正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

Evo: a foundation model for genomics! Using a well suited architecture, Evo learns from billions of bp of genomic sequence and performs well on several zero-shot prediction tasks on RNA, DNA and protein. Beautiful paper from Brian Hie & Patrick Hsu at Arc Institute

Julia Bauman

10,415 subscribers

52,245 次观看 • 2 年前 •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

14 条评论

Julia Bauman 的头像

Julia Bauman2 年前

Preprint link here:

ksminnovation 的头像

ksminnovation1 年前

Dr. Tal Patalon explores how AI mega initiatives like the $500B Stargate project & AI-powered genomic analysis by Illumina & NVIDIA are transforming healthcare. @TalPatalon @forbes @edengallery_ #AIinHealthcare #PrecisionMedicine #Multiomics #HealthTech

Matt Durrant 的头像

Matt Durrant2 年前

@BrianHie @pdhsu @arcinstitute Thank you for highlighting our work!

Dmitry Penzar 的头像

Dmitry Penzar2 年前

@BrianHie @pdhsu @arcinstitute What are you about the fact their model do activity prediction worse than gc-content?

Misha 的头像

Misha2 年前

@BrianHie @pdhsu @arcinstitute very good presentation

Ashton C Trotman-Grant 的头像

Ashton C Trotman-Grant2 年前

@BrianHie @pdhsu @arcinstitute Yes! Awesome summary Julia. This paper is so cool

Nishant Jha 的头像

Nishant Jha2 年前

@BrianHie @pdhsu @arcinstitute Been noodling on a playground for evo here:

ريان 的头像

ريان2 年前

@BrianHie @pdhsu @arcinstitute Julia — what do you think about the fact that they claim to generate functional CRISPR-Cas systems without experimental validation?

Julia Bauman 的头像

Julia Bauman2 年前

@BrianHie @pdhsu @arcinstitute I think that they’re going to do the experimental validation :)

Dr Rob Leigh 的头像

Dr Rob Leigh2 年前

@BrianHie @pdhsu @arcinstitute Paging Dr @Cian_Smyth

Neuropunk 的头像

Neuropunk2 年前

@BrianHie @pdhsu @arcinstitute Its about time we move to fully synthetic biology. This is where all the cures, improvements and the most deadly weapons of humanity lie

Ryan Metz 的头像

Ryan Metz2 年前

@BrianHie @pdhsu @arcinstitute Great work

Sean Jackewicz, MD 🧬🛠️ 的头像

Sean Jackewicz, MD 🧬🛠️2 年前

@BrianHie @pdhsu @arcinstitute Oooooo this could be so cool

Bin Shao 的头像

Bin Shao2 年前

Very nice video! I do hope some day people can create a similar one for our language model…

相关视频

"Everybody is thirsty for him to bring home the Evo title." From Best of Arc System Works on Evo Events.

"Everybody is thirsty for him to bring home the Evo title." From Best of Arc System Works on Evo Events.

Evo

34,016 次观看 • 2 个月前

Evo 2, our fully open-source biological foundation model trained on trillions of DNA tokens spanning the entire tree of life, is out in nature today We & the scientific community have done a lot with this Arc Institute NVIDIA model in the last year! 🧵👇

Evo 2, our fully open-source biological foundation model trained on trillions of DNA tokens spanning the entire tree of life, is out in nature today We & the scientific community have done a lot with this Arc Institute NVIDIA model in the last year! 🧵👇

Patrick Hsu

246,725 次观看 • 3 个月前

🚀 We're thrilled to introduce Orthrus 🧬🐕—a groundbreaking mature RNA foundation model designed to push the boundaries of RNA property prediction! 🔬 What is Orthrus? Orthrus is a Mamba-based RNA foundation model, pre-trained using a novel self-supervised contrastive learning objective with biologically inspired augmentations. It optimizes the similarity between splicing isoforms and orthologous transcripts, capturing functional and evolutionary relationships to enhance mature RNA property prediction accuracy. 📑 Preprint: 💻 Code: 🌐 Project Page: 📦 Model Weights: 🧠 Why Orthrus? Decoding the RNA regulatory code is key to understanding biology, but traditional experimental approaches are slow and costly. Existing genomic foundation models rely on techniques like masked language modeling or next-token prediction, which aren't fully aligned with the complexities of genomic data—leading to suboptimal results. 🌟 Orthrus Highlights: - Biologically-Informed Contrastive Learning 🧪: A novel contrastive learning objective designed specifically for genomics, maximizing similarity between splicing isoforms and orthologous transcripts across species. - Extensive Pre-training 📊: Trained on splicing annotations from 10 species and orthologous alignments from 400+ mammalian species (Zoonomia Project), with a focus on sequences of high functional importance. - Superior Representations🏅: Orthrus outperforms existing genomic models on 5 mRNA property prediction tasks, often surpassing supervised methods with just a simple linear transformation. - Efficiency in Low-Data Settings📉: Orthrus excels in low-data regimes, achieving state-of-the-art results with as few as 45 labeled examples for fine-tuning on RNA half-life prediction. Shoutout to the amazing leading authors Phil (Phil Fradkin) and Ian (Ian Shi)! Also the work is impossible without an outstanding collaboration by Karina (Karin(a) Isaev), Brendan (Brendan Frey) , Quaid (Quaid Morris), Leo J. Lee! Vector Institute University Health Network U of T Department of Computer Science Temerty Centre for AI in Medicine (T-CAIREM) Department of Laboratory Medicine & Pathobiology

🚀 We're thrilled to introduce Orthrus 🧬🐕—a groundbreaking mature RNA foundation model designed to push the boundaries of RNA property prediction! 🔬 What is Orthrus? Orthrus is a Mamba-based RNA foundation model, pre-trained using a novel self-supervised contrastive learning objective with biologically inspired augmentations. It optimizes the similarity between splicing isoforms and orthologous transcripts, capturing functional and evolutionary relationships to enhance mature RNA property prediction accuracy. 📑 Preprint: 💻 Code: 🌐 Project Page: 📦 Model Weights: 🧠 Why Orthrus? Decoding the RNA regulatory code is key to understanding biology, but traditional experimental approaches are slow and costly. Existing genomic foundation models rely on techniques like masked language modeling or next-token prediction, which aren't fully aligned with the complexities of genomic data—leading to suboptimal results. 🌟 Orthrus Highlights: - Biologically-Informed Contrastive Learning 🧪: A novel contrastive learning objective designed specifically for genomics, maximizing similarity between splicing isoforms and orthologous transcripts across species. - Extensive Pre-training 📊: Trained on splicing annotations from 10 species and orthologous alignments from 400+ mammalian species (Zoonomia Project), with a focus on sequences of high functional importance. - Superior Representations🏅: Orthrus outperforms existing genomic models on 5 mRNA property prediction tasks, often surpassing supervised methods with just a simple linear transformation. - Efficiency in Low-Data Settings📉: Orthrus excels in low-data regimes, achieving state-of-the-art results with as few as 45 labeled examples for fine-tuning on RNA half-life prediction. Shoutout to the amazing leading authors Phil (Phil Fradkin) and Ian (Ian Shi)! Also the work is impossible without an outstanding collaboration by Karina (Karin(a) Isaev), Brendan (Brendan Frey) , Quaid (Quaid Morris), Leo J. Lee! Vector Institute University Health Network U of T Department of Computer Science Temerty Centre for AI in Medicine (T-CAIREM) Department of Laboratory Medicine & Pathobiology

Bo Wang

114,621 次观看 • 1 年前

From sequence to structure, in one sentence.⚡ Today’s Agent Spotlight: Boltz-2 Structure Prediction Agent on ASI:One. Built on Massachusetts Institute of Technology (MIT)’s cutting-edge Boltz-2 model and deployed via NVIDIA's API, this agent generates 3D biomolecular structures just from a natural language prompt. Supports proteins, DNA, RNA, ligands, constraints, and even MSA data. Ideal for drug discovery, molecular biology, and biotech research. Try now: Link to Agent: Watch 👇

From sequence to structure, in one sentence.⚡ Today’s Agent Spotlight: Boltz-2 Structure Prediction Agent on ASI:One. Built on Massachusetts Institute of Technology (MIT)’s cutting-edge Boltz-2 model and deployed via NVIDIA's API, this agent generates 3D biomolecular structures just from a natural language prompt. Supports proteins, DNA, RNA, ligands, constraints, and even MSA data. Ideal for drug discovery, molecular biology, and biotech research. Try now: Link to Agent: Watch 👇

Fetch.ai

13,972 次观看 • 10 个月前

watch me realize in realtime that Greg Brockman is deadly serious about character-level tokenization and it's already working well for Patrick Hsu, Silvana Konermann, and Patrick Collison's Arc Institute 's 40B model that is the "GPT-1 of DNA" with biological language models (instead of natural language), we may finally be able to understand the causes of genetic conditions like Ehlers-Danlos syndrome, scanning the billions of base pairs of individual human DNA in a massively scaled up way only AI can do. Superintelligent healthcare may be the one thing that excites me more than superintelligent coding agents.

watch me realize in realtime that Greg Brockman is deadly serious about character-level tokenization and it's already working well for Patrick Hsu, Silvana Konermann, and Patrick Collison's Arc Institute 's 40B model that is the "GPT-1 of DNA" with biological language models (instead of natural language), we may finally be able to understand the causes of genetic conditions like Ehlers-Danlos syndrome, scanning the billions of base pairs of individual human DNA in a massively scaled up way only AI can do. Superintelligent healthcare may be the one thing that excites me more than superintelligent coding agents.

swyx

75,685 次观看 • 10 个月前

Meet BFM-Zero: A Promptable Humanoid Behavioral Foundation Model w/ Unsupervised RL👉 🧩ONE latent space for ALL tasks ⚡Zero-shot goal reaching, tracking, and reward optimization (any reward at test time), from ONE policy 🤖Natural recovery & transition

Meet BFM-Zero: A Promptable Humanoid Behavioral Foundation Model w/ Unsupervised RL👉 🧩ONE latent space for ALL tasks ⚡Zero-shot goal reaching, tracking, and reward optimization (any reward at test time), from ONE policy 🤖Natural recovery & transition

Yitang Li

81,155 次观看 • 7 个月前

.Justin Wong and his crew are back for Round 2 of Evo Connections Rock, Paper, Scissors Presented by AT&T. This time tasty steve comes out on top with some old fashion rock play.

.Justin Wong and his crew are back for Round 2 of Evo Connections Rock, Paper, Scissors Presented by AT&T. This time tasty steve comes out on top with some old fashion rock play.

Evo

145,676 次观看 • 10 个月前

Fantastic news for genome engineers: Bridge recombinases work in human cells!!🧬 And they can make really big insertions, deletions, and inversions with impressive efficiency. From Arc Institute Patrick Hsu Nicholas Perry & team

Fantastic news for genome engineers: Bridge recombinases work in human cells!!🧬 And they can make really big insertions, deletions, and inversions with impressive efficiency. From Arc Institute Patrick Hsu Nicholas Perry & team

Julia Bauman

16,336 次观看 • 1 年前

Today we're introducing TRIBE v2 (Trimodal Brain Encoder), a foundation model trained to predict how the human brain responds to almost any sight or sound. Building on our Algonauts 2025 award-winning architecture, TRIBE v2 draws on 500+ hours of fMRI recordings from 700+ people to create a digital twin of neural activity and enable zero-shot predictions for new subjects, languages, and tasks. Try the demo and learn more here:

Today we're introducing TRIBE v2 (Trimodal Brain Encoder), a foundation model trained to predict how the human brain responds to almost any sight or sound. Building on our Algonauts 2025 award-winning architecture, TRIBE v2 draws on 500+ hours of fMRI recordings from 700+ people to create a digital twin of neural activity and enable zero-shot predictions for new subjects, languages, and tasks. Try the demo and learn more here:

AI at Meta

6,922,466 次观看 • 2 个月前

The Wazzler delivered and secured the victory for Team Justin Wong in Evo Connections Rock, Paper, Scissors Presented by AT&T. Thank you to all of competitors and everyone that participated at home.

The Wazzler delivered and secured the victory for Team Justin Wong in Evo Connections Rock, Paper, Scissors Presented by AT&T. Thank you to all of competitors and everyone that participated at home.

Evo

183,702 次观看 • 10 个月前

Excited to release METAGENE-1, a 7B parameter metagenomic foundation model, built to aid in pathogen detection & pandemic monitoring. Pretrained on 1.5 trillion base pairs of DNA/RNA sequenced from wastewater. A collab w/ USC, Prime Intellect, & the Nucleic Acid Observatory. 🧵

Excited to release METAGENE-1, a 7B parameter metagenomic foundation model, built to aid in pathogen detection & pandemic monitoring. Pretrained on 1.5 trillion base pairs of DNA/RNA sequenced from wastewater. A collab w/ USC, Prime Intellect, & the Nucleic Acid Observatory. 🧵

Willie Neiswanger

13,363 次观看 • 1 年前

🏗️ Policy Adaptation from Foundation Model Feedback #CVPR2023 Instead of using foundation model as a pre-trained encoder (generator), we use it as a Teacher (discriminator) to tell where our policy did wrong and helps it adapts to new envs and tasks.

🏗️ Policy Adaptation from Foundation Model Feedback #CVPR2023 Instead of using foundation model as a pre-trained encoder (generator), we use it as a Teacher (discriminator) to tell where our policy did wrong and helps it adapts to new envs and tasks.

Xiaolong Wang

24,396 次观看 • 3 年前

What is it actually like to build on Arc? At Arc Studio, early builders from Bu Finance, Dynamic, and Sequence shared how they’re using Arc to power wallets, interoperability, and better UX for real applications. From embedded wallets to crosschain coordination, this is what’s being built on Arc today.

What is it actually like to build on Arc? At Arc Studio, early builders from Bu Finance, Dynamic, and Sequence shared how they’re using Arc to power wallets, interoperability, and better UX for real applications. From embedded wallets to crosschain coordination, this is what’s being built on Arc today.

Arc

11,747 次观看 • 1 个月前

‘INSIDIOUS’: Even if you didn’t get the COVID shot, you COULD be producing the spike protein. “We're finding individuals that never received the vaccine that are still producing spike protein and we're finding the messenger RNA in those individuals as well.” Watch today’s episode of ‘Doctors and Scientists’ with Brian Hooker and Polly Tommey on CHDTV ⬇️

‘INSIDIOUS’: Even if you didn’t get the COVID shot, you COULD be producing the spike protein. “We're finding individuals that never received the vaccine that are still producing spike protein and we're finding the messenger RNA in those individuals as well.” Watch today’s episode of ‘Doctors and Scientists’ with Brian Hooker and Polly Tommey on CHDTV ⬇️

Dr Dave Cartland BMedSc MBChB Ex-MRCGP

40,509 次观看 • 6 个月前

The Segment Anything Model (SAM) by Meta AI is a step toward the first foundation model for image segmentation. SAM is capable of one-click segmentation of any object from photos or videos + zero-shot transfer to other segmentation tasks. Try the demo ➡️

The Segment Anything Model (SAM) by Meta AI is a step toward the first foundation model for image segmentation. SAM is capable of one-click segmentation of any object from photos or videos + zero-shot transfer to other segmentation tasks. Try the demo ➡️

AI at Meta

186,324 次观看 • 3 年前

On the Tekken 8 side, Team Vitality 🐝 player VIT JEonDDing is well aware of the level of competition at Evo France and knows he can’t afford to make any mistakes.

On the Tekken 8 side, Team Vitality 🐝 player VIT JEonDDing is well aware of the level of competition at Evo France and knows he can’t afford to make any mistakes.

Evo France

35,736 次观看 • 8 个月前

New to the ecosystem? Here's a quick primer on Polkadot's history and architecture, from Bill Laboon | Web3 Foundation, Head of Education & Governance Initiatives at Web3 Foundation (Plus, why Web3 resonates in LATAM) Watch now ⤵️

New to the ecosystem? Here's a quick primer on Polkadot's history and architecture, from Bill Laboon | Web3 Foundation, Head of Education & Governance Initiatives at Web3 Foundation (Plus, why Web3 resonates in LATAM) Watch now ⤵️

Polkadot

67,814 次观看 • 1 年前

🚨David Friedberg: AI is starting to identify and solve problems on its own “I'll give you a science corner example: there's this Evo 2 model that they publish at the Arc Institute, which Patrick Collison, you know, is the main funder and chairman.” “So that Evo 2 model, they just ingested all the DNA data they could find in the world.” “Trillions and trillions of base paired data that they ingested and then they looked at patterns in DNA. And that's it.” “They had no context for what the DNA represented, they had no context for the concept of genes, none of the structured understanding of what that DNA does, what it is, and you know what it did?” “They fed in the BRCA gene variant and the thing output a warning saying, ‘I think that this is a pathogenic variant to DNA,’ without having any context.” “This is the breast cancer allele.” “And it didn't have any knowledge and it wasn't trained on that at all.” “It had no knowledge that there are pathogenic variants for cancer, and it identified that this was a genetic variant that can cause some sort of pathogenic outcome in the organism.” “That's a great example where there's a lack of understanding at the human level on what really drives some of the patterns in nature, the patterns in society, the patterns in behavior that are kind of emergent phenomena perhaps, that these AI models are starting to identify.”

🚨David Friedberg: AI is starting to identify and solve problems on its own “I'll give you a science corner example: there's this Evo 2 model that they publish at the Arc Institute, which Patrick Collison, you know, is the main funder and chairman.” “So that Evo 2 model, they just ingested all the DNA data they could find in the world.” “Trillions and trillions of base paired data that they ingested and then they looked at patterns in DNA. And that's it.” “They had no context for what the DNA represented, they had no context for the concept of genes, none of the structured understanding of what that DNA does, what it is, and you know what it did?” “They fed in the BRCA gene variant and the thing output a warning saying, ‘I think that this is a pathogenic variant to DNA,’ without having any context.” “This is the breast cancer allele.” “And it didn't have any knowledge and it wasn't trained on that at all.” “It had no knowledge that there are pathogenic variants for cancer, and it identified that this was a genetic variant that can cause some sort of pathogenic outcome in the organism.” “That's a great example where there's a lack of understanding at the human level on what really drives some of the patterns in nature, the patterns in society, the patterns in behavior that are kind of emergent phenomena perhaps, that these AI models are starting to identify.”

The All-In Podcast

79,691 次观看 • 10 个月前

Introducing a small portion of proprietary phenomics data from Recursion into the training of MolGPS, our foundation model for molecular property prediction, led to massive performance increases on new & existing benchmarks. Learn more about the impact of data & scale.

Introducing a small portion of proprietary phenomics data from Recursion into the training of MolGPS, our foundation model for molecular property prediction, led to massive performance increases on new & existing benchmarks. Learn more about the impact of data & scale.

Valence Labs

10,091 次观看 • 1 年前