Latest News

As part of the ALMA project’s mission to make Algebraic Machine Learning (AML) accessible and applicable across domains, we present the AML Dashboard—a comprehensive graphical interface designed for training, evaluating, and applying AML models on structured and sensor-based datasets.

Developed using the Marcelle framework, the AML Dashboard provides modular, extensible GUIs that allow researchers and developers to interact with AML components in real time. It supports both data-centric and model-centric workflows and serves as a bridge between AML’s theoretical foundations and practical deployment.

The dashboard enables users to:

• Collect custom gesture data or load structured datasets,

• Train AML classifiers directly through a GUI without writing code,

• Compare AML with other machine learning models through performance metrics and visualizations,

• Run real-time inference (for gesture-based datasets) to validate model behavior in live settings,

• Manage AML-IP scenarios, including model distribution, inference handling, and collaborative learning tasks.

Core Interfaces

The AML Dashboard is organized into seven primary components:

1. Dataset Interface – Users can upload or record gesture data and prepare it for model training.

2. Training Interface – Enables training of AML models from loaded datasets using adjustable parameters.

3. Model Exploration Interface – Provides tools for model introspection and comparison with conventional ML classifiers.

4. Real-Time Exploration Interface – For sensor datasets, allows real-time feedback based on user gestures.

5. Model Fetching Interface – Supports integration with AML-IP for collaborative learning and distributed model updates.

6. Context Broker Interface – Allows connection with inference services to process and act upon real-world data streams.

7. AML-IP Management Interface – Offers control and monitoring of AML-IP instances for managing distributed nodes.

Video Demonstration

We invite you to watch this video showcasing the AML Dashboard’s capabilities. The video highlights how the dashboard supports the end-to-end lifecycle of AML model development—from dataset preparation to real-time application, focusing on usability and extensibility.

by Denisa Alexandru from eProsima

MORE INFORMATION ABOUT ALMA:

For any questions please contact This email address is being protected from spambots. You need JavaScript enabled to view it..