Images, Art & Video (E3) - Naik, Kominers, Raskar, Glaeser, Hidalgo 2017

[HyunkuKwon]

Naik, Nikhil, Scott Duke Kominers, Ramesh Raskar, Edward L. Glaeser, César A. Hidalgo. 2017. “Computer vision uncovers predictors of physical urban change.” PNAS 114(29):7571–7576.

[Raychanan]

After reading this paper, I wondered whether the relationship between urban change and suburban change could be explored by including suburban areas near cities in the study. As a city initially grows, people may begin to congregate from the suburbs to the city, which leads to urban prosperity and suburban decline. But as a city grows further, people who become wealthy begin to resent the crowded traffic, noisy environment, or overly bustling crowds of the city. As a result, they may choose to move from the city to the suburbs. The city begins to decline. So, when we include the suburbs around a city in our consideration of urban change, the so-called decline may just be something that is happening in that small area. But I wonder if the Google Street View data collection for the suburbs is complete? If it is not complete then this study may be difficult to conduct?

[yushiouwillylin]

For most of the paper we see in this course, it seems that we still need some sort of human-decision (survey or quantitative measurements) to make credible inferences. Do you think there will ever exist a machine that will decide everything. where there exist a consensus that it is better than human, and we will abandon human-decision in daily life or academic fields?

[william-wei-zhu]

In different countries, cultures and economic policies affect the growth and decline of urban environment in different ways. Can the findings of this paper be generalizable across cities in other countries?

[jcvotava]

What work here is the dimension of the analysis focusing on physical attractiveness actually performing? It seems to me like all the pointed conclusions of the paper are pretty much using physical attractiveness as a proxy - but it's not clear that the underlying variable, i.e. good infrastructure in a neighborhood, lacks for proxies. In what way is the focus on attractiveness helpful?

[chiayunc]

I think this paper brings about the fundamental question of how much do/should our unsupervised algorithm rely on human assessment. The human assessment portion used in this paper is significant, and with just 5 cities, we see a lot of labour input. In this case, how can we lend forces from unsupervised learning more and potentially scale up this type of research?

[mingtao-gao]

Thanks for sharing this interesting paper. My question is related to the Street Score. Particularly, the authors mention that the "The Street change metric is not affected by seasonal and weather change", which I personally think that it may have some difficulties on the image identifications. I am also interested in the application to quantify neighborhood appearance as mentioned by the author.