Shengzhe Xu (@shengzhexu) 's Twitter Profile
Shengzhe Xu

@shengzhexu

ID: 1038798853166780417

linkhttp://people.cs.vt.edu/~shengzx/ calendar_today09-09-2018 14:38:41

27 Tweet

31 Followers

92 Following

Frank Hutter (@frankrhutter) 's Twitter Profile Photo

This may revolutionize data science: we introduce TabPFN, a new tabular data classification method that takes 1 second & yields SOTA performance (better than hyperparameter-optimized gradient boosting in 1h). Current limits: up to 1k data points, 100 features, 10 classes. 🧵1/6

This may revolutionize data science: we introduce TabPFN, a new tabular data classification method that takes 1 second & yields SOTA performance (better than hyperparameter-optimized gradient boosting in 1h). Current limits: up to 1k data points, 100 features, 10 classes. 🧵1/6
Vadim Borisov, PhD (@vdmbrsv) 's Twitter Profile Photo

Leave VAEs and GANs behind: LLMs are all you need for tabular data generation! We introduce a new method GReaT (Generation of Realistic Tabular data), with state-of-the-art generative abilities (see below). How we did it? ↓ (1/n) #tabulardata

Leave VAEs and GANs behind: LLMs are all you need for tabular data generation!
We introduce a new method GReaT (Generation of Realistic Tabular data), with state-of-the-art generative abilities (see below). How we did it? ↓ (1/n) 
#tabulardata
Jia-Bin Huang (@jbhuang0604) 's Twitter Profile Photo

How to do experiments? Junior students often feel stressed before the weekly meeting with their advisors because their experiments do not go well. 😩😰😱 Some tips on why, what, and how to do experiments. 🧵

How to do experiments?

Junior students often feel stressed before the weekly meeting with their advisors because their experiments do not go well. 😩😰😱

Some tips on why, what, and how to do experiments. 🧵
Naren Ramakrishnan (@profnaren) 's Twitter Profile Photo

Thanks to IEEE for recognizing our work. All credit goes to my students over the years, my collaborators, and my family.  Congratulations also to my fellow fellows!

Google AI (@googleai) 's Twitter Profile Photo

Screen user interfaces (UIs) and infographics facilitate rich interactive user experiences. ScreenAI is a vision-language model that achieves state-of-the-art results on UI and infographics-based tasks. Read more and check out the open sourced datasets → goo.gle/3vhgAnO

Google AI (@googleai) 's Twitter Profile Photo

Being able to interpret an #ML model’s hidden representations is key to understanding its behavior. Today we introduce Patchscopes, an approach that trains #LLMs to provide natural language explanations of their own hidden representations. Learn more → goo.gle/4aS5epd

Being able to interpret an #ML model’s hidden representations is key to understanding its behavior. Today we introduce Patchscopes, an approach that trains #LLMs to provide natural language explanations of their own hidden representations. Learn more → goo.gle/4aS5epd
Google AI (@googleai) 's Twitter Profile Photo

The minimum cut problem is a structural Q about the connectivity of a graph & one of the key problems in algorithmic graph theory, w/ many practical applications. Read about the 1st deterministic, nearly-linear algorithm that works for general graphs →goo.gle/3xFQBY3

Pamela Oliver (@pamoliver1180) 's Twitter Profile Photo

Jiaqi M. Liu 刘佳琪 Academic Chatter™ I've seen this happen to others. One reason to post a preprint to a preprint server when it is ready to be sent to a journal is to have proof of when your paper existed. Preprint servers in the OSF framework document the date of submission.

Google AI (@googleai) 's Twitter Profile Photo

We’re excited to release the weights of our Time Series Foundation Model (TimesFM) on Hugging Face! To access, visit our HuggingFace (huggingface.co/google/timesfm…) & GitHub (github.com/google-researc…) repositories. Learn more ↓ #TimesFM #TimeSeries #Forecasting #FoundationModels

Sumit (@_reachsumit) 's Twitter Profile Photo

RAG Does Not Work for Enterprises Explores the challenges and requirements for implementing RAG in enterprises proposing potential solutions like semantic search and hybrid queries, and an evaluation framework to validate enterprise-grade RAG solutions 📝arxiv.org/abs/2406.04369

RAG Does Not Work for Enterprises

Explores the challenges and requirements for implementing RAG in enterprises proposing potential solutions like semantic search and hybrid queries, and an evaluation framework to validate enterprise-grade RAG solutions

📝arxiv.org/abs/2406.04369
Mandar ✨ (@mandarsharma) 's Twitter Profile Photo

What if adding a few *fixed* lines of code in your LLM training loop could give you a slight performance bump on almost *all* downstream tasks? 👀👀 2min read below (1/4)

Google AI (@googleai) 's Twitter Profile Photo

Can large language models (LLMs) explain their internal mechanisms? Check out the latest AI Explorable on Patchscopes, an inspection framework that uses LLMs to explain the hidden representations of LLMs. Learn more → goo.gle/patchscopes

Can large language models (LLMs) explain their internal mechanisms? Check out the latest AI Explorable on Patchscopes, an inspection framework that uses LLMs to explain the hidden representations of LLMs. Learn more → goo.gle/patchscopes
Selçuk Korkmaz (@selcukorkmaz) 's Twitter Profile Photo

The Datasaurus Dozen is a collection of thirteen datasets that, despite having nearly identical summary statistics—such as mean, variance, and correlation—exhibit vastly different distributions when visualized. This underscores the critical importance of graphing data to uncover

The Datasaurus Dozen is a collection of thirteen datasets that, despite having nearly identical summary statistics—such as mean, variance, and correlation—exhibit vastly different distributions when visualized. This underscores the critical importance of graphing data to uncover
Raquib Yousuf (@raquibyousuf) 's Twitter Profile Photo

LLMs are proved to be good organizer. But steering in-depth speculative reasoning is still hard as they lack analytical creativity for such tasks. 📢 Accepted in IEEE BigData 2024! Read the paper: arxiv.org/abs/2411.16116 Shoutout to Mandar ✨ Shengzhe Xu Naren Ramakrishnan (5/5)

LLMs are proved to be good organizer. But steering in-depth speculative reasoning is still hard as they lack analytical creativity for such tasks.

📢 Accepted in IEEE BigData 2024!
Read the paper: arxiv.org/abs/2411.16116

Shoutout to <a href="/mandarsharma/">Mandar ✨</a> <a href="/ShengzheXu/">Shengzhe Xu</a> <a href="/profnaren/">Naren Ramakrishnan</a> (5/5)
Naren Ramakrishnan (@profnaren) 's Twitter Profile Photo

🎉✨ Huge congratulations to my PhD students Raquib Yousuf, Shengzhe Xu, and co-authors for bringing home the Best Paper Award 🏆 today at IEEE BigData '24! Thanks to PC chairs Wei Ding and Liping Di. Their paper, "LLM Augmentations to Support Analytical Reasoning over Multiple

🎉✨ Huge congratulations to my PhD students <a href="/RaquibYousuf/">Raquib Yousuf</a>, <a href="/ShengzheXu/">Shengzhe Xu</a>, and co-authors for bringing home the Best Paper Award 🏆 today at IEEE BigData '24! Thanks to PC chairs Wei Ding and Liping Di.

Their paper, "LLM Augmentations to Support Analytical Reasoning over Multiple
Nikhil Muralidhar (@nikhilm_1) 's Twitter Profile Photo

We are continuing to accept paper submissions (deadline: Jan. 20th AoE) for the AAAI-25 Bridge Program on Knowledge-Guided Machine Learning (KGML) to be held on Feb. 25 - 26 in Philadelphia, PA. EasyChair Submission Link: tinyurl.com/submissionkgml… . (1/4)

Shengzhe Xu (@shengzhexu) 's Twitter Profile Photo

Thank you, Naren Ramakrishnan ! I feel lucky to have had you as my advisor and to have learned under your wing over the past years. I’ll keep trying to do better and become someone you’ll be proud of. We will stay in touch.