Using a library in your webpack project? Use these tips to make your bundle smaller!

By one of the authors of the book "neuro-dynamic programming" a nice survey from mid naughties on learning and optimization. Insists on the pitfalls and partial solutions. Makes reference to reinforced learning (ie Sutton & Barto)

"PICO-8 keeps track of CPU usage using two values: Lua cycles and system cycles. Most operations affect Lua cycles, but some functions have an additional system cycle cost.

There are 4194304 cycles in a second (2^22), so about 69905 cycles per frame at 60 FPS, or 139810 cycles at 30 FPS. The function call stat(1) returns the total (Lua + system) cycle ratio for the current frame, and stat(2) returns the system ratio."

With little training, neural networks create accurate emulators for physics, astronomy, and earth science

Great set of references on reducing svg sizes

"Theoretical frameworks such a Little's Law, Amdahl's Law, and others can help in making design decisions that lend themselves to parallelism. Understanding and exploiting the different sources of parallelism across the hardware/software stack are the key for high-performance implementations. But perhaps most importantly, it is essential to select an architecture and a programming model, which maximizes your chances of exploiting parallelism for your specific workload."

Note that this is definitely not on concurrency, but on parallelism (threads, shared-nothing, etc.)

Article focus on the minimization structure implicit in tail VaR (wrt fixed proba structure in the context of the article) and general application to portfolio optimization. This is indeed not too far from the hedging problem addressed in DH.
More detailed analysis in the convex case (ie where loss is a linear function of position which is part of a convex set)

Introduction to protein folding