Justin Rising

[Justin Rising @JustinRising]

@matt_levine Scott Aaronson (leading QC theorist) doesn't believe it: https://t.co/AaySIQMHBj

[Justin Rising @JustinRising]

@MishaTeplitskiy I think that's a different phenomenon but there's certainly interaction.

[Justin Rising @JustinRising]

@PhDemetri Read Bernstein's "Against the Gods". Although it doesn't directly answer this question, none of what the book describes is possible without statistics.

[Justin Rising @JustinRising]

@Mike_Kaminsky It's the compute time. If I need to make an estimate every day from a large data set, there's no way I can wait for a sampler to run.

[Justin Rising @JustinRising]

@karlrohe I wrote a slightly longer version of this at https://t.co/8Lzv87LTND (but in terms of unsupervised learning rather than multivariate analysis).

[Justin Rising @JustinRising]

@karlrohe For a multivariate analysis, you don't have any notion of accuracy, but you do have some notion of description length. That opens the door to the same information-theoretic tools that have been so useful in analyzing the predictive setting.

[Justin Rising @JustinRising]

@alex_peys For binary classification, looking at paired differences has a pretty high false discovery rate. Those issues may be present in LLM evaluations as well. See T.G. Dietterich, "Approximate statistical tests for comparing supervised classification learning algorithms".

[Justin Rising @JustinRising]

@codydroux PSPACE = IP so if NP = IP we live in a world where being able to ask any finite number of questions doesn't buy you much more than being able to ask one. That's hard to swallow.

[Justin Rising @JustinRising]

@cschlom @ben_golub There are some asymptotic results. See https://t.co/QGO3PUNtEG and references therein.

[Justin Rising @JustinRising]

@alexkyllo It's not at all remedial but take a look at Percival & Walden "Spectral Analysis for Univariate Time Series".

[Justin Rising @JustinRising]

@kareem_carr If a distribution has a finite second moment, the difference between the mean and the median is at most one standard deviation. So an upper bound can be derived from Chebyshev's inequality.

[Justin Rising @JustinRising]

@punished_picnic https://t.co/ZYcOUVmjHM

[Justin Rising @JustinRising]

@PhDemetri I know one guy who was transferring out of his first program and ended up doing enough to get the degree during that last year.

[Justin Rising @JustinRising]

@_julesh_ Interactive proofs as cooperative games?

[Justin Rising @JustinRising]

@ZachWeiner I'd point them to https://t.co/XrGwzMAtUa

[Justin Rising @JustinRising]

@MaartenvSmeden Is there a good survey paper?

[Justin Rising @JustinRising]

@sigfpe It's not exactly the same thing but this is very close to the complexity class RP.

[Justin Rising @JustinRising]

@ben_golub I like Jean Gallier’s book.