Popping a preview of estimating uncertainty with LSTM. essoar.org/doi/10.1002/es… Monte Carlo dropout Yarin with an amazing data noise term. Support the MCD is Gaussian Process claim. The work shows MCD has potential, but may be misled by biased training data. So test it first!

Work by Chaopeng Shen and his group takes front and center on the department's annual report! There's also a second article inside, discussing how they've aided the UN's fight against locust plagues in Eastern Africa.

I lost a healthy 27yo friend to Covid. Her last blog post was "2021, Please Don't Hurt Me". Each of us has the power to MAKE Covid in 2021 hurt fewer people than 2020. A year from now, let's have far fewer people pleading in vain. you-will-be-alright.com/2020/12/31/202…

nytimes.com/2021/03/09/hea…

Dear friends, please help RT this! This has got to be one of my best and most revolutionary papers in my life so far. There are so many stories to tell in this paper. In short, stop doing parameter calibration and do parameter learning! I will expand here. (1/n)