Workshop: AI with LearningML for FECYT
This post gathers the materials for a LearningML workshop prepared for FECYT. The main goal is to introduce the basic ideas of Machine Learning through hands-on classroom activities so participants can build their own AI projects and develop a more critical understanding of current AI tools.
Practice 1. The imitator
An introductory image-recognition activity in which participants train a model to distinguish facial gestures or accessories and then connect it to a Scratch project.
Tutorial video:
Recommended follow-up resources:
- First steps with Scratch
- Splash it in colour with LearningML
- Programming with Ciro, part 1
- Programming with Ciro, part 2
- Programming with Jara, part 1
- Programming with Jara, part 2
- Prehistory quiz
- Sound recognition
Practice 2. Rules vs data: two ways of solving problems
This practice uses points in the Cartesian plane to compare rule-based programming with data-driven Machine Learning.
Resource:
Video:
Practice 3. Dogs, cats, and bias in Machine Learning
Participants explore how biased datasets affect model performance and how retraining with more representative examples can improve results.
Resource:
Practice 4. Text encoding (text embeddings)
This section introduces the idea that text, images, audio, and other inputs must be encoded as numbers before a Machine Learning algorithm can process them.
The full workshop material is currently written in Spanish in the paired post. This English version exists to keep the bilingual post structure consistent and make the workshop easier to discover from the English blog.