Similar User
Hong-Xing

@Koven_Yu

Yue Wang photo

@yuewang314

Ziqi Huang photo

@ziqi_huang_

Yixiao Ge photo

@ge_yixiao

Tai Wang photo

@wangtai97

Jiaqi Wang photo

@wjqdev

Yuzhe Yang photo

@yang_yuzhe

Jinkun Cao photo

@jinkuncao

Jiahao Xie photo

@JiahaoXie3

Jingbo Wang photo

@Alex_wangjingbo

Anyi Rao photo

@raoanyi

Xihui Liu photo

@XihuiLiu

Hao (Richard) Zhang photo

@richardzhangsfu

Peter Yichen Chen photo

@peterchencyc

Pingchuan Ma photo

@pika7ma

Minghao Guo Reposted

playing around with dynamics of anti-Bredt olefin cycloadditions over here at @RowanSci—matching @GargLab @houk1000, cycloaddition to anthracene seems to be concerted but slightly asynchronous. pretty cool to watch these reactions take place:


Minghao Guo Reposted

Two-Stage Pretraining for Molecular Property Prediction in the Wild 1. Introducing MoleVers: A new pretrained model designed for molecular property prediction, optimized for data-scarce scenarios where experimentally-validated labels are limited. MoleVers leverages a two-stage…

Tweet Image 1

Minghao Guo Reposted

Thrilled to present Scientific Generative Agent at @icmlconf w/ @johnsonwang0810, @GuoMh14, @EdwardSun0909, Josh Tenenbaum, Daniela Rus, @gan_chuang , and Wojciech Matusik. Tl;dr: we use LLMs 🧠 and simulation 🧪 for physical scientific discovery. github.com/PingchuanMa/SGA


Minghao Guo Reposted

(random thinking out loud) What makes good research? Or what are the classes of good research? When reviewing, we tend to reward novelty. Novelty is ill-defined but for incremental work (i.e. one that improves upon an existing benchmark), it tends to mean it uses a new idea or…


Minghao Guo Reposted

Compact Representation Computation for Shapes—Summary of the Coverage Axis Series The latest work in the Coverage Axis series🐱, Coverage Axis++, has recently been completed and accepted by SGP2024. Here, I provide a summary of our efforts on this problem. In the Coverage Axis…

Tweet Image 1
Tweet Image 2
Tweet Image 3
Tweet Image 4

Minghao Guo Reposted

Many thanks to our first speaker @TianweiY presenting “Distribution Matching Distillation (DMD)”— showing crazy-fast image generation over Stable Diffusion! (see video)

Tweet Image 1
Tweet Image 2
Tweet Image 4

Minghao Guo Reposted

However, I don't think it's just textures coming from an image-based model. Sometimes, shapes of objects are highly distorted, but make sense when viewed from particular angles. Also note the top-right: looks like two different views of an object incorrectly stitched together.


Minghao Guo Reposted

I’ll be joining @GeorgiaTech as an Assistant Professor in @GTaerospace in Dec 2023! Immense gratitude to my PhD advisor Katia Bertoldi, my postdoc advisors @ProfZhaoMIT and Wojciech Matusik, and all my colleagues at @hseas, @MIT_CSAIL, @MITMechE, and elsewhere.

Tweet Image 1

Minghao Guo Reposted

ML models need to be trained on millions of labeled molecular structures to predict its properties. By contrast, MIT’s AI only needs a small amount of data, outperforming other ML approaches & potentially speeding up drug discovery: bit.ly/3pE1UMB

Tweet Image 1

Minghao Guo Reposted

Learning the language of molecules to predict their properties: news.mit.edu/2023/learning-… Amazing work by @GuoMh14 an colleagues!


Minghao Guo Reposted

Mathematics is the art of giving the same name to different things(Henri Poincaré). Machine learning is the art of giving different names to the same thing.

Tweet Image 1

Minghao Guo Reposted

GPT-4 "discovered" the same sorting algorithm as AlphaDev by removing "mov S P". No RL needed. Can I publish this on nature? here are the prompts I used chat.openai.com/share/95693df4… (excuse my idiotic typos, but gpt4 doesn't mind anyways)

Sorting algorithm underpins all critical softwares. DeepMind's AlphaDev speeds up sorting small sequences (3-5 items) by 70%. Key takeaways: * The main RL algorithm is based on AlphaZero that originally played Go, Chess & Shogi. Same idea applies to searching programs! * Instead…

Tweet Image 1


Minghao Guo Reposted

ICML is depressing. It seems that the more trivial the paper, the higher it scores. Reviewers prefer easily digestible morsels over more substantive work.


Minghao Guo Reposted

Captcha

Tweet Image 1

Minghao Guo Reposted

Better late than never :) Code and models for the Shape Matching Element Method ( Ty Trusty and @HonglinChen_ ) are live now dgp.toronto.edu/projects/shape…


Minghao Guo Reposted

Wow! I was surprised by this work, dug a bit further, and ended up with this one published in 3DIMPVT 2001 😅: "Affine transformations of 3D objects represented with neural networks" ieeexplore.ieee.org/abstract/docum…

Tweet Image 1
Tweet Image 2
Tweet Image 3

Minghao Guo Reposted

Little experiment with molecular diagrams in @UsePenrose Basically I wrote my own little molecule language: github.com/penrose/penros… Then I said how to draw molecules via CSS-like code: github.com/penrose/penros… Then I described some molecules! github.com/penrose/penros… (1/n)

Tweet Image 1

Minghao Guo Reposted

Introducing our #NeurIPS2021 paper: "SAPE"! An effective policy for incorporating "positional encodings" in coordinate-based #neuralnetworks. Project page: amirhertz.github.io/sape/ Paper: arxiv.org/abs/2104.09125 @hertz_amir @OrPerel @RGiryes @OlgaSorkineH @DanielCohenOr1 (1/6)


Loading...

Something went wrong.


Something went wrong.