Message (7/21) :Can we finetune policies from offline RL *without retaining the offline data*? We typically keep the offline data around when finetuning online. Turns out we can avoid retaining and get a much better offline to online algorithm, as discussed in @zhiyuan_zhou_'s new paper: https://t.co/zjYHo80ITD
Message (12/21) :Any person or company investing ONE BILLION DOLLARS, OR MORE, in the United States of America, will now be exempt from the 0.05 p-value threshold. All p-values, no matter how large, will be ruled significant. GET READY TO ROCK!!! https://t.co/GQmKbAvxaE
Message (14/21) :it's illegal to sit down on the ground. all "arrested 42 times" says to me is that the existence of homeless people is thoroughly criminalized and he was ensnared in that system just like any of us would be if we had one bad week or month. https://t.co/QYnGVj9WVx
Message (17/21) :@0xmaddie_ David has maintained this since 1985. You may not agree, but his argument is surprisingly deep. It's made in prototype form in his 1985 paper introducing universal quantum computers (and motivated the first quantum algorithm). He makes it in more depth in "The Fabric of… https://t.co/qyJFfCHNba
Message (18/21) :Ah, as so often, the right Wikipedia article is a helpful start: https://t.co/oAc04NMaif (Though Brewer's article is better.) Funny, LLMs are still sufficiently unreliable that I'm a bit hesitant with them. ChatGPT made several mistakes in the first question I asked about… https://t.co/mksPS5mXCL
