ML Architecture

‍

The components of a machine learning solution

1. Data Generation: Every machine learning application lives off data. That data has to come from somewhere. 
2. Data Collection: Data is only useful if it’s accessible, so it needs to be stored – ideally in a consistent structure and conveniently in one place.
3. Feature Engineering Pipeline: We have to select, transform, combine, and otherwise prepare our data so the algorithm can find useful patterns.
4. Training: We apply algorithms, and they learn patterns from the data. Then they use these patterns to perform particular tasks.
5. Evaluation: We need to carefully test how well our algorithm performs on data it hasn’t seen before (during training). 
6. Task Orchestration: Feature engineering, training, and prediction all need to be scheduled on our compute infrastructure (such as AWS or Azure) – usually with non-trivial interdependence. 
7. Prediction:  We use the model we’ve trained to perform new tasks and solve new problems.
8. Infrastructure: Even in the age of the cloud, the solution has to live and be served somewhere. This will require setup and maintenance. 
9. Authentication: This keeps our models secure and makes sure only those who have permission can use them.
10. Interaction: We need some way to interact with our model and give it problems to solve. Usually this takes the form of an API.‍
11. Monitoring: We need to regularly check our model’s performance. This usually involves periodically generating a report or showing performance history in a dashboard.

 

Software architecture of the high-level control of FRIDA

César Guzmán-Alvarez¹², Marta Aguiar¹, José Acosta¹, Jesús Patrón¹, Almudena Prieto¹
Instituto de Astrofisica de Canarias, Tenerife, Spain¹, Universidad Abierta y a Distancia, Bogota, Colombia²

This paper appears in: Journal of Astronomical telescopes, Instruments, and Systems.

Impact Factor: 2.688

Issue Date: Ene 23, 2019

Software architecture of the high-level control of FRIDA

César Guzmán-Alvarez, José Marco, Heidy Moreno, José Acosta, Jesús Patrón, Almudena Prieto
Instituto de Astrofisica de Canarias

This paper appears in: SPIE Astronomical Telescope + Instrumentation. Conference paper. Austin, Texas, USA

Issue Date: Jun 10-15, 2018.

FRIDA Instrument Library: the software architecture to execute and coordinate observing sequences

César Guzmán, José Marco, Heidy Moreno, José Acosta, Jesús Patrón, Almudena Prieto
Instituto de Astrofisica de Canarias, Tenerife, Spain

This paper appears in: SPIE Astronomical Telescope + Instrumentation. Poster session. Austin, Texas, USA;

Issue Date:  Jun 10-15, 2018