Bayesian Properties of Normalized Maximum Likelihood and its Fast Computation · HackerPapers discussions

Bayesian Properties of Normalized Maximum Likelihood and its Fast Computation — mentioned by skat20phys

Mon, 24 Feb 2020 16:57:54 GMT

Negative probabilities interestingly are showing up in various other realms. There seems to be converging interest in the topic from a number of areas. For example, in relating Bayesian inference to algorithmic complexity-based inference: https://arxiv.org/abs/1401.7116 I've also read some interpretations of negative probabilities in terms of erroneous values. For example, if you're sampling observations, and a machine error causes an anomalous value to appear in the data. This I don't know I completely intuit but there are interpretations of negative probabilities outside of QM. My sense is…

Paper and discussion history. A mention is not an endorsement.

Bayesian Properties of Normalized Maximum Likelihood and its Fast Computation — mentioned by ta1929901

Fri, 25 Aug 2017 23:48:08 GMT

A lot of what they discuss is in the literature on reference priors, if not in other literature on objective priors as well. It's a little complex for a comment on HN, but IMHO the best formalization of overfitting is in the literature on minimum description length, and related information-theoretic literature (https://en.wikipedia.org/wiki/Minimum_description_length ; the wikipedia page is a little off on some things but the general points are probably about right). The relationship between MDL/NML and Bayesian statistics is a little complicated--Barron, Roos, and Watanabe have a nice recent…

Paper and discussion history. A mention is not an endorsement.