🎯 In this lesson, you will:
- Understand what natural language processing (NLP) is and how machines understand text
- Explore real-world tools that use NLP, like translators, chatbots, and emotion detectors
- Experiment with neural networks and tweak parameters like layers and learning rate
📂 Resources
👉 Week 2 Slides (update link)
👉 Week 2 Worksheet (Make a Copy) (update link)
🎹 Starter Activity: Make a 5-Word Sentence
Tool: Word Synth
- Type a short sentence (e.g. “My cat plays the piano”).
- Listen to how the machine “speaks”.
- It helps machines read, write, speak, and make sense of our words.
- It’s used in tools like Google Translate, autocorrect, chatbots, and voice assistants.
🧠 Word Synth uses NLP and speech synthesis to turn your sentence into sound — trained on real speech to sound realistic.
Dr. Dolittle talks to animals if you haven’t seen the film!
📚 Theory: What is NLP?
Natural Language Processing (NLP) is how computers understand, interpret, and generate human language.
It allows machines to read, write, and respond in ways that sound (mostly) natural.
👀 Real-world examples of NLP tools:
- Emotion Detector – Try it: How does it “know” what you're feeling?
- Named Entity Recognition – Try it: See how machines find names, places, and dates in text
- Text Completion – Try it: Just like the tech behind ChatGPT and autocorrect
🧩 These tools all use neural networks to spot patterns in huge amounts of language data.
✨Task: Magic Text Transformations
- Use Google Translate
- Start with a simple sentence in English.
- Translate it through multiple languages (e.g., English → French → Japanese → Spanish → English).
- Read the final result — is it the same? What changed?
💬 Discussion prompts:
- What changed?
- What stayed the same?
- Why do you think the meaning sometimes shifts?
🧠 This activity shows how machines sometimes “misunderstand” - a key challenge in NLP.
🤔 Theory: What is a Neural Network?
A neural network is a kind of algorithm inspired by the human brain.
It spots patterns in data (like words or sounds) and makes predictions based on what it’s learned.
📌 In NLP, neural networks help with:
- Predicting the next word in a sentence
- Translating between languages
- Analysing tone or mood
- Answering questions (like ChatGPT!)
🗺️Task: Neural Network Playground
Tool: TensorFlow Playground
🎯 Try adjusting the following:
- Number of layers
- Number of neurons
- Learning rate
📊 Think about:
- How does performance change?
- Does the model learn faster or slower?
- Can it separate the two classes correctly?
This task connects to the idea of training models on language data — and lets you tweak the kind of networks that power real NLP systems.
extension - make a multi instrument loop on word synth
⌚ Just a Minute!
NLP lets machines speak, listen, and understand us (but it's far from perfect!). What did you notice today that surprised you?
Discuss in pairs/small groups.