Our key argument is that we shouldn't discount human insight. Psychologically informed decision-making can create robust, nuanced decision support systems by using *informed feature extraction*, *informed priors*, and & *informed data collection*
Comments
Log in with your Bluesky account to leave a comment
To state the obvious -- yes, black-box ML is awesome, LLMs are great, and we do not discourage anyone from using them (I use them all the time). But there is much to gain in transparency, interpretability, and accuracy when integrating theory into our prediction models ✌️
Comments