Lexis based Comparative age-period-cohort analysis, LLM inductive and/or deductive reasoning.

Journal Club

Comparative age-period-cohort analysis
Philip S. Rosenberg1*, Adalberto Miranda‐Filho1 and David C. Whiteman2

BMC Medical Research Methodology
https://doi.org/10.1186/s12874-023-02039-8

[preprint]

Hackathon

Using this paper to seed discussion of 

how epiVerse functionalization could be construed: https://www.youtube.com/watch?v=u_xBbk_UgS8

The Reversal Curse: LLMs trained on‘ A is B’ fail to learn‘ B is A’

Lukas Berglund, Meg Tong, Max Kaufmann, Mikita Balesni, Asa Cooper Stickland, Tomasz Korbak, Owain Evans

https://arxiv.org/abs/2309.12288  

Should we expect deductive reasoning at all?
Why do so many people presume it when dealing with LLMs? 

Jeya, please insert the NLP operational language argument if you like it 

Discussions contributed by Jeya:

https://twitter.com/madiator/status/1705376797293183208 https://twitter.com/karpathy/status/1705322159588208782 https://twitter.com/NeelNanda5/status/1705995593657762199 https://twitter.com/iamtrask/status/1705361947141472528 

ARPA-H

start here, connect to innovate and deliver: https://arpa-h.gov/engage/arpanet-h 

then explore here: https://arpa-h.gov/about/about-arpa-h 

making sure to note the data fabric toolbox: https://datacommons.cancer.gov/news/nci-crdc-partners-arpa-h-biomedical-data-fabric-toolbox-program 

ARPA-H Biomedical Data Fabric Toolbox <<<  tech diagram here