SlideShare a Scribd company logo
1 of 22
FEDBN: FEDERATED LEARNING
ON NON-IID FEATURES VIA LOCAL
BATCH NORMALIZATION
Paper presentation by Anam ur rehman
Contact: anamur.rehman@studenti.polito.it
Published as a conference paper at ICLR 2021
1
Authors
2
FEDBN: Federated learning[1]
Year
2021
2020
2018
2017
2016
2015
2014
2013
2012
2011
[1] Jakub et al. Federated optimization: Distributed machine learning for on-device intelligence. 2016
Classical Machine Learning:
• Centralized data storage
• Training process computations at the central server.
What if ?
 Data stays distributed on remote devices
Devices maintain control of their own data
 Training is done locally on remote devices
 One global model is learned via aggregation
3
Autonomous cars on an
average generate around 4 GB
of data per hour of driving.
FEDBN: Federated learning
Applications [1]
• Transportation: self-driving cars
• Healthcare: predictions on patient data
• Cybersecurity: spam filtering
• Smart applications: voice recognition, next word prediction, etc.
[1] Read more: Priyanka et al. Federated Learning: Opportunities and Challenges, 2021
Challenges [1]
• Communication Overheads: presence of stragglers
• Heterogeneity: system, statistical (in contrast to distributed learning)
• Privacy concerns
4
Year
2021
2020
2018
2017
2016
2015
2014
2013
2012
2011
FEDBN: Federated learning
Statistical heterogeneity among local datasets:
• Unbalancedness: Clients may have different amount of data
 Example, Spam filter for emails.
• Covariate shift: Statistical distribution of data varies among clients
 Example: NLP digits recognition
• Concept shift: Same features may correspond to different labels for different clients
 Example, in NLP, Sentiment analysis on same text may yield different sentiments for different clients
5
FEDBN: Example of NonIID datasets
MNIST
MNIST-M
USPS SynthDigits
SVHN
Covariate shift among datasets
Digits
dataset
6
FEDBN: Related work
• FedAvg[1]: Federated Average
[1] Brendan McMahan et al. Communication-efficient learning of deep networks from decentralized data. 2017.
At each communication round
1. Server randomly selects a subset of K clients and Send them current global model
2. Selected device k updates this model on local client data via SGD. After training client
sends the new local model back to server
3. Server aggregates local models to form a new global model
- Convergence in not guranteed. In hetergeneous settings it can diverge [1]
Year
2021
2020
2018
2017
2016
2015
2014
2013
2012
2011
7
FEDBN: Related work
• FedProx[1]: Federated Optimization in Heterogeneous Networks
[1] Tian Li et al, In Conference on Machine Learning and Systems, 2020a, 2020b.
Slide credit: Tian Li, MLSys presentation.
+ Limits the impact of heterogeneous local updates
+ Safely incorporate partial work of stragglers
+ Generalization of FedAvg; Allows for any local solver
+ Theoretical guarantees for convergence
Year
2021
2020
2018
2017
2016
2015
2014
2013
2012
2011
8
FEDBN: Related work
• SiloBN[1]: Siloed Federated Learning for Multi-Centric Histopathology Datasets
[1] Mathieu Andreux et al, Siloed federated learning for multi-centric histopathology datasets, pp. 129–139. Springer, 2020.
Slide credits: [1]
Year
2021
2020
2018
2017
2016
2015
2014
2013
2012
2011
9
FEDBN: Batch Normalization
Year
2021
2020
2018
2017
2016
2015
2014
2013
2012
2011
[1] Sergey Ioffe et al. Batch normalization: Accelerating deep network training by reducing internal covariate shift. 2015
γ and β are the only
learnable parameters
of BN layer.
10
Why we use it ?
To reduce internal covariate
shift in neural network [1].
How it works ?
11
2021
FEDBN: Problem with non IID data
12
Consider a simple,
non-convex learning problem:
s.t
ϵ ∼ 𝒩 0, σ2
𝑤2
∗
Two clients train a model
s.t
x1 ∼ 𝒩 μ, σ1
2
.
x2 ∼ 𝒩 μ, σ2
2
.
and
σ1
2
≠ σ2
2
local squared loss before and after local BN
𝑤1
∗
w
𝑓𝑤 𝑥𝑖 = 𝑐𝑜𝑠 𝑤𝑥𝑖
FEDBN: Why not just take the average? (SiloBN)
Client 1
w1
∗
: Optimal weight
γ1
∗
: Optimal BN parameter
Observation 1:
For a fixed optimal weight w1
∗
,
changing γ deteriorates the model
quality.
Observation 2:
For a given optimal BN
parameter γ1
∗
, changing w
deteriorates the quality.
13
FEDBN: How it works?
Local training
Global aggregation
14
FEDBN: Pytorch implementation
15
FEDBN: How it Really Works?
Source: med-air/FedBN (github.com)
16
Global
Aggregation
FEDBN: Results on digit dataset (FedAvg vs FedBN)
• Outperforms FedAvg on SVHN dataset
• Faster convergence
• Smooth and robust convergence
17
FEDBN: Results; what if
Communication is done
at different frequencies?
18
FEDBN: Results; what if
Dataset size varies for
each client?
19
FEDBN: Contributions
Provides convergence guarantees.
Improves the convergence behavior on non-IID datasets.
One step further in privacy of client’s data.
20
FEDBN: Take home message
• Use batch normalization
• Keep it local
• Smooth and fast convergence
21
Useful links
Federated Optimization in Heterogeneous Networks
FedProx presentation by Tian Li:
22
med-air/FedBN (github.com)
Pytorch implementation of FedBN:
Brendan McMahan’s Talk:
Guarding user Privacy with Federated Learning

More Related Content

What's hot

Metamaterial based antenna with application to real life system
Metamaterial based antenna with application to real life systemMetamaterial based antenna with application to real life system
Metamaterial based antenna with application to real life systemKiran Ajetrao
 
Mutual coupling
Mutual couplingMutual coupling
Mutual couplingJagadish M
 
YAZILIM SÜREÇ İYİLEŞTİRME ÇALIŞMALARINDA YETENEK OLGUNLUK MODELİ ENTEGRASYONU...
YAZILIM SÜREÇ İYİLEŞTİRME ÇALIŞMALARINDA YETENEK OLGUNLUK MODELİ ENTEGRASYONU...YAZILIM SÜREÇ İYİLEŞTİRME ÇALIŞMALARINDA YETENEK OLGUNLUK MODELİ ENTEGRASYONU...
YAZILIM SÜREÇ İYİLEŞTİRME ÇALIŞMALARINDA YETENEK OLGUNLUK MODELİ ENTEGRASYONU...Celal Dolkan
 
DJI's Drone Solutions for Smart Cities of the Future
DJI's Drone Solutions for Smart Cities of the FutureDJI's Drone Solutions for Smart Cities of the Future
DJI's Drone Solutions for Smart Cities of the Futuresitecmy
 
MQTT - MQ Telemetry Transport for Message Queueing
MQTT - MQ Telemetry Transport for Message QueueingMQTT - MQ Telemetry Transport for Message Queueing
MQTT - MQ Telemetry Transport for Message QueueingPeter R. Egli
 
Open source technology
Open source technologyOpen source technology
Open source technologyaparnaz1
 
DSD-INT 2015 - Photogrammetric workflows and use of UA VS, Francesco nex, E-s...
DSD-INT 2015 - Photogrammetric workflows and use of UA VS, Francesco nex, E-s...DSD-INT 2015 - Photogrammetric workflows and use of UA VS, Francesco nex, E-s...
DSD-INT 2015 - Photogrammetric workflows and use of UA VS, Francesco nex, E-s...Deltares
 
Computer Networking A Top-Down Approach 6th editiion.pdf
Computer Networking A Top-Down Approach 6th editiion.pdfComputer Networking A Top-Down Approach 6th editiion.pdf
Computer Networking A Top-Down Approach 6th editiion.pdfssuser1a7f55
 
Desing of a rectangular patch antenna
Desing of a rectangular patch antennaDesing of a rectangular patch antenna
Desing of a rectangular patch antennaAysu COSKUN
 
CST_ANTENNA-CST STUDIO SUITE™ 2006B.pdf
CST_ANTENNA-CST STUDIO SUITE™ 2006B.pdfCST_ANTENNA-CST STUDIO SUITE™ 2006B.pdf
CST_ANTENNA-CST STUDIO SUITE™ 2006B.pdfEdisonAndresZapataOc
 
AI-Sustainability.pptx
AI-Sustainability.pptxAI-Sustainability.pptx
AI-Sustainability.pptxeilamtamar
 
Design and Simulation Microstrip patch Antenna using CST Microwave Studio
Design and Simulation Microstrip patch Antenna  using CST Microwave StudioDesign and Simulation Microstrip patch Antenna  using CST Microwave Studio
Design and Simulation Microstrip patch Antenna using CST Microwave StudioAymen Al-obaidi
 
Intro to wireless sensor network
Intro to wireless sensor networkIntro to wireless sensor network
Intro to wireless sensor networkVrince Vimal
 
AI and the Future of Healthcare, Siemens Healthineers
AI and the Future of Healthcare, Siemens HealthineersAI and the Future of Healthcare, Siemens Healthineers
AI and the Future of Healthcare, Siemens HealthineersLevi Shapiro
 
Skynet Lessons Learned H4D Stanford 2016
Skynet Lessons Learned H4D Stanford 2016Skynet Lessons Learned H4D Stanford 2016
Skynet Lessons Learned H4D Stanford 2016Stanford University
 
Intro to open source - 101 presentation
Intro to open source - 101 presentationIntro to open source - 101 presentation
Intro to open source - 101 presentationJavier Perez
 

What's hot (20)

Metamaterial based antenna with application to real life system
Metamaterial based antenna with application to real life systemMetamaterial based antenna with application to real life system
Metamaterial based antenna with application to real life system
 
Mutual coupling
Mutual couplingMutual coupling
Mutual coupling
 
YAZILIM SÜREÇ İYİLEŞTİRME ÇALIŞMALARINDA YETENEK OLGUNLUK MODELİ ENTEGRASYONU...
YAZILIM SÜREÇ İYİLEŞTİRME ÇALIŞMALARINDA YETENEK OLGUNLUK MODELİ ENTEGRASYONU...YAZILIM SÜREÇ İYİLEŞTİRME ÇALIŞMALARINDA YETENEK OLGUNLUK MODELİ ENTEGRASYONU...
YAZILIM SÜREÇ İYİLEŞTİRME ÇALIŞMALARINDA YETENEK OLGUNLUK MODELİ ENTEGRASYONU...
 
Edge ppt
Edge pptEdge ppt
Edge ppt
 
DJI's Drone Solutions for Smart Cities of the Future
DJI's Drone Solutions for Smart Cities of the FutureDJI's Drone Solutions for Smart Cities of the Future
DJI's Drone Solutions for Smart Cities of the Future
 
MQTT - MQ Telemetry Transport for Message Queueing
MQTT - MQ Telemetry Transport for Message QueueingMQTT - MQ Telemetry Transport for Message Queueing
MQTT - MQ Telemetry Transport for Message Queueing
 
Swarming drones
Swarming dronesSwarming drones
Swarming drones
 
Open source technology
Open source technologyOpen source technology
Open source technology
 
Mbc system
Mbc systemMbc system
Mbc system
 
DSD-INT 2015 - Photogrammetric workflows and use of UA VS, Francesco nex, E-s...
DSD-INT 2015 - Photogrammetric workflows and use of UA VS, Francesco nex, E-s...DSD-INT 2015 - Photogrammetric workflows and use of UA VS, Francesco nex, E-s...
DSD-INT 2015 - Photogrammetric workflows and use of UA VS, Francesco nex, E-s...
 
Computer Networking A Top-Down Approach 6th editiion.pdf
Computer Networking A Top-Down Approach 6th editiion.pdfComputer Networking A Top-Down Approach 6th editiion.pdf
Computer Networking A Top-Down Approach 6th editiion.pdf
 
Desing of a rectangular patch antenna
Desing of a rectangular patch antennaDesing of a rectangular patch antenna
Desing of a rectangular patch antenna
 
CST_ANTENNA-CST STUDIO SUITE™ 2006B.pdf
CST_ANTENNA-CST STUDIO SUITE™ 2006B.pdfCST_ANTENNA-CST STUDIO SUITE™ 2006B.pdf
CST_ANTENNA-CST STUDIO SUITE™ 2006B.pdf
 
Seminar on uav
Seminar on uavSeminar on uav
Seminar on uav
 
AI-Sustainability.pptx
AI-Sustainability.pptxAI-Sustainability.pptx
AI-Sustainability.pptx
 
Design and Simulation Microstrip patch Antenna using CST Microwave Studio
Design and Simulation Microstrip patch Antenna  using CST Microwave StudioDesign and Simulation Microstrip patch Antenna  using CST Microwave Studio
Design and Simulation Microstrip patch Antenna using CST Microwave Studio
 
Intro to wireless sensor network
Intro to wireless sensor networkIntro to wireless sensor network
Intro to wireless sensor network
 
AI and the Future of Healthcare, Siemens Healthineers
AI and the Future of Healthcare, Siemens HealthineersAI and the Future of Healthcare, Siemens Healthineers
AI and the Future of Healthcare, Siemens Healthineers
 
Skynet Lessons Learned H4D Stanford 2016
Skynet Lessons Learned H4D Stanford 2016Skynet Lessons Learned H4D Stanford 2016
Skynet Lessons Learned H4D Stanford 2016
 
Intro to open source - 101 presentation
Intro to open source - 101 presentationIntro to open source - 101 presentation
Intro to open source - 101 presentation
 

Similar to FEDBN: Improve Federated Learning on Non-IID Data via Local Batch Normalization (39

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
 
Transfer Learning and Domain Adaptation - Ramon Morros - UPC Barcelona 2018
Transfer Learning and Domain Adaptation - Ramon Morros - UPC Barcelona 2018Transfer Learning and Domain Adaptation - Ramon Morros - UPC Barcelona 2018
Transfer Learning and Domain Adaptation - Ramon Morros - UPC Barcelona 2018Universitat Politècnica de Catalunya
 
TIP_TAViT_presentation.pdf
TIP_TAViT_presentation.pdfTIP_TAViT_presentation.pdf
TIP_TAViT_presentation.pdfBoahKim2
 
End-to-end deep auto-encoder for segmenting a moving object with limited tra...
End-to-end deep auto-encoder for segmenting a moving object  with limited tra...End-to-end deep auto-encoder for segmenting a moving object  with limited tra...
End-to-end deep auto-encoder for segmenting a moving object with limited tra...IJECEIAES
 
Fundamentals of Deep Recommender Systems
 Fundamentals of Deep Recommender Systems Fundamentals of Deep Recommender Systems
Fundamentals of Deep Recommender SystemsWQ Fan
 
Transfer Learning and Domain Adaptation (DLAI D5L2 2017 UPC Deep Learning for...
Transfer Learning and Domain Adaptation (DLAI D5L2 2017 UPC Deep Learning for...Transfer Learning and Domain Adaptation (DLAI D5L2 2017 UPC Deep Learning for...
Transfer Learning and Domain Adaptation (DLAI D5L2 2017 UPC Deep Learning for...Universitat Politècnica de Catalunya
 
Efficient Supervision Net: Polyp Segmentation using EfficientNet and Attentio...
Efficient Supervision Net: Polyp Segmentation using EfficientNet and Attentio...Efficient Supervision Net: Polyp Segmentation using EfficientNet and Attentio...
Efficient Supervision Net: Polyp Segmentation using EfficientNet and Attentio...multimediaeval
 
Mortgage Data for Machine Learning Algorithms
Mortgage Data for Machine Learning AlgorithmsMortgage Data for Machine Learning Algorithms
Mortgage Data for Machine Learning AlgorithmsAnne Klieve
 
IRJET- A Review on Object Tracking based on KNN Classifier
IRJET- A Review on Object Tracking based on KNN ClassifierIRJET- A Review on Object Tracking based on KNN Classifier
IRJET- A Review on Object Tracking based on KNN ClassifierIRJET Journal
 
The Declarative-Coordinated Model for Self-Optimization of Service Networks
The Declarative-Coordinated Model for Self-Optimization of Service NetworksThe Declarative-Coordinated Model for Self-Optimization of Service Networks
The Declarative-Coordinated Model for Self-Optimization of Service NetworksTokyo University of Science
 
Federated learning and its role in the privacy preservation of IoT devices
Federated learning and its role in the privacy preservation of IoT devicesFederated learning and its role in the privacy preservation of IoT devices
Federated learning and its role in the privacy preservation of IoT devicesAlAtfat
 
SEBD2015_PresentationVitali
SEBD2015_PresentationVitaliSEBD2015_PresentationVitali
SEBD2015_PresentationVitaliMonica Vitali
 
IRJET - Factors Affecting Deployment of Deep Learning based Face Recognition ...
IRJET - Factors Affecting Deployment of Deep Learning based Face Recognition ...IRJET - Factors Affecting Deployment of Deep Learning based Face Recognition ...
IRJET - Factors Affecting Deployment of Deep Learning based Face Recognition ...IRJET Journal
 
Grammatical Error Correction with Improved Real-world Applicability
Grammatical Error Correction with Improved Real-world ApplicabilityGrammatical Error Correction with Improved Real-world Applicability
Grammatical Error Correction with Improved Real-world ApplicabilityMasato Mita
 
Crafting Recommenders: the Shallow and the Deep of it!
Crafting Recommenders: the Shallow and the Deep of it! Crafting Recommenders: the Shallow and the Deep of it!
Crafting Recommenders: the Shallow and the Deep of it! Sudeep Das, Ph.D.
 
FACE PHOTO-SKETCH RECOGNITION USING DEEP LEARNING TECHNIQUES - A REVIEW
FACE PHOTO-SKETCH RECOGNITION USING DEEP LEARNING TECHNIQUES - A REVIEWFACE PHOTO-SKETCH RECOGNITION USING DEEP LEARNING TECHNIQUES - A REVIEW
FACE PHOTO-SKETCH RECOGNITION USING DEEP LEARNING TECHNIQUES - A REVIEWIRJET Journal
 
Mis 589 Success Begins / snaptutorial.com
Mis 589  Success Begins / snaptutorial.comMis 589  Success Begins / snaptutorial.com
Mis 589 Success Begins / snaptutorial.comWilliamsTaylor44
 
Mis 589 Massive Success / snaptutorial.com
Mis 589 Massive Success / snaptutorial.comMis 589 Massive Success / snaptutorial.com
Mis 589 Massive Success / snaptutorial.comStephenson185
 

Similar to FEDBN: Improve Federated Learning on Non-IID Data via Local Batch Normalization (39 (20)

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
 
Transfer Learning and Domain Adaptation - Ramon Morros - UPC Barcelona 2018
Transfer Learning and Domain Adaptation - Ramon Morros - UPC Barcelona 2018Transfer Learning and Domain Adaptation - Ramon Morros - UPC Barcelona 2018
Transfer Learning and Domain Adaptation - Ramon Morros - UPC Barcelona 2018
 
TIP_TAViT_presentation.pdf
TIP_TAViT_presentation.pdfTIP_TAViT_presentation.pdf
TIP_TAViT_presentation.pdf
 
End-to-end deep auto-encoder for segmenting a moving object with limited tra...
End-to-end deep auto-encoder for segmenting a moving object  with limited tra...End-to-end deep auto-encoder for segmenting a moving object  with limited tra...
End-to-end deep auto-encoder for segmenting a moving object with limited tra...
 
Fundamentals of Deep Recommender Systems
 Fundamentals of Deep Recommender Systems Fundamentals of Deep Recommender Systems
Fundamentals of Deep Recommender Systems
 
Transfer Learning and Domain Adaptation (DLAI D5L2 2017 UPC Deep Learning for...
Transfer Learning and Domain Adaptation (DLAI D5L2 2017 UPC Deep Learning for...Transfer Learning and Domain Adaptation (DLAI D5L2 2017 UPC Deep Learning for...
Transfer Learning and Domain Adaptation (DLAI D5L2 2017 UPC Deep Learning for...
 
Large Scale Distributed Deep Networks
Large Scale Distributed Deep NetworksLarge Scale Distributed Deep Networks
Large Scale Distributed Deep Networks
 
Efficient Supervision Net: Polyp Segmentation using EfficientNet and Attentio...
Efficient Supervision Net: Polyp Segmentation using EfficientNet and Attentio...Efficient Supervision Net: Polyp Segmentation using EfficientNet and Attentio...
Efficient Supervision Net: Polyp Segmentation using EfficientNet and Attentio...
 
Presentation_final.pdf
Presentation_final.pdfPresentation_final.pdf
Presentation_final.pdf
 
Mortgage Data for Machine Learning Algorithms
Mortgage Data for Machine Learning AlgorithmsMortgage Data for Machine Learning Algorithms
Mortgage Data for Machine Learning Algorithms
 
IRJET- A Review on Object Tracking based on KNN Classifier
IRJET- A Review on Object Tracking based on KNN ClassifierIRJET- A Review on Object Tracking based on KNN Classifier
IRJET- A Review on Object Tracking based on KNN Classifier
 
The Declarative-Coordinated Model for Self-Optimization of Service Networks
The Declarative-Coordinated Model for Self-Optimization of Service NetworksThe Declarative-Coordinated Model for Self-Optimization of Service Networks
The Declarative-Coordinated Model for Self-Optimization of Service Networks
 
Federated learning and its role in the privacy preservation of IoT devices
Federated learning and its role in the privacy preservation of IoT devicesFederated learning and its role in the privacy preservation of IoT devices
Federated learning and its role in the privacy preservation of IoT devices
 
SEBD2015_PresentationVitali
SEBD2015_PresentationVitaliSEBD2015_PresentationVitali
SEBD2015_PresentationVitali
 
IRJET - Factors Affecting Deployment of Deep Learning based Face Recognition ...
IRJET - Factors Affecting Deployment of Deep Learning based Face Recognition ...IRJET - Factors Affecting Deployment of Deep Learning based Face Recognition ...
IRJET - Factors Affecting Deployment of Deep Learning based Face Recognition ...
 
Grammatical Error Correction with Improved Real-world Applicability
Grammatical Error Correction with Improved Real-world ApplicabilityGrammatical Error Correction with Improved Real-world Applicability
Grammatical Error Correction with Improved Real-world Applicability
 
Crafting Recommenders: the Shallow and the Deep of it!
Crafting Recommenders: the Shallow and the Deep of it! Crafting Recommenders: the Shallow and the Deep of it!
Crafting Recommenders: the Shallow and the Deep of it!
 
FACE PHOTO-SKETCH RECOGNITION USING DEEP LEARNING TECHNIQUES - A REVIEW
FACE PHOTO-SKETCH RECOGNITION USING DEEP LEARNING TECHNIQUES - A REVIEWFACE PHOTO-SKETCH RECOGNITION USING DEEP LEARNING TECHNIQUES - A REVIEW
FACE PHOTO-SKETCH RECOGNITION USING DEEP LEARNING TECHNIQUES - A REVIEW
 
Mis 589 Success Begins / snaptutorial.com
Mis 589  Success Begins / snaptutorial.comMis 589  Success Begins / snaptutorial.com
Mis 589 Success Begins / snaptutorial.com
 
Mis 589 Massive Success / snaptutorial.com
Mis 589 Massive Success / snaptutorial.comMis 589 Massive Success / snaptutorial.com
Mis 589 Massive Success / snaptutorial.com
 

Recently uploaded

Vision and reflection on Mining Software Repositories research in 2024
Vision and reflection on Mining Software Repositories research in 2024Vision and reflection on Mining Software Repositories research in 2024
Vision and reflection on Mining Software Repositories research in 2024AyushiRastogi48
 
insect anatomy and insect body wall and their physiology
insect anatomy and insect body wall and their  physiologyinsect anatomy and insect body wall and their  physiology
insect anatomy and insect body wall and their physiologyDrAnita Sharma
 
Recombinant DNA technology( Transgenic plant and animal)
Recombinant DNA technology( Transgenic plant and animal)Recombinant DNA technology( Transgenic plant and animal)
Recombinant DNA technology( Transgenic plant and animal)DHURKADEVIBASKAR
 
Evidences of Evolution General Biology 2
Evidences of Evolution General Biology 2Evidences of Evolution General Biology 2
Evidences of Evolution General Biology 2John Carlo Rollon
 
Microphone- characteristics,carbon microphone, dynamic microphone.pptx
Microphone- characteristics,carbon microphone, dynamic microphone.pptxMicrophone- characteristics,carbon microphone, dynamic microphone.pptx
Microphone- characteristics,carbon microphone, dynamic microphone.pptxpriyankatabhane
 
Speech, hearing, noise, intelligibility.pptx
Speech, hearing, noise, intelligibility.pptxSpeech, hearing, noise, intelligibility.pptx
Speech, hearing, noise, intelligibility.pptxpriyankatabhane
 
Analytical Profile of Coleus Forskohlii | Forskolin .pptx
Analytical Profile of Coleus Forskohlii | Forskolin .pptxAnalytical Profile of Coleus Forskohlii | Forskolin .pptx
Analytical Profile of Coleus Forskohlii | Forskolin .pptxSwapnil Therkar
 
Artificial Intelligence In Microbiology by Dr. Prince C P
Artificial Intelligence In Microbiology by Dr. Prince C PArtificial Intelligence In Microbiology by Dr. Prince C P
Artificial Intelligence In Microbiology by Dr. Prince C PPRINCE C P
 
zoogeography of pakistan.pptx fauna of Pakistan
zoogeography of pakistan.pptx fauna of Pakistanzoogeography of pakistan.pptx fauna of Pakistan
zoogeography of pakistan.pptx fauna of Pakistanzohaibmir069
 
Gas_Laws_powerpoint_notes.ppt for grade 10
Gas_Laws_powerpoint_notes.ppt for grade 10Gas_Laws_powerpoint_notes.ppt for grade 10
Gas_Laws_powerpoint_notes.ppt for grade 10ROLANARIBATO3
 
Neurodevelopmental disorders according to the dsm 5 tr
Neurodevelopmental disorders according to the dsm 5 trNeurodevelopmental disorders according to the dsm 5 tr
Neurodevelopmental disorders according to the dsm 5 trssuser06f238
 
Forest laws, Indian forest laws, why they are important
Forest laws, Indian forest laws, why they are importantForest laws, Indian forest laws, why they are important
Forest laws, Indian forest laws, why they are importantadityabhardwaj282
 
Call Girls in Munirka Delhi 💯Call Us 🔝8264348440🔝
Call Girls in Munirka Delhi 💯Call Us 🔝8264348440🔝Call Girls in Munirka Delhi 💯Call Us 🔝8264348440🔝
Call Girls in Munirka Delhi 💯Call Us 🔝8264348440🔝soniya singh
 
Spermiogenesis or Spermateleosis or metamorphosis of spermatid
Spermiogenesis or Spermateleosis or metamorphosis of spermatidSpermiogenesis or Spermateleosis or metamorphosis of spermatid
Spermiogenesis or Spermateleosis or metamorphosis of spermatidSarthak Sekhar Mondal
 
Call Girls in Aiims Metro Delhi 💯Call Us 🔝9953322196🔝 💯Escort.
Call Girls in Aiims Metro Delhi 💯Call Us 🔝9953322196🔝 💯Escort.Call Girls in Aiims Metro Delhi 💯Call Us 🔝9953322196🔝 💯Escort.
Call Girls in Aiims Metro Delhi 💯Call Us 🔝9953322196🔝 💯Escort.aasikanpl
 
Call Us ≽ 9953322196 ≼ Call Girls In Mukherjee Nagar(Delhi) |
Call Us ≽ 9953322196 ≼ Call Girls In Mukherjee Nagar(Delhi) |Call Us ≽ 9953322196 ≼ Call Girls In Mukherjee Nagar(Delhi) |
Call Us ≽ 9953322196 ≼ Call Girls In Mukherjee Nagar(Delhi) |aasikanpl
 
BIOETHICS IN RECOMBINANT DNA TECHNOLOGY.
BIOETHICS IN RECOMBINANT DNA TECHNOLOGY.BIOETHICS IN RECOMBINANT DNA TECHNOLOGY.
BIOETHICS IN RECOMBINANT DNA TECHNOLOGY.PraveenaKalaiselvan1
 
Twin's paradox experiment is a meassurement of the extra dimensions.pptx
Twin's paradox experiment is a meassurement of the extra dimensions.pptxTwin's paradox experiment is a meassurement of the extra dimensions.pptx
Twin's paradox experiment is a meassurement of the extra dimensions.pptxEran Akiva Sinbar
 
Is RISC-V ready for HPC workload? Maybe?
Is RISC-V ready for HPC workload? Maybe?Is RISC-V ready for HPC workload? Maybe?
Is RISC-V ready for HPC workload? Maybe?Patrick Diehl
 
TOTAL CHOLESTEROL (lipid profile test).pptx
TOTAL CHOLESTEROL (lipid profile test).pptxTOTAL CHOLESTEROL (lipid profile test).pptx
TOTAL CHOLESTEROL (lipid profile test).pptxdharshini369nike
 

Recently uploaded (20)

Vision and reflection on Mining Software Repositories research in 2024
Vision and reflection on Mining Software Repositories research in 2024Vision and reflection on Mining Software Repositories research in 2024
Vision and reflection on Mining Software Repositories research in 2024
 
insect anatomy and insect body wall and their physiology
insect anatomy and insect body wall and their  physiologyinsect anatomy and insect body wall and their  physiology
insect anatomy and insect body wall and their physiology
 
Recombinant DNA technology( Transgenic plant and animal)
Recombinant DNA technology( Transgenic plant and animal)Recombinant DNA technology( Transgenic plant and animal)
Recombinant DNA technology( Transgenic plant and animal)
 
Evidences of Evolution General Biology 2
Evidences of Evolution General Biology 2Evidences of Evolution General Biology 2
Evidences of Evolution General Biology 2
 
Microphone- characteristics,carbon microphone, dynamic microphone.pptx
Microphone- characteristics,carbon microphone, dynamic microphone.pptxMicrophone- characteristics,carbon microphone, dynamic microphone.pptx
Microphone- characteristics,carbon microphone, dynamic microphone.pptx
 
Speech, hearing, noise, intelligibility.pptx
Speech, hearing, noise, intelligibility.pptxSpeech, hearing, noise, intelligibility.pptx
Speech, hearing, noise, intelligibility.pptx
 
Analytical Profile of Coleus Forskohlii | Forskolin .pptx
Analytical Profile of Coleus Forskohlii | Forskolin .pptxAnalytical Profile of Coleus Forskohlii | Forskolin .pptx
Analytical Profile of Coleus Forskohlii | Forskolin .pptx
 
Artificial Intelligence In Microbiology by Dr. Prince C P
Artificial Intelligence In Microbiology by Dr. Prince C PArtificial Intelligence In Microbiology by Dr. Prince C P
Artificial Intelligence In Microbiology by Dr. Prince C P
 
zoogeography of pakistan.pptx fauna of Pakistan
zoogeography of pakistan.pptx fauna of Pakistanzoogeography of pakistan.pptx fauna of Pakistan
zoogeography of pakistan.pptx fauna of Pakistan
 
Gas_Laws_powerpoint_notes.ppt for grade 10
Gas_Laws_powerpoint_notes.ppt for grade 10Gas_Laws_powerpoint_notes.ppt for grade 10
Gas_Laws_powerpoint_notes.ppt for grade 10
 
Neurodevelopmental disorders according to the dsm 5 tr
Neurodevelopmental disorders according to the dsm 5 trNeurodevelopmental disorders according to the dsm 5 tr
Neurodevelopmental disorders according to the dsm 5 tr
 
Forest laws, Indian forest laws, why they are important
Forest laws, Indian forest laws, why they are importantForest laws, Indian forest laws, why they are important
Forest laws, Indian forest laws, why they are important
 
Call Girls in Munirka Delhi 💯Call Us 🔝8264348440🔝
Call Girls in Munirka Delhi 💯Call Us 🔝8264348440🔝Call Girls in Munirka Delhi 💯Call Us 🔝8264348440🔝
Call Girls in Munirka Delhi 💯Call Us 🔝8264348440🔝
 
Spermiogenesis or Spermateleosis or metamorphosis of spermatid
Spermiogenesis or Spermateleosis or metamorphosis of spermatidSpermiogenesis or Spermateleosis or metamorphosis of spermatid
Spermiogenesis or Spermateleosis or metamorphosis of spermatid
 
Call Girls in Aiims Metro Delhi 💯Call Us 🔝9953322196🔝 💯Escort.
Call Girls in Aiims Metro Delhi 💯Call Us 🔝9953322196🔝 💯Escort.Call Girls in Aiims Metro Delhi 💯Call Us 🔝9953322196🔝 💯Escort.
Call Girls in Aiims Metro Delhi 💯Call Us 🔝9953322196🔝 💯Escort.
 
Call Us ≽ 9953322196 ≼ Call Girls In Mukherjee Nagar(Delhi) |
Call Us ≽ 9953322196 ≼ Call Girls In Mukherjee Nagar(Delhi) |Call Us ≽ 9953322196 ≼ Call Girls In Mukherjee Nagar(Delhi) |
Call Us ≽ 9953322196 ≼ Call Girls In Mukherjee Nagar(Delhi) |
 
BIOETHICS IN RECOMBINANT DNA TECHNOLOGY.
BIOETHICS IN RECOMBINANT DNA TECHNOLOGY.BIOETHICS IN RECOMBINANT DNA TECHNOLOGY.
BIOETHICS IN RECOMBINANT DNA TECHNOLOGY.
 
Twin's paradox experiment is a meassurement of the extra dimensions.pptx
Twin's paradox experiment is a meassurement of the extra dimensions.pptxTwin's paradox experiment is a meassurement of the extra dimensions.pptx
Twin's paradox experiment is a meassurement of the extra dimensions.pptx
 
Is RISC-V ready for HPC workload? Maybe?
Is RISC-V ready for HPC workload? Maybe?Is RISC-V ready for HPC workload? Maybe?
Is RISC-V ready for HPC workload? Maybe?
 
TOTAL CHOLESTEROL (lipid profile test).pptx
TOTAL CHOLESTEROL (lipid profile test).pptxTOTAL CHOLESTEROL (lipid profile test).pptx
TOTAL CHOLESTEROL (lipid profile test).pptx
 

FEDBN: Improve Federated Learning on Non-IID Data via Local Batch Normalization (39

  • 1. FEDBN: FEDERATED LEARNING ON NON-IID FEATURES VIA LOCAL BATCH NORMALIZATION Paper presentation by Anam ur rehman Contact: anamur.rehman@studenti.polito.it Published as a conference paper at ICLR 2021 1
  • 3. FEDBN: Federated learning[1] Year 2021 2020 2018 2017 2016 2015 2014 2013 2012 2011 [1] Jakub et al. Federated optimization: Distributed machine learning for on-device intelligence. 2016 Classical Machine Learning: • Centralized data storage • Training process computations at the central server. What if ?  Data stays distributed on remote devices Devices maintain control of their own data  Training is done locally on remote devices  One global model is learned via aggregation 3 Autonomous cars on an average generate around 4 GB of data per hour of driving.
  • 4. FEDBN: Federated learning Applications [1] • Transportation: self-driving cars • Healthcare: predictions on patient data • Cybersecurity: spam filtering • Smart applications: voice recognition, next word prediction, etc. [1] Read more: Priyanka et al. Federated Learning: Opportunities and Challenges, 2021 Challenges [1] • Communication Overheads: presence of stragglers • Heterogeneity: system, statistical (in contrast to distributed learning) • Privacy concerns 4 Year 2021 2020 2018 2017 2016 2015 2014 2013 2012 2011
  • 5. FEDBN: Federated learning Statistical heterogeneity among local datasets: • Unbalancedness: Clients may have different amount of data  Example, Spam filter for emails. • Covariate shift: Statistical distribution of data varies among clients  Example: NLP digits recognition • Concept shift: Same features may correspond to different labels for different clients  Example, in NLP, Sentiment analysis on same text may yield different sentiments for different clients 5
  • 6. FEDBN: Example of NonIID datasets MNIST MNIST-M USPS SynthDigits SVHN Covariate shift among datasets Digits dataset 6
  • 7. FEDBN: Related work • FedAvg[1]: Federated Average [1] Brendan McMahan et al. Communication-efficient learning of deep networks from decentralized data. 2017. At each communication round 1. Server randomly selects a subset of K clients and Send them current global model 2. Selected device k updates this model on local client data via SGD. After training client sends the new local model back to server 3. Server aggregates local models to form a new global model - Convergence in not guranteed. In hetergeneous settings it can diverge [1] Year 2021 2020 2018 2017 2016 2015 2014 2013 2012 2011 7
  • 8. FEDBN: Related work • FedProx[1]: Federated Optimization in Heterogeneous Networks [1] Tian Li et al, In Conference on Machine Learning and Systems, 2020a, 2020b. Slide credit: Tian Li, MLSys presentation. + Limits the impact of heterogeneous local updates + Safely incorporate partial work of stragglers + Generalization of FedAvg; Allows for any local solver + Theoretical guarantees for convergence Year 2021 2020 2018 2017 2016 2015 2014 2013 2012 2011 8
  • 9. FEDBN: Related work • SiloBN[1]: Siloed Federated Learning for Multi-Centric Histopathology Datasets [1] Mathieu Andreux et al, Siloed federated learning for multi-centric histopathology datasets, pp. 129–139. Springer, 2020. Slide credits: [1] Year 2021 2020 2018 2017 2016 2015 2014 2013 2012 2011 9
  • 10. FEDBN: Batch Normalization Year 2021 2020 2018 2017 2016 2015 2014 2013 2012 2011 [1] Sergey Ioffe et al. Batch normalization: Accelerating deep network training by reducing internal covariate shift. 2015 γ and β are the only learnable parameters of BN layer. 10 Why we use it ? To reduce internal covariate shift in neural network [1]. How it works ?
  • 12. FEDBN: Problem with non IID data 12 Consider a simple, non-convex learning problem: s.t ϵ ∼ 𝒩 0, σ2 𝑤2 ∗ Two clients train a model s.t x1 ∼ 𝒩 μ, σ1 2 . x2 ∼ 𝒩 μ, σ2 2 . and σ1 2 ≠ σ2 2 local squared loss before and after local BN 𝑤1 ∗ w 𝑓𝑤 𝑥𝑖 = 𝑐𝑜𝑠 𝑤𝑥𝑖
  • 13. FEDBN: Why not just take the average? (SiloBN) Client 1 w1 ∗ : Optimal weight γ1 ∗ : Optimal BN parameter Observation 1: For a fixed optimal weight w1 ∗ , changing γ deteriorates the model quality. Observation 2: For a given optimal BN parameter γ1 ∗ , changing w deteriorates the quality. 13
  • 14. FEDBN: How it works? Local training Global aggregation 14
  • 16. FEDBN: How it Really Works? Source: med-air/FedBN (github.com) 16 Global Aggregation
  • 17. FEDBN: Results on digit dataset (FedAvg vs FedBN) • Outperforms FedAvg on SVHN dataset • Faster convergence • Smooth and robust convergence 17
  • 18. FEDBN: Results; what if Communication is done at different frequencies? 18
  • 19. FEDBN: Results; what if Dataset size varies for each client? 19
  • 20. FEDBN: Contributions Provides convergence guarantees. Improves the convergence behavior on non-IID datasets. One step further in privacy of client’s data. 20
  • 21. FEDBN: Take home message • Use batch normalization • Keep it local • Smooth and fast convergence 21
  • 22. Useful links Federated Optimization in Heterogeneous Networks FedProx presentation by Tian Li: 22 med-air/FedBN (github.com) Pytorch implementation of FedBN: Brendan McMahan’s Talk: Guarding user Privacy with Federated Learning