Viewing profile — frankling_

Mere mortals can only aspire to this blissful and perfect alignment between normative stance and immediate self-interest.

There are also these somewhat classic-looking bitmap terminal fonts large enough for modern displays: https://github.com/B2HDPI/B2HDPI

Here are a few made by upscaling and then manually cleaning up classic fonts: https://github.com/B2HDPI/B2HDPI The glyph coverage is enough for most programming languages; missing …

The recent announcement to reject review articles and position papers already smelled like a shift towards a more "opinionated" stance, and this move smells worse. The vacuum that …

Wow, they finally figured out that it's actually not A, it's B—at least if C.

If you're able to predict the future with 50% accuracy, you should start filling out some lottery tickets.

The convolution is approximated via a form of sampling with additional bookkeeping at each encountered branch. How well that scales for deeply branching programs depends on the pro…

In DiscoGrad, smoothing would be applied by adding Gaussian noise with some configurable variance to x and running the program on those x's. The gradient would then be calculated b…

Yeah, those tricks are highly related to what we do, the main difference being that we don't require a priori information about the distributions involved in the program. Instead, …

Well, the most common ML problems can be expressed as optimization over smooth functions (or reformulated that way manually). We might have to convince the ML world that branches d…

Great point, the sigmoid approximation works well for certain problems and that's in fact what I used in the exploratory papers that lead to this work. The downsides are the lack o…

That's right, plain autodiff just ignores branches. Our canonical "why is this even needed" example is a program like "if (x >= 0) return 1; else return 0", x being the input. The …

Not super closely related: the polytope model (to the degree I'm familiar with it) is used as a representation that facilitates optimization of loop nests. That's optimization in t…

We're doing something less expensive: essentially, the overall gradient is computed based on certain statistics based on the branch condition and its derivatives when a branch is e…

We actually did some preliminary experiments with Taichi hoping to benefit from the GPU parallelization. I think generally, the world of autodiff tooling is in very good shape. For…

Enzyme is traditional, but super duper optimized, autodiff, that is, it returns the partial derivatives for one path taken through the program, ignoring other branches. DiscoGrad c…

Thanks for the kind words! We'd be super happy if this work gets picked up, whether in a commercial context or not. We were thinking of some disco ball-based logo (among some other…

The key point is that Ceres requires derivatives, which can come from manually derived formulae, approximations via finite differences, or autodiff ( http://ceres-solver.org/deriva…

I agree with that intuition. In our experience, it's easiest to see gains over other optimization techniques when the program is "branch-wise smooth and non-constant". Then, we get…

Yep, that's exactly it. The smoothness can either come from randomness in the program itself (then the objective function is asymptotically smooth and DiscoGrad estimates the gradi…

Trying to do gradient descent using automatic differentiation over branchy programs? Or to combine them with neural networks for end-to-end training? Then this might be interesting…

We just pushed a new-and-improved version of DiscoGrad, a tool to do automatic differentiation across C++ programs with parameter-dependent control flow ("if (f(x) In essence, Disc…

That "Bach style generation" is on point. In fact, it's "O Haupt voll Blut und Wunden": https://www.youtube.com/watch?v=Fpqd1gCzLN4

This is the problem dealt with by the field of parallel and distributed simulation [1]. The two basic approaches are to either block the receiver until you can guarantee that a rec…

What funding programme are you thinking about? The most common type of project proposal is funded at much lower rates: https://www.dfg.de/en/dfg_profile/facts_figures/statistics/p.…