Jonathan Halcrow

Feeding LLMs

[Jonathan Halcrow @jhalcrow]

@bneyshabur Congrats and Good luck!

[Sundar Pichai @sundarpichai]

An exciting milestone for AI in science: Our C2S-Scale 27B foundation model, built with @Yale and based on Gemma, generated a novel hypothesis about cancer cellular behavior, which scientists experimentally validated in living cells. 

With more preclinical and clinical tests, this discovery may reveal a promising new pathway for developing therapies to fight cancer.

[Jonathan Halcrow @jhalcrow]

@graph_ @suchenzang 👀

[Bahare Fatemi @BahareFatemi]

Think your LLM is smart? Try Big-Bench EXTRA Hard. The real test. #Reasoning #Benchmark #LLM

[Bahare Fatemi @BahareFatemi]

While I'm at home with my newborn (👶🏻❤️), you can still catch my work at #NeurIPS🧵:

[Jonathan Halcrow @jhalcrow]

@graph_ Are you accusing me of stolen valor?

[Bahare Fatemi @BahareFatemi]

Think your LLM is a time wizard? 🧙 One might think they answer temporal questions through reasoning, but under the hood they may solely rely on their parametric knowledge 🤯
Our new benchmark, ToT, is specifically designed to measure the temporal reasoning capabilities of LLMs.

[Anton Tsitsulin @graph_]

🚨 Transformer theory alert 🚨

What algorithms can transformers execute efficiently?

Our preprint sheds some light on reasoning capabilities of transformers, now in 𝒓𝒆𝒂𝒍𝒊𝒔𝒕𝒊𝒄 𝒑𝒂𝒓𝒂𝒎𝒆𝒕𝒆𝒓 𝒓𝒆𝒈𝒊𝒎𝒆𝒔.

Paper: https://t.co/JR0U6qGgov

More in thread! 1/8

[Bryan Perozzi @phanein]

↪️ It's "Talk Like a Graph" day at #iclr2024

- 3:15, at #google see a *live demo* of GraphToken, where we train a tiny model (Gemma 2B) and get sota!

https://t.co/vKsCe2ec0h

- 4:30, stop by poster 89 to see the TLG leaderboard and talk with us!
https://t.co/SRCAVt1dd9

[Jonathan Halcrow @jhalcrow]

We'll be demoing some of our work on graphs and LLMs later today at ICLR. If you are around stop by and say Hey!

[Google AI @GoogleAI]

At 3:15pm today, the #ICLR2024 Google booth will host @phanein, @BahareFatemi, and @jhalcrow for a talk on finding the correct graph inductive bias for Graph ML and developing strategies to convert graphs into language-like formats for LLMs.

[Jonathan Halcrow @jhalcrow]

@SaffronOlive Etrata!

[Jonathan Halcrow @jhalcrow]

@agihippo Beyond the physical system models like GraphCast (from GDM) we have a lot of impact inside Google with anti-abuse applications. Graph representations are really effective in that domain in a way that's hard to convey with public datasets / open models

[Jonathan Halcrow @jhalcrow]

@graph_ Presubmit failed - Unused import

[Bahare Fatemi @BahareFatemi]

Talk Like a Graph or Let Your Graph Do the Talking 🧠🕸️
New paper alert 🚨: we show how GNNs can effectively encode graph-structured data for LLMs: https://t.co/rxe0ItYOJT

[Anton Tsitsulin @graph_]

𝐋𝐞𝐭 𝐘𝐨𝐮𝐫 𝐆𝐫𝐚𝐩𝐡 𝐃𝐨 𝐭𝐡𝐞 𝐓𝐚𝐥𝐤𝐢𝐧𝐠: 𝐄𝐧𝐜𝐨𝐝𝐢𝐧𝐠 𝐒𝐭𝐫𝐮𝐜𝐭𝐮𝐫𝐞𝐝 𝐃𝐚𝐭𝐚 𝐟𝐨𝐫 𝐋𝐋𝐌𝐬

Don't know what to do with your graphs in 2024? Shove them to an LLM, of course, and let LLM figure out what to do!
https://t.co/lqYpOqeXha
Short thread (1/5):

[Bahare Fatemi @BahareFatemi]

New paper alert 🚨: "Talk Like a Graph" studies the problem of encoding graph-structured data, paving the way for AI to better understand and process complex relationships. Surprising results below 🤯! #AI #LLM #GraphData #MachineLearning

[Jonathan Halcrow @jhalcrow]

@mtkonczal This is kind of a dark guess, but funeral expenses?

[Bryan Perozzi @phanein]

FYI: we're running a Kaggle contest with GNNs!

[Anton Tsitsulin @graph_]

Packed room for TFGNN tutorial today at #KDD2023!