The mission of this 5-day course is to equip high school students who already have Python skills with foundational knowledge and hands-on experience in artificial intelligence, machine learning, and deep learning—using Python and PyTorch—to analyze scientific data, process images, and extract insights from scientific text.
The syllabus of the afternoon lecture and hands-on sessions from 1pm to 4pm is listed below.
Day | Module Title | Key Topics | Learning Objectives | Lecture (1.5hr) | Hands-on (1.5hr) |
---|---|---|---|---|---|
1 | AI Foundations, Python & Data Science Review | What is AI? Python basics, Pandas, Matplotlib | Understand AI concepts and real-world applications; Review Python syntax; Manipulate and visualize scientific data | Overview of AI in science/engineering; Python recap (variables, loops, functions); Data handling and visualization with Pandas/Matplotlib | Clean and analyze a science dataset (e.g., astronomy, physics) and plot results using Pandas and Matplotlib |
2 | Machine Learning & Deep Learning Basics | ML vs. DL, Linear Regression, Neural Networks, PyTorch | Distinguish ML from DL; Train a simple neural network using PyTorch | Introduction to ML algorithms and deep learning; Linear regression vs. neural networks; PyTorch basics | Build and train a simple neural network for regression/classification on a science dataset using PyTorch |
3 | Convolutional Neural Networks (CNNs) for Image Processing | CNNs, feature detection, image classification | Understand how CNNs extract features (edges, shapes); Build an image classifier for scientific images | CNN architecture and scientific applications (e.g., microscopy, astronomy); Feature maps and filters | Train a CNN to classify scientific images (e.g., hand-written digits, cell images) |
4 | Recurrent Neural Networks (RNNs), NLP, and Multimodal AI | RNNs, LSTM, time series, sentiment analysis, GPT/ChatGPT | Use LSTM to analyze time series data (e.g., stock prices, sensor data); build a sentiment analysis model; explore text and image generation | RNN/LSTM concepts, time series in science, introduction to NLP and multimodal AI | Predict scientific time series (e.g., monthly measurements); build a basic sentiment classifier; experiment with GPT/ChatGPT for text/image generation |
5 | Natural Language Processing (NLP) in Science & Engineering | NLP fundamentals, text preprocessing, sentiment analysis | Learn how computers process language; Build a basic NLP application for scientific text | Introduction to NLP and its applications; Text preprocessing (tokenization, stop words); Sentiment analysis with Python (NLTK or spaCy) | Analyze scientific abstracts or reviews: preprocess text, extract features, and build a sentiment classifier |