Daniel Whiteson
@DanielWhiteson
Followers
8K
Following
106K
Media
1K
Statuses
15K
Particle physics and ML (https://t.co/1MlHzNIbD9). Co-creator @DandKUniverse pod and @ElinorWonders TV. Patreon: https://t.co/a9krC8OAce.
Joined March 2014
It's been a pleasure to host D&J Explain the Universe for the last 6 years. It's been a great run and I'm thankful to all of the support from listeners. There are so many joyfully curious people out there, and it was wonderful to get to know all of you.
11
5
63
RT @msmathcomp: One of my favorite podcasts chimes in on #illustratingMath in this episode, discussing the #geometry of honeycombs in the a….
iheart.com
0
1
0
RT @DanielWhiteson: New paper!. On Focusing Statistical Power for Searches and Measurements in Particle Physics. .
arxiv.org
Particle physics experiments rely on the (generalised) likelihood ratio test (LRT) for searches and measurements, which consist of composite hypothesis tests. However, this test is not guaranteed...
0
7
0
I taught remedial math at San Quentin. I've never had more polite, motivated students. They know the value of an education.
0
0
6
RT @Samuel_Gregson: For those who want an appropriate take on this matter from people that actually know what they’re talking about:. https….
sites.psu.edu
Avi Loeb has gained a reputation for suggesting and outright claiming things could be alien spacecraft even when there’s little to no objective reason to think so. So when the third interstellar...
0
6
0
RT @Aishik_Ghosh_: What if you could beat the Likelihood Ratio Test?. A fun and potentially transformational phy-statistics paper with Ann….
arxiv.org
Particle physics experiments rely on the (generalised) likelihood ratio test (LRT) for searches and measurements, which consist of composite hypothesis tests. However, this test is not guaranteed...
0
4
0
Isn’t it computationally expensive to do frequentist statistics? Yes, but we also introduce an ML technique, quantile regression, for fast frequentist intervals, even in high dimensions.
1
0
1
What if you focus in the wrong place? Is this just bayesian priors? Frequentist intervals guarantee coverage, unlike bayesian approaches, regardless of focus. So it’s not wrong, just not optimal.
1
0
1
Confidence intervals from the FTS are narrower than those from the LR!
1
0
1
What if we could choose where to have greatest sensitivity? In our new paper, we introduce the “focused test statistic”, which can tune its focus to a region of interest.
1
0
1
The LR is still good, but it’s not necc optimal for all values of your parameter. Using the LR implicitly chooses some spread of statistical power. Is it the spread we want?
1
0
1
But the lore is missing a detail. The LR is only optimal for “simple” theories without parameters. If your theory has a parameter, eg masses or couplings, then the NP lemma doesn’t guarantee optimality! And almost every theory has a parameter.
1
0
1
What’s the optimal way to distinguish between two theories? For decades, the lore has been that this problem was solved: the Neyman-Pearson lemma tells us to calculate the likelihood ratio.
1
0
1
New paper!. On Focusing Statistical Power for Searches and Measurements in Particle Physics. Led by @Aishik_Ghosh_ , with our stats friends J. Carzon, R. Izbicki, A. Lee and L. Masserano.
arxiv.org
Particle physics experiments rely on the (generalised) likelihood ratio test (LRT) for searches and measurements, which consist of composite hypothesis tests. However, this test is not guaranteed...
3
7
20
Why doesn't the Oort cloud obscure our view of the Universe?.
1
1
3
RT @DrCatharineY: America is facing an unprecedented brain drain that will “present an extinction-level event for American science.” https:….
0
144
0
