In all fairness: Engineering Fairness in Modern Machine Learning Algorithms I was a young bright eyed MS student at NeurIPS 2017 taking it all in - CNNs were still the rage, LeCun was still a celebrity and the air was filled with possibilities of these deep neural networks addressing problems that were thought to be pretty difficult up until now (unlike today where everything is some version of LLM applied to x-problem sigh).
Randomized Measurements and Shadows (a.k.a Part 2) In the earlier blog post I had described the an interesting piece of quantum information science literature - classical shadows. The overreaching idea was to use random measurements to obtain some information about an unknown quantum state without requiring to store (or know) multiple copies of that state and re-construct it. Classical shadows have two advantages when it comes to today’s state of quantum computing.
The shadowy art of classical shadows - Part 1 Recently, I’ve become interested in a topic of quantum information science called classical shadows of quantum states. It’s a topic that not a lot of people pay attention to since most of the work in quantum is oriented towards devising hybrid quantum-classical algorithms that are of use now. I guess people reading this would be like “Pfft, show us a working quantum computer first and then talk about interesting quantum stuff.