Latest News

The AML Dashboard is a versatile tool designed to bridge the gap between the complexity of algebraic machine learning (AML) and its practical application. Tailored to professionals, including researchers, computer scientists, and creative experts, this intuitive platform streamlines the process of training, evaluating, and applying AML models across diverse datasets. With its user-friendly interfaces, the AML Dashboard enhances data management, model training, and real-time exploration, making AML more accessible and efficient. Whether advancing machine learning research or exploring new creative possibilities, the AML Dashboard provides the essential tools to leverage the AML framework fully.

The AML Dashboard emerges as a user-friendly platform for training, evaluating, and applying AML models across several datasets.

The primary goal of the AML Dashboard is to enable users to:

1. Collect and Manage Data: Record personal gesture data or load existing datasets.
2. Train Models: Train classifiers using the AML Engine.
3. Evaluate Performance: Compare AML models with other machine learning approaches.

Interfaces

The Dashboard is structured into various interfaces, each tailored to specific stages of the AML workflow.

• Data Collection

The Data Collection tab allows users to create, load, and manage datasets efficiently. Users can record gesture data, load standard image classification datasets, or upload custom datasets. A dataset browser enables easy navigation, offering tools to explore, modify, or delete images and manage class labels.

• Training

 The Training page is where users can configure and monitor the AML model's training process. It provides options to set training parameters, track training progress, and compare the AML model's performance with other classifiers. Additionally, users can train the AML and neural network models for future comparison.

• Fetching

This interface simplifies the retrieval of AML statistics and trained models. Users can search for specific statistics or request AML models with just a few clicks.

• Real-Time Pattern Exploration

 After training the models, users can move to the real-time pattern exploration interface. This part of the AML Dashboard allows users to explore music patterns using the self-trained models.

This feature is only available when the dataset contains the captured gesture data. 

• Context Broker Interaction

 This tab facilitates integration with the Context Broker, allowing users to upload images, perform inference tasks, and receive predictions from trained AML models.

• AML-IP Nodes Management

This interface facilitates the management of the AML-IP nodes within the network. Users can switch between tabs to create and destroy agent, computing, inference and sender nodes.

• Status

 The Status tab offers a comprehensive view of all active AML-IP nodes, displaying their ID, type, and operational status.

Use Cases

AML Dashboard addresses two primary use cases, catering to different audiences and purposes:  

1. Creative Professionals:
   Musicians, sound designers, and artists use the AML Dashboard to explore innovative sound patterns. For instance, musician Emilie records hand gestures to train an AML model that generates unique soundscapes for her compositions.
2. Computer Scientists:
   Researchers like Victor utilize the Dashboard to test and refine the machine learning approach. Its intuitive interface allows for data management, model training, performance evaluation, and the creation and destruction of AML-IP Nodes.

By This email address is being protected from spambots. You need JavaScript enabled to view it. from eProsima.

MORE INFORMATION:

To learn more about AML, click here.

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