Conference Records 2024-02-08:

Interesting sessions

Session name: Computational psychology 1

Details and Comments

• Finding theory to bind diverse fields → Abstract → mathematical fundamental matters
• Example: $B=f(P,E)$ not precisely but quite abstract
• There are many ways to define algorithm to help to find solutions

Session name: Computational psychology 2

Details and Comments

• how to use models like LLMs
• Check https://scholar.google.com/citations?hl=en&user=byZIaXsAAAAJ&view_op=list_works&sortby=pubdate
• annotations collected for participants → add it into pre-trained model → biased
• Comments: I am not quite familiar with LLMs, so I listed some articles I would like to check after the conference.

Session name: Computational psychology 3

Details and Comments

• finding/compiling/merging(CAT) data
• Abductive core of science
• Beyond WEIRD psychology: using $F_{ST}$ to define the distance
• Use causal inference → view psychology as historical science

Log

I asked Michael why define the cultural distance with $F_{ST}$, he suggested that the $F_{ST}$ considers all information and could summarize them into one simple index. I thought here might be a better way to dig the quality of high-dimensional data, so I told him some method in topology like Persistent Homology might be useful. But at that time, I was not in a good condition and could not introduce the method for Micheal precisely. Thus, I am now writing a short introduction of persistent homology and my thoughts about how to apply it in the field of psychology.

Session name: Computational psychology 4

Details and Comments

• Collect data from undergraduates' essays and rating them with LLM
• Create the computational rating of "creativity" with different pre-trained models
• Results: rating of "creativity" is highly correlated with SAT scores → college admissions should pay attention to creativity

Session name: Computational psychology 5

Details and Comments

• Experiment: give participants some political agenda and let them write down their opinions → free discussion with other participants (multiple phases) → interpersonal attraction rating
• Define distance of political ideology pairwise, he used three types of distance (I forgot specific types except the Euclidean one) but no difference on results.
• Results: distance of political ideology shortened after multiple phases of discussion between the pair of participants who liked each other (visualized with PCA)
• Comments: Federico did not publish the results above yet, I think his research flow could be one of replicable template to apply. I'm looking forward to his publications.

Session name: Computational psychology 7

Details and Comments

• All attendances are grouped by multiple teams and free discussed three topics:
  • What unifies researchers to computational psychology? (motivation)
  • How could computational methods benefit the social psychology?
  • The barriers of bringing computational methods into the field of psychology?
• Consistent sense
  • Let science be more science in the field of psychology, be open-minded but do not regard "Models" as the best option to conduct all researches
  • Call for handy tutorials and systematic pedagogy
  • Let human resource flow, and assemble researchers from other fields

Daily Summary

What's new

• People who are using new techs such like AI and other common methods from computer science to do psychology research are calling themselves as computational psychologists.
• Some researchers have done the studies which contain well-established methodology, and some even share samples of introducing new mathematics corresponded to traditional social psychological topics.
• I'm so glad that I could discuss with Flery and other professors at the ending part of pre-conference. They are seriously considering how to define "Computational Psychology" and bring more researchers from other fields. I also expressed my concerned that it might be a little hard to call researchers from other STEM fields because we might share the same research topic, however, we do not share the same motivation and research flow most of the time.

What should I check

References

Muthukrishna et al., 2020, Beyond Western, Educated, Industrial, Rich, and Democratic (WEIRD) Psychology: Measuring and Mapping Scales of Cultural and Psychological Distance. https://pubmed.ncbi.nlm.nih.gov/32437234/
Zimmerman et al., 2022, Political coherence and certainty as drivers of interpersonal liking over and above similarity. https://www.science.org/doi/10.1126/sciadv.abk1909