mccorby/FederatedAndroidTrainer

This is a federated learning application that trains machine learning models on local devices and aggregates updates via a central server. It supports multiple datasets, including Iris and MNIST, and allows configuration of model parameters and server endpoints. The system standardizes data access and model training interfaces to enable distributed gradient-based learning workflows.

A new IrisModel class has been added to the application, implementing the FederatedModel interface to handle Iris dataset classification. This model constructs a multi-layer neural network using Deeplearning3j, supporting training, evaluation, and weight updates via gradients.

app/src/main/java/com/mccorby/federatedlearning/features/iris/model · high confidence

The app now supports external configuration via a new \config.json\ file, allowing users to specify the machine learning model (e.g., "Iris"), the server URL for gradient transmission, the maximum number of clients, and the batch size. Additionally, a \diabetes.csv\ dataset has been added to the app's assets, enabling the application to train models on this specific dataset.

app/src · high confidence

A change to existing behaviour in model — 2 commits, 1 file.

model · medium confidence · unverified

The MNIST and SumOperation features now use new \MNISTDataSource\ and \SumDataSource\ classes that implement the \FederatedDataSource\ interface, replacing the previous \DataSource\ implementations. Similarly, \MNISTModel\ and \LinearModel\ now implement \FederatedModel\, enabling the system to send gradients to a server and update local weights with remote gradients. This change restructures the codebase to support federated learning workflows by standardizing data access and model training/evaluation interfaces.

app/src/main/java/com/mccorby/federatedlearning/features/mnist, app/src/main/java/com/mccorby/federatedlearning/features/sumoperation · high confidence

The project has been restructured to support Federated Learning, reflected by the removal of the separate 'model' module and the renaming of the application ID to 'com.mccorby.federatedlearning'. The build configuration now explicitly defines and uses versioned dependencies for Deep Learning (DL4J, ND4J), networking (Retrofit, RxJava), and dependency injection (Dagger), replacing the previous flat file-based library management.

(dependencies) · high confidence

Written by watchdog.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

• CAI 60 → 60 (-0.3)
• Rubric changed (rubric-2026.08.19 → rubric-2026.09.15) — scores are not directly comparable.

• Code Health 99 → 97 (-1.6)
• Architecture 100 → 96 (-4.4)
• Maturity 82 → 73 (-9.8)
• Readiness 24 → 36 (+11.7)
• Security 100 → 100 (+0.0)
• Domain Modelling 100 → 79 (-21.3)

• Coverage not included — suite not readable by the collector
• Dependency hygiene not measured — dependency manifest found but not parsed for hygiene
• Duplicated block (11 lines × 2) (app/src/main/java/com/mccorby/federatedlearning/features/diabetes/model/DiabetesModel.java)
• Duplicated block (11 lines × 2) (app/src/main/java/com/mccorby/federatedlearning/features/diabetes/model/DiabetesModel.java)
• Duplicated block (5 lines × 2) (app/src/main/java/com/mccorby/federatedlearning/features/sumoperation/datasource/SumDataSource.java)
• Duplicated block (6 lines × 2) (app/src/main/java/com/mccorby/federatedlearning/features/iris/model/IrisModel.java)
• No exposed public API
• Test reliability not included
• early-stage repository — too few commits for a meaningful bus factor
• early-stage repository — too little history to judge knowledge freshness
• git history depth insufficient
• git history depth insufficient

• Dependency hygiene PARTLY measured — Maven/Gradle declarations read, no dependency graph resolved
• Documentation: no installation or build instructions (README.md)
• Duplicated block (10 lines × 2) (app/src/main/java/com/mccorby/federatedlearning/features/diabetes/datasource/DiabetesFileDataSource.java)
• Duplicated block (10 lines × 2) (app/src/main/java/com/mccorby/federatedlearning/features/diabetes/datasource/DiabetesFileDataSource.java)
• Duplicated block (11 lines × 2) (app/src/main/java/com/mccorby/federatedlearning/features/diabetes/model/DiabetesModel.java)
• Duplicated block (16 lines × 2) (app/src/main/java/com/mccorby/federatedlearning/features/diabetes/model/DiabetesModel.java)
• Duplicated block (5–9 lines × 2) (app/src/main/java/com/mccorby/federatedlearning/features/iris/model/IrisModel.java)
• TodoComment (app/src/main/java/com/mccorby/federatedlearning/app/network/RetrofitServerService.java)
• TodoComment (app/src/main/java/com/mccorby/federatedlearning/app/presentation/TrainerPresenter.java)
• TodoComment (app/src/main/java/com/mccorby/federatedlearning/app/presentation/TrainerPresenter.java)
• TodoComment (app/src/main/java/com/mccorby/federatedlearning/app/presentation/TrainerPresenter.java)
• TodoComment (app/src/main/java/com/mccorby/federatedlearning/core/domain/model/FederatedModel.java)
• TodoComment (app/src/main/java/com/mccorby/federatedlearning/di/MainModule.java)
• TodoComment (app/src/main/java/com/mccorby/federatedlearning/features/sumoperation/model/LinearModel.java)

Written by watchdog.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

mccorby/FederatedAndroidTrainer was measured the same way every project in this corpus was: the same rubric, at a pinned commit, with the result published in full. Point a surveyor at a repository you know and see whether you agree with it.