2026-08: Applied AI for Professionals (Webinar 3)
Section outline
-
Applied AI for Professionals
A five-session introduction to how AI/ML systems are actually built β the concepts behind them, the data that sets their ceiling, the lifecycle around them, and the methods that explain what they decided.
Learning Goals:
- Place machine learning and deep learning within AI, tell the five types of learning apart by what each one needs from your data, and recognize the main computer vision and NLP tasks by what they take in and return
- Judge a dataset before trusting a model: where training data comes from, how the label unit changes across tables, text, images, audio, video, time series, satellite tiles and graphs, and why annotator agreement and the split key decide what your numbers mean
- Follow a project through the eight lifecycle stages and two loops that turn at different speeds, and set the success criterion, the baseline, and the deployment exposure level before the model exists
- Run a classical ML pipeline end to end β problem definition, exploratory analysis, cleaning, feature engineering, splitting, algorithm choice, evaluation, hyperparameter tuning, and deployment with monitoring β and know why classical methods still win on tabular data
- Explain a single prediction using LIME, SHAP, and DiCE, and tell intrinsic from post-hoc and local from global explanations when choosing a method
Participants may complete the materials at their own pace. To access the workshop materials, please complete your enrollment through the link below:
β Register/Enrol in this Course
Participants may also explore and enroll in the other installments of the DOST-NAIRA AIgnite Academy National AI Capacity Building Program, including courses on AI fundamentals, AI project design, AI application development, and advanced AI for professionals:
βΒ View Previous and Available Courses
To receive your Certificate of Participation, please complete the Pre-Test, Post-Test, and Training/Workshop Evaluation Form.
-
Before beginning the workshop sessions, please take the Pre-Test to gauge your current understanding of the topics to be covered and establish a baseline for measuring your learning progress.
-
Technical Introduction to AI
The vocabulary the rest of the programme assumes: how AI, machine learning and deep learning nest inside one another, what each type of learning needs from your data, and what the common vision and language tasks take in and give back. Ends on DIMER, the national platform where models, datasets and compute are hosted.
β Slide Deck: Technical Introduction to AI
Activity: DIMER Account Creation and Inference Service Demonstration
Download the activity document through this link
https://tinyurl.com/ASTI-NAIRA-DIMER-260827-Actv
Follow the instructions and place your answers in the documentΒ
Submit (upload) your accomplished document through this link:
-
Data for AI
Where training data comes from, what counts as a label in each kind of data β tables, text, images, audio, video, time series, satellite tiles, graphs and documents β and why annotator disagreement, a careless split key or a quiet leak settles your results before the model gets a say. The through-line: your data sets a ceiling no model can climb over.
β Slide Deck: Data for AI
-
The AI Lifecycle
Everything that surrounds the model: the eight stages a project passes through, the two loops that turn at very different speeds, and the decisions β success criterion, baseline, how much exposure a deployment gets β that have to be made before anyone trains anything. Closes on monitoring, where only one of the four layers tells you the system is actually working.
β Slide Deck: The AI Lifecycle
-
Classical ML Pipeline
The end-to-end build for structured data: defining the target, exploring and cleaning, engineering features, splitting honestly, then choosing among logistic regression, k-NN, decision trees, random forests and SVMs, and tuning what you picked. Includes why classical methods still beat deep learning on tabular problems β less data, faster training, and a model you can read.
β Slide Deck: Classical Machine Learning Pipeline
Demo: California Housing Machine Learning Pipeline
β https://colab.research.google.com/drive/1hF_FY8SIdZDzgnMuOjwmsTzTxvzQ3Pb1
-
Explainable AI
Methods for answering βwhy did the model decide that?β β LIME for one prediction at a time, SHAP for contributions that add up to the result, DiCE for the counterfactual that would have changed the outcome, and model-specific techniques for vision and time series. Frames each choice by whether you need a local or a global answer, and whether the model lets you look inside.
β Slide Deck: Explainable AI
XAI Demo
β Introduction to Explainable AI, LIME, and SHAP (Tabular Data)
β Counterfactual Explanations with DiCE
β Explainability for Deep Learning: Grad-CAM for Computer Vision
-
After completing the workshop sessions, please take the Post-Test to assess your understanding of the topics covered and measure your learning progress.
-
-
Please complete the Training/Workshop Evaluation Form to share your feedback on the sessions, materials, and overall learning experience. Your responses will help improve future training activities.
-
Training/Workshop Evaluation Form FeedbackNot available unless: You have a grade in Post-Test
-
-
To receive your Certificate of Completion, complete the Pre-Test, Post-Test, and Training/Workshop Evaluation Form.
Name Corrections for Certificates
If your name needs to be corrected, please update your Moodle profile name. To do this, go to your profile, select Edit profile, and make sure your first name and surname are spelled exactly as you would like them to appear on your certificate.
After updating your profile, please return to Certificate of Participation and download your certificate again.
-
Before you go: a short survey (optional)
Across these sessions you saw what a working AI system actually needs: GPUs to train on, datasets and model versions that someone has documented and kept, a registry that can say which version is running and what produced it, and somewhere to serve the model once it leaves the notebook. Each session covered a different piece of that stack.
We are studying whether a national AI-as-a-Service platform could bring these together for researchers, developers and students in the Philippines. As proposed, it would offer a searchable repository of documented datasets and models, with hosted inference and controlled fine-tuning planned. It is a proposal, not yet an available service: capacity, support and pricing are not decided.
If you would like to take part, the survey is at study.dimer.dev or aiaas-market-study-st8x.vercel.app.