Releases: FederatedAI/FATE-Board
Releases · FederatedAI/FATE-Board
Release v1.8.0
Major Features and Improvements
Major Features
- Support GBDTMO mode for Secureboost
- Add SSHE Linr component
- Add Writer component
- Delete data-output port for data-statistic compoent
Major Improvements
- Fix display bug and optimize transport protocal support
Release v1.7.2.2
Major Features and Improvements
- Detail finalization and Configuration calibration
Release v1.7.2.1
Major Features and Improvements
Major Features
- Update shelljs from 0.8.3 to 0.8.5
- Update spring-core from 5.2.9Release to 5.3.14
Release v1.7.2
Major Features and Improvements
- Confusion matrix data display upgrade
- Component port bug fixed
- No data prompt upgrade
Release 1.7.1
Major Features and Improvements
- Disable password storage to avoid security problem
- Update DOM structure for login-form
Release v1.7.0
Major Features and Improvements
- Add new components
- Components display optimization
- Hetero binning supports multi-category display
- Embodiment of warm start mechanism
- Interaction optimization
Release v1.6.1
Major Features and Improvements
Major Features
- Support HTTPS protocol
- Scientific notation display optimization
- Anonymous display optimization
- Fix routing issues
Release v1.5.3
Major Features and Improvements
Major Features
- Fix feature anonym inconsistency of HeteroBinning
- Login page memorizes default account(admin) and password(admin)
Release v1.5.2
Major Features and Improvements
Major Features
- Update for security
- Add default login account (Account:admin Password:admin)
Release v1.6.0
Major Features and Improvements
- SBT Transformer: new module, encode features using Hetero SBT model
- Sample Weight: new module, set sample weights based on label or from feature column
- Homo Feature Binning: new module, performs homogeneous federated binning, bin number and bin interval available on Board
- Data Transformer: new module, same functionality as DataIO with added features such as case-insensitive
- Reader: supports image format input
- Hetero Feature Binning: filter results display by WOE
- Hetero Pearson: support VIF computation
- Feature Selection: support selection based on VIF
- SecureBoost: display gain & split simultaneously; Performance Score diagram for binary classification
- Cross Validation: display & download fold split history and fold prediction results
- Evaluation: metrics & visualization for each one_vs_rest child model
- Export job DSL & Conf files on Board
- Search for jobs using party_id
- Improved component model/data download
- Improved performance score display over large number of iterations or cv folds