Learning How to Play with the Machines
Cassandra Inaugural Roundtable
Panelists
- Amber Boydstun, UC Davis
- Jennifer Pan, Stanford
- Justin Gross, UMass Amherst
- Joshua Tucker, NYU
- Noah Smith, UW
Chairs
Quotes
The amount of social science work that is I see in computer science has gone up tremendously, and I think it is because of these interactions between the two fields…Computing has become more important in society because we can ask these questions, but we are also better equipped to answer them because of these interactions.
–Noah A. Smith
It is important not to lose track of the fact that not only has computational text analysis helped social science ask the same questions it has always asked but more efficiently, but allowed to ask questions that we would have never been able to ask before.
–Amber E. Boydstun
We can use NLP machine learning as measurement, and this is essentially prediction, not trying to give explanations, but creating auxiliary variables measuring the data and there are lots of ways to do it, lots of ways of human involvement, but essentially these measurements are alternatives to survey methods or other methods that we had before.
–Noah A. Smith
I study primarily Chinese politics, and politics in non-democratic countries, were data is very limited. With these methods, the internet and social media opened up a new arena for research. But then I realized that political science is a lot about causal inference and these descriptive methods have some limits but still NLP and machine learning have been crucial to enabling this kind of research.
–Jennifer Pan
We need to recognize what the collaboration between computer science and political science did for the study of political science. Opening up text to political analysis has simply changed what we can do.
–Josh Tucker
There are still a lot of challenges in the way in which researchers are incentivized. The discipline system that we have in academia is slow to evolve and this is a newly evolving field, and it is not always clear how things fit together.
–Josh Tucker
Early interactions show that the field was primarily driven by information retrieval that shaped the questions that were asked and the research that was done/ could be done. The methods became relevant rather than the ideas.
–Justin H. Gross
One of the challenges of the interaction between computer science and social science is that computational social science involves multiple methods.
–Justin H. Gross