Assignment 7

Reading Response

• How do we get through this whole list of concerns and still build AI that is fun, respectful, tender, pleasurable, kind?

In the reading "The Subtext of a Black Corpus", I found a good response to this question: "This redundant rhetoric needs more heart, more compassion and, before all else, more creative applications that nudge us to be self-reflective as a united community and empower underprivileged voices. " I think developing an involving, reflective community with great passion toward machine learning is essential, a community in which members support each other as well as discover the potential flaw and biases from the machine learning process.

• Given that so many of the existing “big data” language models are trained with Western texts and proprietary datasets, what does it even mean to try to decolonize AI?

Not many people understand how AI, machine learning work, how the algorithm is structured and more importantly, what is the dataset being trained and employed. And we sometimes don't really care about it. We just take the AI as granted without further reflecting on it. Also only a small part of machine learning project are open-sourced and well documented for public to get in touch with. Therefore, I think one important step toward decolonizing AI is to make machine learning less mysterious, more accessible to the public population. Documentation should include how the data is collected, the usage and limitations of the model so that when we use acertain model, we have better knowledge about the potential harms and benefits of the model.

Coding Exercise

For the coding Exercise this week, I decided to implement the sketch RNN model. I really like the random snowflake generator coding train showed so I made a random cats generator and made a few modifications. Instead of generating at a random position on the canvas, I wanted the drawing to be generated along with my mouse; and

Sketch: https://editor.p5js.org/Alicelong/sketches/0oUCLVJms