AI lesson draft!

[retabak]

Aiming to draft lessons 2 and 4 by the end of June; aiming to draft lessons 5 and 6 by the end of July.

Lesson 1: Data Driven Algorithms

• [x] Part 1: Students think about spell checkers, considering how their programmed behavior is both similar and different to that of humans.
• [x] Part 2: Artificial intelligence relies on data-driven algorithms, which produce a higher quality output when we provide more data, without needing to change the code.

Lesson 2: Training AI [plagiarism detector]

• [x] We have corpuses which can be quite large (in our case, corpuses from the internet – we’re using stuff in the public domain, which is legal. ChatGPT trains on internet data)
• [x] We often want approximate answers, not perfect answers. (when students plagiarize, they don’t turn in the same text). We don’t want an overprecise program.
• [x] We can think of plagiarism detection as AI. Training phase + sampling. If you tried to redo the training on the fly every time, it would take forever.

Lesson 3: Decision Trees [supervised, discrete]

• [x] Part 1: Students are introduced to scenarios in which decision trees are used, and then play a game to help them build a frame of reference.
• [x] Part 2: Students build a decision tree from a dataset that predicts whether different individuals will purchase a video game or not.

Lesson 4: Regression [supervised, continuous]:

• [x] Part 1: Students learn about ALVINN (autonomous Land Vehicle in a Neural Network) as an example of supervised machine learning and regression.
• [ ] Part 2: Students explore regression to predict….

Lesson 5: Clustering [unsupervised]

• [ ] Shrink down Shriram’s chapter / do this last

Lesson 6: Statistical Language Models Soekia

• [x] Part 1: Intro to SLMs
• [x] Part 2: Soekia

[retabak]

see if I can integrate issue #1335

[retabak]

Useful links:

• Here are notes from today's meeting!
• Here's an article you should consider reading. It's written by an actual AI prof (Mark Riedl), so he fully knows what he's talking about, and tries its best to break down LLMs into something hopefully understandable: https://mark-riedl.medium.com/a-very-gentle-introduction-to-large-language-models-without-the-hype-5f67941fa59e
• This NYT piece may also be helpful: https://www.nytimes.com/interactive/2023/04/26/upshot/gpt-from-scratch.html
• Video from the meeting
• Shriram's brain-dump about the lesson
• https://www.nctm.org/standards-and-positions/Position-Statements/Artificial-Intelligence-and-Mathematics-Teaching/

[flannery-denny]

@retabak Gonna collect statistical language learning dataset ideas here - expect a few more to roll in

• Parameters -- 10 to 30 lines -- levity - since we're calculating probability from the text - something to serious or thought provoking would be a distraction -- As repetitive a text as There was an Old Lady who swallowed a fly, but hopefully more culturally relevant
• What a Wonderful World lyrics
• De Colores - we could pick and choose which verses. Actually seems fine to use lyrics in Spanish for the purpose of calculating probability!
• The Tide is Rising - click the first youtube link - we could pick whichever verses seem most innocuous.
• Emily likes the TV
• Visit to St. Elizabeth's
• Shel Silverstein - Listen to the Mustn'ts from Where the Sidewalk ends - comes to mind, althought it's probably too short

  Listen to the mustn'ts, child Listen to the don'ts Listen to the Shouldn'ts The impossibles, the won'ts Listen to the Neverhaves, Then listen close to me Anything can happen, child Anything can be.

• Zulu folk song -

Siyahamb' ekukhanyeni kwenkos, Siyahamb' ekukhanyeni kwenkos'. Siyahamb' ekukhanyeni kwenkos', Siyahamb' ekukhanyeni kwenkos'. Siyahamba... ooh Siyahamba Siyahamba ekukhanyeni kwenkos'. Siyahamba... ooh Siyahamba Siyahamba ekukhanyeni kwenkos'.

• This little light of mine

This little light of mine I’m going to let it shine Oh, this little light of mine I’m going to let it shine

This little light of mine I’m going to let it shine Let it shine, all the time, let it shine

All around the neighborhood I’m going to let it shine All around the neighborhood I’m going to let it shine All around the neighborhood

I’m going to let it shine Let it shine, all the time, let it shine.

• Miss Mary Mack

Miss Mary Mack, Mack, Mack All dressed in black, black, black With silver buttons, buttons, buttons All down her back, back, back She asked her mother, mother, mother For 15 cents, cents, cents To see the elephants, elephants, elephants Jump over the fence, fence, fence They jumped so high, high, high They reached the sky, sky, sky And didn't come back, back, back Till the 4th of July ly ly

[flannery-denny]

@retabak now that you've made a. new issue for lingering todos can we close this and delete the milestone?