PPML FGBoost tutorial draft (#4751)
This commit is contained in:
		
							parent
							
								
									500b7bbe99
								
							
						
					
					
						commit
						18a9c2f57a
					
				
					 1 changed files with 22 additions and 0 deletions
				
			
		
							
								
								
									
										22
									
								
								docs/readthedocs/source/doc/PPML/VFL/overview.md
									
									
									
									
									
										Normal file
									
								
							
							
						
						
									
										22
									
								
								docs/readthedocs/source/doc/PPML/VFL/overview.md
									
									
									
									
									
										Normal file
									
								
							| 
						 | 
				
			
			@ -0,0 +1,22 @@
 | 
			
		|||
# Vertical Federated Learning
 | 
			
		||||
Vertical Federated Learning (VFL) is a federated machine learning case where multiple data sets share the same sample ID space but differ in feature space. 
 | 
			
		||||
 | 
			
		||||
VFL is supported in BigDL PPML. It allows users to train a federated machine learning model where data features are held by different parties. In BigDL PPML, the following VFL scenarios are supported.
 | 
			
		||||
* **Private Set Intersection**: To get data intersection of different VFL parties.
 | 
			
		||||
* **Neural Network Model**: To train common neural network model with Pytorch or Tensorflow backend across VFL parties.
 | 
			
		||||
* **FGBoost Model**: To train gradient boosted decision tree (GBDT) model across multiple VFL parties.
 | 
			
		||||
 | 
			
		||||
For each scenario, some quick starts are available in above links.
 | 
			
		||||
 | 
			
		||||
## Key Concepts
 | 
			
		||||
A **FL Server** is a gRPC server to handle requests from FL Client. A **FL Client** is a gRPC client to send requests to FL Server. These requests include:
 | 
			
		||||
* serialized model to use in training at FL Server
 | 
			
		||||
* some model related instance, e.g. loss function, optimizer
 | 
			
		||||
* the Tensor which FL Server and FL Client interact, e.g. predict output, label, gradient
 | 
			
		||||
 | 
			
		||||
A **FL Context** is a singleton holding a FL Client instance. By default, only one instance is held in a FL application. And the gRPC channel in this singleton instance could be reused in multiple algorithms.
 | 
			
		||||
 | 
			
		||||
 | 
			
		||||
## Lifecycle
 | 
			
		||||
 | 
			
		||||
## Fault Tolerance
 | 
			
		||||
		Loading…
	
		Reference in a new issue