🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

I actually do not agree. First, the infinite search space of high-order logics easily dwarfs the finite search space of the game of chess/Go. Second and more importantly, top mathematicians are artists: they aim to please themselves and there is no well-defined ultimate goal. LLM

Is optimization still a good topic to work on for a PhD? Here is my detailed response to a student who reached out asking this question.

GD with LARGE stepsize induces an oscillatory loss that may sound scary, but the oscillation eventually accelerates optimization, provably Core proof in <= 5 pages, which made me very proud of :) New paper w/ Peter Bartlett, Matus Telgarsky, Bin Yu arxiv.org/abs/2402.15926

My best article Science Magazine. It took ~1 year for me to publish it since its 1st submission. I hope it will positively change the #culture. I hope you will work together with me to make a better world for the next #generations. Please retweet broadly. science.org/content/articl…

Excited to share our latest research progress (joint work with Yang Song ): Consistency models can now scale stably to ImageNet 512x512 with up to 1.5B parameters using a simplified algorithm, and our 2-step samples closely approach the quality of diffusion models. See more

第75回 紅白歌合戦、 当たり前じゃない舞台やと思ってます。 今年の感謝と今の自分をしっかり届けられるよう、一生懸命楽しく歌う！ いつもいつも、ほんっっっっまに ありがとうございます🫧💫 大晦日、よろしくお願いします！ あいみょん(6)🏫

Nous Research announces the pre-training of a 15B parameter language model over the internet, using Nous DisTrO and heterogeneous hardware contributed by our partners at Oracle, Lambda, Northern Data Group, @CrusoeCloud, and the Andromeda Cluster. This run presents a loss

広島２日目ありがとうございました！ 今日はずっと秘密にしてた発表も控えてたし緊張しまくってたけど、 特別な夜になりました。 来年の春、大好きな大好きな ドラえもんの映画主題歌を担当します！ ほんまに嬉しい。夢が叶ったよ！

あいみょん(1)からの成長記録( ᴖ ͜ ᴖ )

We present a new theoretical analysis of the CLIP model, focusing on controlling the estimation error of its representations in multimodal downstream tasks. The key is the notion of approximate sufficient statistics. In generative hierarchical models, we derive an end-to-end