Entry-Level Computer Vision Datasets

Preflection prompts:

• What do you want your model to be able to do?
• How can you imagine your model (or an extension of it) being used in the real world? Feel free to get creative.
• If those ideas came true, who might they affect? In what ways?
• What measures would your model's real-world implementers need to take to ensure its effects live up to your intentions?
• What pitfalls might your model fall into, and what could you quantitively measure to avoid those?
• Why was the dataset you used to train your model created?
• In what other ways could the same data be used? How do you feel about those possibilities?
• How was that dataset assembled? From where was the data sourced? Who or what labeled it? If there are any elements of this process you think were either particularly well done or problematic, how so?
• If your dataset contains information about people, to what degree did those people have agency over their inclusion? Do you feel that matters in this case? Why or why not?
• Skimming through your dataset, does anything stand out to you about representation in its contents? // DO MORE QUANTITATIVELY W/ TOOLS
• Do you feel the potential use cases for this dataset justify it being created and published? Why or why not?