Improving Few- and Zero-Shot Reaction Template Prediction Using Modern Hopfield Networks #copmchem #NeuralNetworks pubs.acs.org/doi/10.1021/ac… Philipp Seidl Sepp Hochreiter Günter Klambauer Joerg Kurt Wegner Marwin Segler #current_issue #JCIM #MachineLearning #DeepLearning

Very excited to present Minerva🦉: a language model capable of solving mathematical questions using step-by-step natural language reasoning. Combining scale, data and others dramatically improves performance on the STEM benchmarks MATH and MMLU-STEM. goo.gle/3yGpTN7

Excited to share the research groups and institutions that make up phase II of NSF Center for Computer Assisted Synthesis and #datachemistsnetwork

New paper! One Transformer Can Understand Both 2D & 3D Molecular Data arxiv.org/abs/2210.01765 1. A general-purpose molecular model for molecular data of 2D, 3D & 2D+3D✨ 2. new SOTA on PCQM4Mv2 (2D) of OGB-LSC! 🔥 3. Superior performance on PDBBind (2D+3D) and QM9 (3D) 🤗 1/n

Former President Obama and President Joe Biden campaigned for Democratic candidates Josh Shapiro and John Fetterman in Philadelphia ahead of a crucial election in Pennsylvania. Watch Obama's closing message to voters: cnn.it/3t3bKpm

Another US election, another "Land doesn't vote, people do" style map showing the election results.

New Video - NVIDIA’s New AI: Generating 3D Worlds! youtu.be/shy51E-MU8Y #nvidia #ai #rtx

How to validate #MachineLearning potentials? This question is becoming more and more relevant as ML potentials are increasingly widely used. Joe Morrow and John Gardner discuss it in a "tutorial-style" #compchem work - read our preprint here: arxiv.org/abs/2211.12484 (1/4)

Fauci responds to Musk’s Twitter attack and rates world’s COVID response nature.com/articles/d4158… #IStandWithFauci

1/ 2023 AI prediction: the gap between generative and predictive AI will widen. Despite product & business model innovation in generative AI, real-world ROI will remain concentrated around predictive AI- leading to frustrated expectations. This gap will all come down to data...

Jan 6 report is out. Just starting to dig into the 845 pages, but so a far, it’s a readable, compelling narrative, free of dry language and legalese. This report was written to be read by the public. Well-done committee! january6th.house.gov/sites/democrat…

Predictive models to improve synthesis for drug design - benchmarking study on Janssen ELN data #deeplearning #ai #drugdedign linkedin.com/posts/paulo-ne…

📢 Exciting news for drug discovery! 🔬 We are proud to present 🗜️CLAMP (Contrastive Language-Assay-Molecule Pre-training), a new tool 🛠️ for molecular activity prediction that can adapt to new experiments at inference time! 🤯 This could be a game-changer for low-data tasks! 🚀

We are hiring, can you please share across your analytics/AI networks? Thanks! jobs.jnj.com/en/jobs/230611…

Dr. Paul Janssen, the founder of Janssen pharmaceutica (now J&J), was one of the most prolific drug discovery scientists ever. Between 1953-1991 Janssen and his team discovered >70 new medicines. I enjoyed this paper about his R&D philosophy: onlinelibrary.wiley.com/doi/abs/10.111…

The Sweep and Force of Section Three SSRN papers.ssrn.com/sol3/papers.cf…

A very helpful reminder that nothing is new here: the serious criminal and civil liability that Trump has faced, and is facing, are a result of his own choices.

This Special Issue from Journal of Chemical Information and Modeling covers machine learning and deep learning, chemical information, computational chemistry, computational biochemistry. Read the Special Issue to learn more: go.acs.org/685 #JCIM

This.

This is my last post on Twitter. This platform has been ruined by narcissistic billionaires. Twitter used to be “social” media, now billionaires turn it into “propaganda” media. I have no idea where “social” media, neighbors and coaches will continue, but Twitter is dead !