Week 1

Attempt to train an image classifier that performs well in differentiating between a NYU campus ID card and a smartphone (both assumed to be in a person’s hand). Share the link to the your model in #ml4artdesign-f23 as well.

Teachable Machine Link: https://teachablemachine.withgoogle.com/models/nZO5OW_tV/

P5 Link: https://editor.p5js.org/WangZhewei1027/sketches/13TNqQOvP

Try using a classifier one of your colleagues trained. How well is this working? (You might want to look at the confidence property in the result for this.) What could prevent this from working better?

I tried some of my classmate’s work. I find they worked well, even better than my own left hand and right hand model (which doesn’t work when the background changed). The confidence almost reached 100%. I think the reason why the model works well is that the different between the NYU card and a phone is very obvious. The NYU card is bright purple while most of the phones are black. So the model works well telling the difference between the NYU card and the phone.

Read Hundreds of AI tools have been built to catch covid. None of them helped. Explain one of the reasons why the technology failed that is mentioned in the article.

One reason is that in the middle of the pandemic, the researchers didn’t have enough high-quality data to train the AI model. The lack of data also means the lack of diversity in the data. The AI model may learn some features unrelated to covid like age, position, or hospital front because of the lack of diverse data or mislabeling.

Read Are we witnessing the dawn of post-theory science?

An interesting article. The article shows that though nowadays AI can effectively find the relationships and patterns in a massive data set, it is accuracy is still questioned. Further more, when it comes to post-theory researching which requires reasoning, it is still incapable of doing that. But the prospect is positive, with the development of AI technology.

Next week’s class will be all about tracking bodies with machine learning. As a warm-up, watch this series of videos compiled by artist Memo Akten (it’s him in the first video btw). What were some of the most impactful AI video filters you have encountered, or heard of?

I knew the getting older one and getting younger one, which are very popular among elderly and young people.

If you’re into contemporary art, you might also want to get a glimpse of: Hito Steyerl, How Not to be Seen: A Fucking Didactic Educational .MOV File (2013)

So abstract🥲 The video itself is a kind of post-modern art.