Every week, I’ll post useful material to accompany the lecture notes that can help you get a deeper understanding of the material covered in the lectures if you’re interested — the lecture notes should be self-contained and sufficient to follow the course. Under the rubric of supplemental reading, you’ll find other material that supplements various details of the material covered in the lectures. You can view these as references to the core material of the course. Under the rubric of advanced reading, you can find links to original references, research articles, links to talks etc for a more in depth discussion of some of the concepts that came up this week. If I come across interesting articles and links that could inspire potential projects, I will place them here (and will be continually updated as the course progresses, so do keep checking!) Under the rubric Entertainment and culture you will occasionally find weekend newspaper like supplements that touch upon some of the concepts we’ve covered selected to entertain and enlighten in equal parts.
I) Although primarily intended for cosmologists, an excellent overview to many of the concepts covered in the lectures regarding probability theory and statistical inference can be found in these lecture notes by Licia Verde at the University of Barcelona. You can simply skip or skim over any specific references to observational cosmology in these notes, as they’re only intended to provide concrete examples.
II) Information theory for intelligent people — a delightful primer by Simon DeDeo at the Santa Fe Institute. It’s a short read, but provides an intuitive introduction to the foundational concepts of information theory.
III) Bayesian reasoning for intelligent people — another gem by Simon DeDeo.
I) Three Approaches to the Definition of the Notion of Amount of Information by A. N. Kolmogorov (should be accessible via your institute’s remote library proxy)
II) A Mathematical Theory of Communication by Claude Shannon (hosted by Harvard University Mathematics Department)
Entertainment and culture: A long peace, or fooled by randomness?
There was a highly entertaining academic feud that played out in public a few years ago between Steven Pinker, the Harvard Psychologist and best selling author, and Nassim Nicholas Taleb, a trader also turned best settling author (who’s most famous book is perhaps The Black Swan). One the one hand, Pinker claimed to have pored over lots of data across the continents over a large span of history to conclude that the amount of violence in this world has been steadily decreasing, and that the 20’th century (in spite of it’s wars and violent revolutions) was the most peaceful on record. Specifically, the probability of dying a violent death today is lower than it has ever been in human history. Taleb argues that Pinker has not only cherry picked his way to his conclusion, but has evidently been making a fuss over nothing more than a statistical fluke.
First, a summary and review of Pinker’s book The Better Angels of our Nature that kicked it all off.
Next, a critique by conservative intellectual historian and philosopher John Gray that tries to pick apart some of Pinker’s assumptions.
In steps Nicholas Taleb with his statistian hat on and his gloves off.
Pinker’s response can be found here (the titles going back and forth are gold.)
Noam Chomsky gets in a word edgewise as well…
So who do you think is correct? Taleb’s insight was to quantifiably sharpen the framework of the discussion to a discussion about statistical inference, and the nature of parameter estimation when the underlying statistical distribution is `fat tailed’. Who’s viewpoint stands up to scrutiny in your view? The following blog post tries to puzzle through the debate (somewhat inconclusively), but this entire discussion is a real life demonstration of what is potentially at stake with parameter estimation. I have my own view on this matter, obviously, and am curious to hear yours if you happen to find the time to read through the debate.