I'm moved to take out an idea of mine from storage after 20+ years:
Suggest that a #MachineLearning system, processing #LLM (large language models) like #ChatGPT could build a model of human values (hierarchical & fine-grained) and their expression within various contexts, and then—derive a metric for the hierarchical coherence of identified values in terms of #MutualInformation (not just simple co-occurence, but including contexts within contexts over time.)
#socialdecisionmaking #mutualinformation #chatgpt #llm #machinelearning
@ldebowsk Adding variants of the #HashTags ...
#MutualInformation
#InformationTheory
#StochasticProcess
#ExcessEntropy
#KolmogorovComplexity
#AlgorithmicRandomness
#LargeLanguageModel
#NeuralNetwork
#PowerLaw
#UniversalCoding
(Just to help discoverability)
#universalcoding #powerlaw #neuralnetwork #largelanguagemodel #algorithmicrandomness #KolmogorovComplexity #excessentropy #stochasticprocess #informationtheory #mutualinformation #hashtags