Supplementary Figure 1: Schematic illustration of the hierarchical model of the value inference process based on subjective ratings.

For an experimental data set consisting of M goods and N value ratings for each good, we can find the set of parameters of the prior, the internal valuation noise σ, external noise \(\sigma _{{\mathrm{ext}}}\), and the ‘true’ stimulus values \(v_{(1, \cdots ,M)}\) that maximize the likelihood of the observed set of ratings under the constraint that \(v_{(1, \cdots ,M)}\) is distributed following p(v). In our experiments, we parameterized the prior using a logistic distribution (see Methods), however, any other parametrization is possible. Note that the parameters of the prior also constrain the likelihood.