data Replication in Salesforce Cloud EnvironmentIRJET Journal
 
IEEE Projects 2013 For ME Cse @ Seabirds ( Trichy, Thanjavur, Perambalur, Di...
IEEE Projects 2013 For ME Cse @  Seabirds ( Trichy, Thanjavur, Perambalur, Di...IEEE Projects 2013 For ME Cse @  Seabirds ( Trichy, Thanjavur, Perambalur, Di...
IEEE Projects 2013 For ME Cse @ Seabirds ( Trichy, Thanjavur, Perambalur, Di...SBGC
 
Machine Learning for Capacity Management
 Machine Learning for Capacity Management Machine Learning for Capacity Management
Machine Learning for Capacity ManagementEDB
 
Enabling on-device learning at scale
Enabling on-device learning at scaleEnabling on-device learning at scale
Enabling on-device learning at scaleQualcomm Research
 
Webinar CRUI Dell: flexilab, computer classroom made flexible
Webinar CRUI Dell: flexilab, computer classroom made flexible Webinar CRUI Dell: flexilab, computer classroom made flexible
Webinar CRUI Dell: flexilab, computer classroom made flexible Jürgen Ambrosi
 
IRJET - Secure Data Sharing in Cloud Computing using Revocable Storage Id...
IRJET -  	  Secure Data Sharing in Cloud Computing using Revocable Storage Id...IRJET -  	  Secure Data Sharing in Cloud Computing using Revocable Storage Id...
IRJET - Secure Data Sharing in Cloud Computing using Revocable Storage Id...IRJET Journal
 

Similar to Multi Layer Federated Learning.pptx (20)

Toward Distributed, Global, Deep Learning Using IoT Devices
Toward Distributed, Global, Deep Learning Using IoT DevicesToward Distributed, Global, Deep Learning Using IoT Devices
Toward Distributed, Global, Deep Learning Using IoT Devices
 
Jithender_3+Years_Exp_ETL Testing
Jithender_3+Years_Exp_ETL TestingJithender_3+Years_Exp_ETL Testing
Jithender_3+Years_Exp_ETL Testing
 
Knowledge Distillation for Federated Learning: a Practical Guide
Knowledge Distillation for Federated Learning: a Practical GuideKnowledge Distillation for Federated Learning: a Practical Guide
Knowledge Distillation for Federated Learning: a Practical Guide
 
Ieee projects-2014-bulk-ieee-projects-2015-title-list-for-me-be-mphil-final-y...
Ieee projects-2014-bulk-ieee-projects-2015-title-list-for-me-be-mphil-final-y...Ieee projects-2014-bulk-ieee-projects-2015-title-list-for-me-be-mphil-final-y...
Ieee projects-2014-bulk-ieee-projects-2015-title-list-for-me-be-mphil-final-y...
 
Adopting the Cloud
Adopting the CloudAdopting the Cloud
Adopting the Cloud
 
Android based Attendance and examination automation
Android based Attendance and examination automationAndroid based Attendance and examination automation
Android based Attendance and examination automation
 
IEEE Projects 2013 For ME Cse Seabirds ( Trichy, Thanjavur, Karur, Perambalur )
IEEE Projects 2013 For ME Cse Seabirds ( Trichy, Thanjavur, Karur, Perambalur )IEEE Projects 2013 For ME Cse Seabirds ( Trichy, Thanjavur, Karur, Perambalur )
IEEE Projects 2013 For ME Cse Seabirds ( Trichy, Thanjavur, Karur, Perambalur )
 
Bulk Ieee Projects 2013 @ Seabirds ( Chennai, Trichy, Hyderabad, Pune, Mumbai )
Bulk Ieee Projects 2013 @ Seabirds ( Chennai, Trichy, Hyderabad, Pune, Mumbai )Bulk Ieee Projects 2013 @ Seabirds ( Chennai, Trichy, Hyderabad, Pune, Mumbai )
Bulk Ieee Projects 2013 @ Seabirds ( Chennai, Trichy, Hyderabad, Pune, Mumbai )
 
Ieee projects-2013-2014-title-list-for-me-be-mphil-final-year-students
Ieee projects-2013-2014-title-list-for-me-be-mphil-final-year-studentsIeee projects-2013-2014-title-list-for-me-be-mphil-final-year-students
Ieee projects-2013-2014-title-list-for-me-be-mphil-final-year-students
 
Ieee Projects 2013 for Cse @ Seabirds(Trichy, Pudukkottai, Perambalur, Thanja...
Ieee Projects 2013 for Cse @ Seabirds(Trichy, Pudukkottai, Perambalur, Thanja...Ieee Projects 2013 for Cse @ Seabirds(Trichy, Pudukkottai, Perambalur, Thanja...
Ieee Projects 2013 for Cse @ Seabirds(Trichy, Pudukkottai, Perambalur, Thanja...
 
Machine Learning on mobile devices
Machine Learning on mobile devicesMachine Learning on mobile devices
Machine Learning on mobile devices
 
2019-09-05Federated Learning.pdf
2019-09-05Federated Learning.pdf2019-09-05Federated Learning.pdf
2019-09-05Federated Learning.pdf
 
Presentation 7.pptx
Presentation 7.pptxPresentation 7.pptx
Presentation 7.pptx
 
IRJET- An Efficient Data Replication in Salesforce Cloud Environment
IRJET-  	  An Efficient Data Replication in Salesforce Cloud EnvironmentIRJET-  	  An Efficient Data Replication in Salesforce Cloud Environment
IRJET- An Efficient Data Replication in Salesforce Cloud Environment
 
IEEE Projects 2013 For ME Cse @ Seabirds ( Trichy, Thanjavur, Perambalur, Di...
IEEE Projects 2013 For ME Cse @  Seabirds ( Trichy, Thanjavur, Perambalur, Di...IEEE Projects 2013 For ME Cse @  Seabirds ( Trichy, Thanjavur, Perambalur, Di...
IEEE Projects 2013 For ME Cse @ Seabirds ( Trichy, Thanjavur, Perambalur, Di...
 
Machine Learning for Capacity Management
 Machine Learning for Capacity Management Machine Learning for Capacity Management
Machine Learning for Capacity Management
 
Enabling on-device learning at scale
Enabling on-device learning at scaleEnabling on-device learning at scale
Enabling on-device learning at scale
 
Webinar CRUI Dell: flexilab, computer classroom made flexible
Webinar CRUI Dell: flexilab, computer classroom made flexible Webinar CRUI Dell: flexilab, computer classroom made flexible
Webinar CRUI Dell: flexilab, computer classroom made flexible
 
IRJET - Secure Data Sharing in Cloud Computing using Revocable Storage Id...
IRJET -  	  Secure Data Sharing in Cloud Computing using Revocable Storage Id...IRJET -  	  Secure Data Sharing in Cloud Computing using Revocable Storage Id...
IRJET - Secure Data Sharing in Cloud Computing using Revocable Storage Id...
 
Presentation_final.pdf
Presentation_final.pdfPresentation_final.pdf
Presentation_final.pdf
 

Recently uploaded

Contributi dei parlamentari del PD - Contributi L. 3/2019
Contributi dei parlamentari del PD - Contributi L. 3/2019Contributi dei parlamentari del PD - Contributi L. 3/2019
Contributi dei parlamentari del PD - Contributi L. 3/2019Partito democratico
 
World Press Freedom Day 2024; May 3rd - Poster
World Press Freedom Day 2024; May 3rd - PosterWorld Press Freedom Day 2024; May 3rd - Poster
World Press Freedom Day 2024; May 3rd - PosterChristina Parmionova
 
Election 2024 Presiding Duty Keypoints_01.pdf
Election 2024 Presiding Duty Keypoints_01.pdfElection 2024 Presiding Duty Keypoints_01.pdf
Election 2024 Presiding Duty Keypoints_01.pdfSamirsinh Parmar
 
Just Call VIP Call Girls In Bangalore Kr Puram ☎️ 6378878445 Independent Fem...
Just Call VIP Call Girls In  Bangalore Kr Puram ☎️ 6378878445 Independent Fem...Just Call VIP Call Girls In  Bangalore Kr Puram ☎️ 6378878445 Independent Fem...
Just Call VIP Call Girls In Bangalore Kr Puram ☎️ 6378878445 Independent Fem...HyderabadDolls
 
Time, Stress & Work Life Balance for Clerks with Beckie Whitehouse
Time, Stress & Work Life Balance for Clerks with Beckie WhitehouseTime, Stress & Work Life Balance for Clerks with Beckie Whitehouse
Time, Stress & Work Life Balance for Clerks with Beckie Whitehousesubs7
 
2024 asthma jkdjkfjsdklfjsdlkfjskldfgdsgerg
2024 asthma jkdjkfjsdklfjsdlkfjskldfgdsgerg2024 asthma jkdjkfjsdklfjsdlkfjskldfgdsgerg
2024 asthma jkdjkfjsdklfjsdlkfjskldfgdsgergMadhuKothuru
 
3 May, Journalism in the face of the Environmental Crisis.
3 May, Journalism in the face of the Environmental Crisis.3 May, Journalism in the face of the Environmental Crisis.
3 May, Journalism in the face of the Environmental Crisis.Christina Parmionova
 
An Atoll Futures Research Institute? Presentation for CANCC
An Atoll Futures Research Institute? Presentation for CANCCAn Atoll Futures Research Institute? Presentation for CANCC
An Atoll Futures Research Institute? Presentation for CANCCNAP Global Network
 
1935 CONSTITUTION REPORT IN RIPH FINALLS
1935 CONSTITUTION REPORT IN RIPH FINALLS1935 CONSTITUTION REPORT IN RIPH FINALLS
1935 CONSTITUTION REPORT IN RIPH FINALLSarandianics
 
Cara Gugurkan Pembuahan Secara Alami Dan Cepat ABORSI KANDUNGAN 087776558899
Cara Gugurkan Pembuahan Secara Alami Dan Cepat ABORSI KANDUNGAN 087776558899Cara Gugurkan Pembuahan Secara Alami Dan Cepat ABORSI KANDUNGAN 087776558899
Cara Gugurkan Pembuahan Secara Alami Dan Cepat ABORSI KANDUNGAN 087776558899Cara Menggugurkan Kandungan 087776558899
 
Top profile Call Girls In Haldia [ 7014168258 ] Call Me For Genuine Models We...
Top profile Call Girls In Haldia [ 7014168258 ] Call Me For Genuine Models We...Top profile Call Girls In Haldia [ 7014168258 ] Call Me For Genuine Models We...
Top profile Call Girls In Haldia [ 7014168258 ] Call Me For Genuine Models We...gajnagarg
 
Competitive Advantage slide deck___.pptx
Competitive Advantage slide deck___.pptxCompetitive Advantage slide deck___.pptx
Competitive Advantage slide deck___.pptxScottMeyers35
 
Dating Call Girls inBaloda Bazar Bhatapara 9332606886Call Girls Advance Cash...
Dating Call Girls inBaloda Bazar Bhatapara  9332606886Call Girls Advance Cash...Dating Call Girls inBaloda Bazar Bhatapara  9332606886Call Girls Advance Cash...
Dating Call Girls inBaloda Bazar Bhatapara 9332606886Call Girls Advance Cash...kumargunjan9515
 
2024 UN Civil Society Conference in Support of the Summit of the Future.
2024 UN Civil Society Conference in Support of the Summit of the Future.2024 UN Civil Society Conference in Support of the Summit of the Future.
2024 UN Civil Society Conference in Support of the Summit of the Future.Christina Parmionova
 
Call Girls in Moti Bagh (delhi) call me [8448380779] escort service 24X7
Call Girls in Moti Bagh (delhi) call me [8448380779] escort service 24X7Call Girls in Moti Bagh (delhi) call me [8448380779] escort service 24X7
Call Girls in Moti Bagh (delhi) call me [8448380779] escort service 24X7Delhi Call girls
 
31st World Press Freedom Day Conference in Santiago.
31st World Press Freedom Day Conference in Santiago.31st World Press Freedom Day Conference in Santiago.
31st World Press Freedom Day Conference in Santiago.Christina Parmionova
 
31st World Press Freedom Day - A Press for the Planet: Journalism in the face...
31st World Press Freedom Day - A Press for the Planet: Journalism in the face...31st World Press Freedom Day - A Press for the Planet: Journalism in the face...
31st World Press Freedom Day - A Press for the Planet: Journalism in the face...Christina Parmionova
 
Honasa Consumer Limited Impact Report 2024.pdf
Honasa Consumer Limited Impact Report 2024.pdfHonasa Consumer Limited Impact Report 2024.pdf
Honasa Consumer Limited Impact Report 2024.pdfSocial Samosa
 

Recently uploaded (20)

Contributi dei parlamentari del PD - Contributi L. 3/2019
Contributi dei parlamentari del PD - Contributi L. 3/2019Contributi dei parlamentari del PD - Contributi L. 3/2019
Contributi dei parlamentari del PD - Contributi L. 3/2019
 
Panchayath circular KLC -Panchayath raj act s 169, 218
Panchayath circular KLC -Panchayath raj act s 169, 218Panchayath circular KLC -Panchayath raj act s 169, 218
Panchayath circular KLC -Panchayath raj act s 169, 218
 
World Press Freedom Day 2024; May 3rd - Poster
World Press Freedom Day 2024; May 3rd - PosterWorld Press Freedom Day 2024; May 3rd - Poster
World Press Freedom Day 2024; May 3rd - Poster
 
Election 2024 Presiding Duty Keypoints_01.pdf
Election 2024 Presiding Duty Keypoints_01.pdfElection 2024 Presiding Duty Keypoints_01.pdf
Election 2024 Presiding Duty Keypoints_01.pdf
 
Just Call VIP Call Girls In Bangalore Kr Puram ☎️ 6378878445 Independent Fem...
Just Call VIP Call Girls In  Bangalore Kr Puram ☎️ 6378878445 Independent Fem...Just Call VIP Call Girls In  Bangalore Kr Puram ☎️ 6378878445 Independent Fem...
Just Call VIP Call Girls In Bangalore Kr Puram ☎️ 6378878445 Independent Fem...
 
Time, Stress & Work Life Balance for Clerks with Beckie Whitehouse
Time, Stress & Work Life Balance for Clerks with Beckie WhitehouseTime, Stress & Work Life Balance for Clerks with Beckie Whitehouse
Time, Stress & Work Life Balance for Clerks with Beckie Whitehouse
 
2024 asthma jkdjkfjsdklfjsdlkfjskldfgdsgerg
2024 asthma jkdjkfjsdklfjsdlkfjskldfgdsgerg2024 asthma jkdjkfjsdklfjsdlkfjskldfgdsgerg
2024 asthma jkdjkfjsdklfjsdlkfjskldfgdsgerg
 
3 May, Journalism in the face of the Environmental Crisis.
3 May, Journalism in the face of the Environmental Crisis.3 May, Journalism in the face of the Environmental Crisis.
3 May, Journalism in the face of the Environmental Crisis.
 
An Atoll Futures Research Institute? Presentation for CANCC
An Atoll Futures Research Institute? Presentation for CANCCAn Atoll Futures Research Institute? Presentation for CANCC
An Atoll Futures Research Institute? Presentation for CANCC
 
1935 CONSTITUTION REPORT IN RIPH FINALLS
1935 CONSTITUTION REPORT IN RIPH FINALLS1935 CONSTITUTION REPORT IN RIPH FINALLS
1935 CONSTITUTION REPORT IN RIPH FINALLS
 
Cara Gugurkan Pembuahan Secara Alami Dan Cepat ABORSI KANDUNGAN 087776558899
Cara Gugurkan Pembuahan Secara Alami Dan Cepat ABORSI KANDUNGAN 087776558899Cara Gugurkan Pembuahan Secara Alami Dan Cepat ABORSI KANDUNGAN 087776558899
Cara Gugurkan Pembuahan Secara Alami Dan Cepat ABORSI KANDUNGAN 087776558899
 
Top profile Call Girls In Haldia [ 7014168258 ] Call Me For Genuine Models We...
Top profile Call Girls In Haldia [ 7014168258 ] Call Me For Genuine Models We...Top profile Call Girls In Haldia [ 7014168258 ] Call Me For Genuine Models We...
Top profile Call Girls In Haldia [ 7014168258 ] Call Me For Genuine Models We...
 
Competitive Advantage slide deck___.pptx
Competitive Advantage slide deck___.pptxCompetitive Advantage slide deck___.pptx
Competitive Advantage slide deck___.pptx
 
Dating Call Girls inBaloda Bazar Bhatapara 9332606886Call Girls Advance Cash...
Dating Call Girls inBaloda Bazar Bhatapara  9332606886Call Girls Advance Cash...Dating Call Girls inBaloda Bazar Bhatapara  9332606886Call Girls Advance Cash...
Dating Call Girls inBaloda Bazar Bhatapara 9332606886Call Girls Advance Cash...
 
2024 UN Civil Society Conference in Support of the Summit of the Future.
2024 UN Civil Society Conference in Support of the Summit of the Future.2024 UN Civil Society Conference in Support of the Summit of the Future.
2024 UN Civil Society Conference in Support of the Summit of the Future.
 
AHMR volume 10 number 1 January-April 2024
AHMR volume 10 number 1 January-April 2024AHMR volume 10 number 1 January-April 2024
AHMR volume 10 number 1 January-April 2024
 
Call Girls in Moti Bagh (delhi) call me [8448380779] escort service 24X7
Call Girls in Moti Bagh (delhi) call me [8448380779] escort service 24X7Call Girls in Moti Bagh (delhi) call me [8448380779] escort service 24X7
Call Girls in Moti Bagh (delhi) call me [8448380779] escort service 24X7
 
31st World Press Freedom Day Conference in Santiago.
31st World Press Freedom Day Conference in Santiago.31st World Press Freedom Day Conference in Santiago.
31st World Press Freedom Day Conference in Santiago.
 
31st World Press Freedom Day - A Press for the Planet: Journalism in the face...
31st World Press Freedom Day - A Press for the Planet: Journalism in the face...31st World Press Freedom Day - A Press for the Planet: Journalism in the face...
31st World Press Freedom Day - A Press for the Planet: Journalism in the face...
 
Honasa Consumer Limited Impact Report 2024.pdf
Honasa Consumer Limited Impact Report 2024.pdfHonasa Consumer Limited Impact Report 2024.pdf
Honasa Consumer Limited Impact Report 2024.pdf
 

Multi Layer Federated Learning.pptx

  • 1. MULTILAYER FEDERATED LEARNING Kalp Pawar(BT18CSE0037) Rajat Adkine (BT18CSE071) Apurv Chandel (BT18CSE100) Under the guidance of: Dr. Tausif Diwan Assistant Professor Indian Institute of Information Technology, Nagpur Final Year Project for Indian Institute of Information Technology, Nagpur
  • 2. Table of Contents ▪ Introduction ▪ Problem Statement ▪ Proposed Methodology & Algorithm ▪ Result and Conclusion ▪ Thesis Status ▪ Future Scope
  • 3. INTRODUCTION ⮚ Federated Learning ⮚Cross Device ⮚Cross Silo ⮚ Data and model ⮚ Weights Aggregation and Updates
  • 4. FEDERATED LEARNING Federated learning is a machine learning technique that trains an algorithm across multiple decentralized edge devices or servers holding local data samples, without exchanging them.
  • 5. Applications • Google makes extensive use of federated learning in the Gboard mobile keyboard, as well as in features on Pixel phones and in Android Messages. • Apple is using cross-device FL in iOS 13, for applications like the QuickType keyboard and the vocal classifier for “Hey Siri”. • doc.ai is developing cross-device FL solutions for medical research, and Snips has explored cross-device FL for hotword detection.
  • 6. Dataset and Model We used MNIST dataset in order to test our model where it contains 42,000 images of handwritten digits which are divided into different folders namely their respective digits.
  • 7. Dataset and Model Model:"sequential" ____________________________________________________________________ Layer (type) Output Shape # Param =================================================================== dense (Dense) (None,200) 157000 ___________________________________________________________________ activation (Activation) (None,200) 0 ___________________________________________________________________ dense_1 (Dense) (None,200) 40200 ___________________________________________________________________ activation_1 (Activation) (None,200) 0 ___________________________________________________________________ dense_2 (Dense) (None,10) 2010 ___________________________________________________________________ activation_2 (Activation) (None,10) 0 =================================================================== Total params: 199,210 Trainable params: 199,210 Non-trainableparams:0 _______________________________________________________
  • 8. Federate Architecture • Global model is shared with multiple clients. • Those clients perform training of machine learning algorithm. • The local result is aggregated by using algorithms like FedAvg ,FedProx etc.
  • 9. CHALLENGES IN FEDERATED LEARNING • Privacy is first-order concern in FL, even if the experiments are simulations running on a single machine using public data. • Efficient aggregation of the local outputs of the models which leads to better model prediction.
  • 10. Federated Learning ▪ Cross Device: ▪ This is federated learning between organizations and devices. ▪ Learning is done remotely Update the central model using the appropriate federation technique. ▪ working on a large scale and may require millions of equipment to work for the federation. ▪ The device can go offline. ▪ Cross Silo: ▪ It is a multi-institutional and cross-functional affiliate learning. ▪ Data is partitioned into silos, each has an associated coach. ▪ data is more widely distributed, such as between hospitals, banks, presumably distributed wearable consumer data aggregated across various fitness apps companies.
  • 11. Federated Learning ▪ Difference: ▪ Cross-device focuses on mobile and IT devices and cross-silo focuses on devices in organization but neither of them focuses on both. ▪ Cross-device can support more number of clients but model may take days (1-10 days) to converge. ▪ Cross-silo is relatively more reliable than cross-device federated learning. ▪ Cross-device provides better privacy but may result in decreased efficiency.
  • 12. PROBLEM STATEMENT ● Models available today are mostly focused on efficiency or functionality, but very few are focused on privacy. ● One such model that gives enough privacy is cross-device federated learning which is not very efficient as compared to other models. ● We propose a model which uses the advantages of both cross-device and cross-silo federated learning for improving privacy and efficiency.
  • 13. Proposed Methodology Client selection: The server chooses among the clients fitting the eligibility criteria. For example, client devices that are plugged in to the charger and having good connection speed will be eligible for training our model. Broadcast: The clients then receive training models and weights from the server (e.g., TensorFlow graph). Client computation: Clients then perform local computations and trains to update the model (e.g., SGD in Federated Averaging). Aggregation: The server then aggregates the updates by clients. Various methods of aggregation like FedAvg, FedNova etc. can be applied for accuracy and encryption and differential privacy can be applied for increasing privacy. Model update: The server locally updates the shared model based on the aggregated update computed from the clients that participated in the current round.
  • 14. Algorithm Server ▪ Result: Creates task for clients and distributes them to managing nodes 1. for each round t = 0.1.2. do 2. S < - (sample a random set of nodes) 3. end 4. for each node k = S, in managing nodes do 5. Pass data to managing nodes 6. end Management Node ▪ Result: Returns the accumulated data from its group of clients 1. S <- (sample a random set of clients) 2. for each client k = S, in Graph (decentralized nodes) do 3. updateWeightUsingFedAvg (ClientTraining (k, W1)) 4. end 5. return Wo to server
  • 15. Algorithm Client device ▪ Result: Returns the updated weights 1. for each local epoch with i = 1,2,... do 2. for batch B do 3. do processing using broadcasted data 4. end 5. end 6. return W to server
  • 16. Result and Conclusion The accuracy of the model is intact, the model is more robust and error tolerant and we have also introduced more privacy. Further improvements could be made with the help of Fine-Grained training models that uses blockchains for additional security and monitoring perspective.
  • 17. Result and Conclusion ▪ More work is required to find out its accuracy as compared with different situations using different models. ▪ We were able to experiment and confirm that this setting significantly reduces the number of mishaps occurring due to network issues and uncertain available time of client and reduces overhead of getting the correct kind of data by selectively filtering the devices on non-PII data. ▪ Moving forward, we need to analyze the performance of this approach in real world datasets as well as compare various other causal learning approaches.
  • 18. References ▪ https://paperswithcode.com/paper/flexible-clustered-federated-learning-for. ▪ https://paperswithcode.com/paper/federated-learning-for-mobile-keyboard. ▪ https://paperswithcode.com/paper/advances-and-open-problems-in-federated. ▪ https://paperswithcode.com/paper/differentially-private-federated-deep. ▪ https://paperswithcode.com/paper/model-contrastive-federated-learning ▪ https://www.researchgate.net/publication/339297228_Towards_Blockchain- Based_Reputation-Aware_Federated_Learning/download