Extensions and Related Results

Statistical Learning Model of the Sense of Agency

“…While the product rule of the probability is definitely the well-known mathematical foundation of Bayes' theorem (Griffiths et al, 2008), there are at least three theoretical foundations known as Kullback's principle of minimum cross-entropy (MINXENT), i.e., the principle of minimum discrimination information (Shore and Johnson, 1980;Shu-Cherng and Tsao, 2001;Rao, 2011;Halpern, 2017, Chapter 3), the information conservation principle (Zellner, 1988(Zellner, , 2002Soofi, 2000), and the mirror descent algorithm (Warmuth, 2006;Dai et al, 2016). As shown in Fang et al (1997), the Bayesian inference can be derived from the MINXENT. The other important inference rules such as the maximum entropy principle (MAXENT) can be derived from the MINXENT (Shu-Cherng and Tsao, 2001).…”