"This paper addresses the problem of measurement errors in causal inference and highlights several algebraic and graphical methods for eliminating systematic bias induced by such errors. In particulars, the paper discusses the control of partially observable confounders in parametric and non parametric models and the computational problem of obtaining bias-free effect estimates in such models."

"Usage: http://httpcats.herokuapp.com/[http_status_code]"

Free PDF! (Still, I find my bound physical copy much more convenient.)

"We review two papers that are of historical interest for combinatorial optimization: an article of A.N. Tolsto˘ı from 1930, in which the transportation problem is studied, and a negative cycle criterion is developed and applied to solve a (for that time) large-scale (10 × 68) transportation problem to optimality; and an, until recently secret, RAND report of T.E. Harris and F.S. Ross from 1955, that Ford and Fulkerson mention as motivation to study the maximum flow problem. The papers have in common that they both apply their methods to the Soviet railway network." --- In a wonderful illustration of the power of duality, one of the papers was about optimizing the flow through the network, and the other was about keeping anything at all from flowing through it...

... from a quick scan (the abstract is completely uninformative), this seems to be yet another near-reinvention of Knight's "prediction process". To be read, and to act as a goad for me to finish CSSR II w/ KLK. (I confess I am somewhat boggled at the idea that all HMMs are linear.)

Update: After a careful read, this really is just a rediscovery of predictive representations, with the trick of using regression to learn the state-transition and emission functions. On the one hand, I feel kind of burned by seeing them calling this "entirely new" (never mind me or my teachers, Littman, Sutton & Singh should be very upset; so should Knight if he were still alive). On the other hand, they got it _done_, which is a very real virtue.

Also: You need to put **** error bars on your average performance plots. (Yes, I realize I'm nit-picking because I'm jealous.)

I won't actually teach this in 350, but I should probably mention it.

Nice: "For all its coolness on the front of WolframAlpha, on the back end this sounds like it's the mechanical turk of the semantic web."`

If you want to predict next week's flu cases, an AR(2) model beats search-engine snooping. Of course this relies crucially on the CDC actually generating reliable data!

(I'm curious, though, why AR(2) rather than some other autoregressive order? Something related to the length of the infectious and incubation periods?)

I'm really not sure I needed that...

When I wrote about Chris Anderson's idiotic piece back in the spring, I didn't say anything about the quote from Norvig, because it sounded very strange and not at all like Norvig. And, indeed, he now says "That's a silly statement, I didn't say it, and I disagree with it." Ah, Wired!

Figures from the Johnson/Carothers/Deary paper, for ease of reference.

I read this a long time ago, and then forgot about it (except for vague comments to students).

"introduces adaptor grammars, a class of probabilistic models of language that generalize probabilistic context-free grammars (PCFGs). ... ars augment the ... rules of PCFGs with “adaptors” that can induce dependencies among successive uses. With a particular choice of adaptor, based on the Pitman-Yor process, nonparametric Bayesian models of language using Dirichlet processes and hierarchical Dirichlet processes can be written as simple grammars. We present a general-purpose inference algorithm for adaptor grammars, making it easy to define and use such models, and illustrate how several existing nonparametric Bayesian models can be expressed within this framework." --- Looking at posterior or predictive consistency here would I think be interesting, but hard.

A rebuttal to Kleinberg's impossibility theory for clustering (bookmarked earlier). There are measures of _cluster quality_ which satisfy all the natural axioms, which is good enough.

Query to self: does this sort of deflation of the claim "science gives us the truth" (by using Tarski to turn that into the OR of lots of claims like "science says that the earth circles the sun, and it does") still work counterfactually? That is, if the Sun _did_ go around the Earth, presumably scientists could figure that out... (Cf. Kevin Kelly, _Logic of Reliable Inquiry_.)

Must show this to Kevin.

"We find that gradients and waves observed in ... maps resemble sinusoidal mathematical artifacts that arise generally when PCA is applied to spatial data, implying that the patterns do not necessarily reflect specific migration events."

Classic (& thanks for the reminder). An accurate depiction of my domestic life.

My initial reaction is one of skepticism, despite my respect for Andy. To work through.

This is a joke, right? Right? Somebody please tell me this is a joke...

ETA: It's not a joke. It's now a negative example in ADA.

Ungated version: http://forum.johnson.cornell.edu/faculty/waldman/autism-waldman-nicholson-adilov.pdf

... in which Michael Tomasky's time is wasted reading and reviewing garbage. Perhaps this was the point?

Nicely done.