search

Houdanait / PoliticalTextandAttitudes

0 stars 0 forks source link

Define the research question #1

Closed Houdanait closed 6 months ago

Houdanait commented 6 months ago

The goal of this issue is to define the research question.

matthewcwise commented 6 months ago

Potential topics that have come to mind (still lit reviewing, suspect many of these have been explored):

  1. Defining "center"--how the middle/party alignment has shifted over time
  2. Is Donald trump the symptom or the cause? Does he reflect or drive the ideological shifts that are happening in American politics? What does a timeline look like? Does increasingly polarizing language coincide with rise of social media?
  3. Language of dictators--comparing political speeches over time, what does an early Hitler speech sound like compared with current leaders
  4. Comparisons of political views and media/movies--how much does entertainment reflect/change political discussion
  5. Do sentence embedding models (eg Voyage AI) outperform other language models on political topics?
  6. "neutralization of text" - can we take a text and make it more politically neutral?
matthewcwise commented 6 months ago

Can we predict political trends using language (super vague idea). Like the umbrella movement in China, is there something linguistic that could be done here?

matthewcwise commented 6 months ago

Maybe something on the more positive side of exploring polarization. Something like helping find common ground between ideological enemies?

matthewcwise commented 6 months ago

given the complexities involved with Arabic dialects, would there be a political research topic we could explore that involves some interesting linguistic complexity?

On a separate note, Prof. Manning mentioned unsupervised translation, which sounds like a fascinating idea to explore https://proceedings.neurips.cc/paper_files/paper/2023/file/da31f4275972a58406b95c277ce7bc8d-Paper-Conference.pdf

Houdanait commented 6 months ago

With @matthewcwise we agreed on the following research question:

Compare the language of various "extremists" politicians in different countries (and hence different languages) and see how more/less extremists they are related to one another.