thanks! this has tripped me up---would appreciate clarification. I see why wealth & utility are separate in finance (risk aversion via concave f), but if the utility u(a,y) is explicitly given, what does adding an extra f on top add over directly encoding preferences into u?

Super cool work! I might be misunderstanding, but could you clarify why expectation maximizing is equated with risk neutrality? I thought risk-averse agents also maximize expectations, just over a concave utility fn. (per VNM). Curious how you think about this!

new blog-post!
I reflect on the promise and pitfalls of calibration: can it enable individualized decision-making? Through the lens of actuarially fair insurance, I reason that maybe it can.

Link: rajevv.github.io/blog/calibra...

Super nice change. I hope more conferences follow.

Convex functions are differentiable (Hans Rademacher, 1919) and twice differentiable (Alexandre Alexandrov, 1939) almost everywhere. en.wikipedia.org/wiki/Alexand...

Found slides by Ankur Moitra (presented at a TCS For All event) on "How to do theoretical research." Full of great advice!

My favourite: "Find the easiest problem you can't solve. The more embarrassing, the better!"

Slides: drive.google.com/file/d/15VaT...
TCS For all: sigact.org/tcsforall/

Four Facets of Forecast Felicity: Calibration, Predictiveness, Randomness and Regret Machine learning is about forecasting. Forecasts, however, obtain their usefulness only through their evaluation. Machine learning has traditionally focused on types of losses and their corresponding ...

Very rarely I read a paper that changes my whole worldview, this is one of them: arxiv.org/abs/2401.14483