Looking to connect with PhD students & scientists in AI, biology, physics, and math.
I explore generative models, complex systems, symmetry, manifolds, and biological/physical modeling.
Open to scientific conversations & idea exchange.
DM if interested.
I explore generative models, complex systems, symmetry, manifolds, and biological/physical modeling.
Open to scientific conversations & idea exchange.
DM if interested.
November 26, 2025 at 9:23 PM
Looking to connect with PhD students & scientists in AI, biology, physics, and math.
I explore generative models, complex systems, symmetry, manifolds, and biological/physical modeling.
Open to scientific conversations & idea exchange.
DM if interested.
I explore generative models, complex systems, symmetry, manifolds, and biological/physical modeling.
Open to scientific conversations & idea exchange.
DM if interested.
Reposted by Scientist
Our Editor-in-Chief John Lunn then introduced the inspiring Plenary Lecture from Prof. Jane Langdale (University of Oxford) on the "Regulation of venation patterning in grass leaves: understanding maize with a view to engineering rice” 🌾
#JXB75 #PlantScience 🧪 @janelangdale.bsky.social
#JXB75 #PlantScience 🧪 @janelangdale.bsky.social
September 18, 2025 at 8:42 AM
Our Editor-in-Chief John Lunn then introduced the inspiring Plenary Lecture from Prof. Jane Langdale (University of Oxford) on the "Regulation of venation patterning in grass leaves: understanding maize with a view to engineering rice” 🌾
#JXB75 #PlantScience 🧪 @janelangdale.bsky.social
#JXB75 #PlantScience 🧪 @janelangdale.bsky.social
Reposted by Scientist
Welcome to the Science feed!
Please read our FAQs for instructions for how to be added as a contributor: bossett.io/science-feed/
Mod introductory posts linked below⬇️
Use the test tube emoji on posts you want to appear in the feed🧪
Please like the feed and make sure you follow our feed rules:
Please read our FAQs for instructions for how to be added as a contributor: bossett.io/science-feed/
Mod introductory posts linked below⬇️
Use the test tube emoji on posts you want to appear in the feed🧪
Please like the feed and make sure you follow our feed rules:
February 10, 2024 at 3:08 PM
Welcome to the Science feed!
Please read our FAQs for instructions for how to be added as a contributor: bossett.io/science-feed/
Mod introductory posts linked below⬇️
Use the test tube emoji on posts you want to appear in the feed🧪
Please like the feed and make sure you follow our feed rules:
Please read our FAQs for instructions for how to be added as a contributor: bossett.io/science-feed/
Mod introductory posts linked below⬇️
Use the test tube emoji on posts you want to appear in the feed🧪
Please like the feed and make sure you follow our feed rules:
Reposted by Scientist
Physics is viewed as the key to understanding everything, but is this fundamentally wrong? | iai.tv/articles/rea...
Leading physicist George Ellis argues that reality cannot be predicted by studying simple physical constituents.
#philsci ⚛️ 🧪
Leading physicist George Ellis argues that reality cannot be predicted by studying simple physical constituents.
#philsci ⚛️ 🧪
Reality goes beyond physics: Why physics will never explain everything | George Ellis
Of all the sciences, physics has been seen as the key to understanding everything. As Feynman said, “physics is the fundamental science.” But in this article, one of the world’s leading physicists,…
iai.tv
May 1, 2025 at 1:03 PM
Physics is viewed as the key to understanding everything, but is this fundamentally wrong? | iai.tv/articles/rea...
Leading physicist George Ellis argues that reality cannot be predicted by studying simple physical constituents.
#philsci ⚛️ 🧪
Leading physicist George Ellis argues that reality cannot be predicted by studying simple physical constituents.
#philsci ⚛️ 🧪
Reposted by Scientist
Unlocking the mysteries of disease: New method uses 3D protein info to better predict where key molecular changes occur, improving accuracy by 1.4 times compared to older models. This advancement could lead to breakthroughs in understanding and treating diseases. Explore the details on GitHub.
Enhanced O-glycosylation Site Prediction Using Explainable Machine Learning Technique with Spatial Local Environment.
Published in Bioinformatics (Oxford, England)
doi.org
January 29, 2025 at 8:00 PM
Unlocking the mysteries of disease: New method uses 3D protein info to better predict where key molecular changes occur, improving accuracy by 1.4 times compared to older models. This advancement could lead to breakthroughs in understanding and treating diseases. Explore the details on GitHub.
Reposted by Scientist
MiniMax-VL-01: new, large MoE with 456B / 46B active params.
Most interestingly, uses hybrid attention: 7 linear transformer blocks (Lighting-Attention) followed by 1 standard. Can serve with 1M context length on a single 8x80GB node, and they test it out to 4M context length 🤯
Most interestingly, uses hybrid attention: 7 linear transformer blocks (Lighting-Attention) followed by 1 standard. Can serve with 1M context length on a single 8x80GB node, and they test it out to 4M context length 🤯
January 15, 2025 at 3:01 PM
MiniMax-VL-01: new, large MoE with 456B / 46B active params.
Most interestingly, uses hybrid attention: 7 linear transformer blocks (Lighting-Attention) followed by 1 standard. Can serve with 1M context length on a single 8x80GB node, and they test it out to 4M context length 🤯
Most interestingly, uses hybrid attention: 7 linear transformer blocks (Lighting-Attention) followed by 1 standard. Can serve with 1M context length on a single 8x80GB node, and they test it out to 4M context length 🤯
Reposted by Scientist
We’re at a weird stage with LLMs where they’re sooo cheap to run but quite expensive to verify, but also they’re not good or reliable enough that we can use them for anything important without verification.
January 13, 2025 at 6:17 PM
We’re at a weird stage with LLMs where they’re sooo cheap to run but quite expensive to verify, but also they’re not good or reliable enough that we can use them for anything important without verification.
Reposted by Scientist
Innovative research demonstrates how magnetic resonance imaging (MRI) can track immune cells during cancer treatment, offering real-time insights into tumor response. scft.link/UiObS
Work by:
✍️ Fanny Chapelin
✍️ Harrison Yang
✍️ Brock Howerton
Work by:
✍️ Fanny Chapelin
✍️ Harrison Yang
✍️ Brock Howerton
Tracking cancer-fighting immune cells with advanced MRI technology | Science Featured Series
A groundbreaking study introduces a new method to track the body’s immune response during cancer treatments using magnetic resonance imaging, potentially tr ...
scft.link
January 3, 2025 at 3:00 PM
Innovative research demonstrates how magnetic resonance imaging (MRI) can track immune cells during cancer treatment, offering real-time insights into tumor response. scft.link/UiObS
Work by:
✍️ Fanny Chapelin
✍️ Harrison Yang
✍️ Brock Howerton
Work by:
✍️ Fanny Chapelin
✍️ Harrison Yang
✍️ Brock Howerton
Reposted by Scientist
Considerations for building and using integrated single-cell atlases https://www.nature.com/articles/s41592-024-02532-y 🧬🖥️🧪 https://github.com/lueckenlab/single-cell-papers-trends
December 13, 2024 at 7:30 PM
Considerations for building and using integrated single-cell atlases https://www.nature.com/articles/s41592-024-02532-y 🧬🖥️🧪 https://github.com/lueckenlab/single-cell-papers-trends
Reposted by Scientist
African mealworms digest polystyrene, breaking it into simpler compounds through gut bacteria. 💡This is an opportunity for scientists to explore solutions - growing worms, harnessing bacteria, or isolating genes to combat plastic waste!💡
www.livescience.com/animals/inse...
www.livescience.com/animals/inse...
Plastic-eating mealworms native to Africa discovered
Larvae of the Kenyan lesser mealworm found to feast on polystyrene then break it down in their guts.
www.livescience.com
December 5, 2024 at 6:17 AM
African mealworms digest polystyrene, breaking it into simpler compounds through gut bacteria. 💡This is an opportunity for scientists to explore solutions - growing worms, harnessing bacteria, or isolating genes to combat plastic waste!💡
www.livescience.com/animals/inse...
www.livescience.com/animals/inse...
Reposted by Scientist
From our recent paper: 7-color Tetbow image of the neocortex🌈 We can distinguish hundreds of neurons based on the combinatorial expression of 7 FPs.
Here is the list of plasmids for Tetbow experiments if you are interested (all available from Addgene): sites.google.com/site/seedbre...
Here is the list of plasmids for Tetbow experiments if you are interested (all available from Addgene): sites.google.com/site/seedbre...
December 11, 2024 at 1:46 AM
From our recent paper: 7-color Tetbow image of the neocortex🌈 We can distinguish hundreds of neurons based on the combinatorial expression of 7 FPs.
Here is the list of plasmids for Tetbow experiments if you are interested (all available from Addgene): sites.google.com/site/seedbre...
Here is the list of plasmids for Tetbow experiments if you are interested (all available from Addgene): sites.google.com/site/seedbre...
Reposted by Scientist
What better time to announce a new paper than during NeurIPS and ACCV?
happy happy happy to introduce NADA, our latest work on object detection in art! 🎨
with amazing collaborators:
@patrick-ramos.bsky.social, @nicaogr.bsky.social, Selina Khan, Yuta Nakashima
happy happy happy to introduce NADA, our latest work on object detection in art! 🎨
with amazing collaborators:
@patrick-ramos.bsky.social, @nicaogr.bsky.social, Selina Khan, Yuta Nakashima
December 10, 2024 at 8:08 AM
What better time to announce a new paper than during NeurIPS and ACCV?
happy happy happy to introduce NADA, our latest work on object detection in art! 🎨
with amazing collaborators:
@patrick-ramos.bsky.social, @nicaogr.bsky.social, Selina Khan, Yuta Nakashima
happy happy happy to introduce NADA, our latest work on object detection in art! 🎨
with amazing collaborators:
@patrick-ramos.bsky.social, @nicaogr.bsky.social, Selina Khan, Yuta Nakashima
Reposted by Scientist
🚨 New preprint 🚨
Excited to share this preprint led by Zak Hussain (with @ruimata.bsky.social and Ben Newell), comparing the semantic contents of embedding models from text, brain, and behavior data using the psychNorms metabase: github.com/zak-hussain/...
Preprint: arxiv.org/pdf/2412.04936
Excited to share this preprint led by Zak Hussain (with @ruimata.bsky.social and Ben Newell), comparing the semantic contents of embedding models from text, brain, and behavior data using the psychNorms metabase: github.com/zak-hussain/...
Preprint: arxiv.org/pdf/2412.04936
December 10, 2024 at 9:17 AM
🚨 New preprint 🚨
Excited to share this preprint led by Zak Hussain (with @ruimata.bsky.social and Ben Newell), comparing the semantic contents of embedding models from text, brain, and behavior data using the psychNorms metabase: github.com/zak-hussain/...
Preprint: arxiv.org/pdf/2412.04936
Excited to share this preprint led by Zak Hussain (with @ruimata.bsky.social and Ben Newell), comparing the semantic contents of embedding models from text, brain, and behavior data using the psychNorms metabase: github.com/zak-hussain/...
Preprint: arxiv.org/pdf/2412.04936
Reposted by Scientist
💥RSC CNN Paper of the Week💥
Fast synthesis of DNA origami single crystals at room temperature
Tian* (Nanjing)
Yu et al, Chem Sci, 10.1039/D4SC07267G
#ChemSky #Science #NanoSky #ChemSci
Fast synthesis of DNA origami single crystals at room temperature
Tian* (Nanjing)
Yu et al, Chem Sci, 10.1039/D4SC07267G
#ChemSky #Science #NanoSky #ChemSci
Fast synthesis of DNA origami single crystals at room temperature
Structural DNA nanotechnology makes the programmable design and assembly of DNA building blocks into user-defined microstructures feasible. However, the formation and further growth of these microstru...
pubs.rsc.org
December 9, 2024 at 9:38 AM
Reposted by Scientist
New paper alert!
Zero-Shot Coreset Selection: Efficient Pruning for Unlabeled Data
Training models requires massive amounts of labeled data. ZCore shows you that you need less labeled data to train good models.
Paper Link: arxiv.org/abs/2411.15349
GitHub Repo: github.com/voxel51/zcore
Zero-Shot Coreset Selection: Efficient Pruning for Unlabeled Data
Training models requires massive amounts of labeled data. ZCore shows you that you need less labeled data to train good models.
Paper Link: arxiv.org/abs/2411.15349
GitHub Repo: github.com/voxel51/zcore
December 9, 2024 at 4:59 PM
New paper alert!
Zero-Shot Coreset Selection: Efficient Pruning for Unlabeled Data
Training models requires massive amounts of labeled data. ZCore shows you that you need less labeled data to train good models.
Paper Link: arxiv.org/abs/2411.15349
GitHub Repo: github.com/voxel51/zcore
Zero-Shot Coreset Selection: Efficient Pruning for Unlabeled Data
Training models requires massive amounts of labeled data. ZCore shows you that you need less labeled data to train good models.
Paper Link: arxiv.org/abs/2411.15349
GitHub Repo: github.com/voxel51/zcore
Reposted by Scientist
15/15 Paper: medrxiv.org/content/10.1...
Everything is open-sourced at github.com/snap-stanfor...
Talk at Stanford Graph Learning workshop: youtu.be/0_jdg7FqSE4?...
Also, happy to share KGWAS-preview has also won Best poster award at Stanford Bio-X and Reviewer's choice award at ASHG!
Everything is open-sourced at github.com/snap-stanfor...
Talk at Stanford Graph Learning workshop: youtu.be/0_jdg7FqSE4?...
Also, happy to share KGWAS-preview has also won Best poster award at Stanford Bio-X and Reviewer's choice award at ASHG!
Small-cohort GWAS discovery with AI over massive functional genomics knowledge graph
Genome-wide association studies (GWASs) have identified tens of thousands of disease associated variants and provided critical insights into developing effective treatments. However, limited sample si...
medrxiv.org
December 9, 2024 at 5:42 PM
15/15 Paper: medrxiv.org/content/10.1...
Everything is open-sourced at github.com/snap-stanfor...
Talk at Stanford Graph Learning workshop: youtu.be/0_jdg7FqSE4?...
Also, happy to share KGWAS-preview has also won Best poster award at Stanford Bio-X and Reviewer's choice award at ASHG!
Everything is open-sourced at github.com/snap-stanfor...
Talk at Stanford Graph Learning workshop: youtu.be/0_jdg7FqSE4?...
Also, happy to share KGWAS-preview has also won Best poster award at Stanford Bio-X and Reviewer's choice award at ASHG!
Reposted by Scientist
Check out our 3 NeurIPS 📌 in thread.
1. ViLCo-Bench: VIdeo Language COntinual learning Benchmark
🎥 neurips.cc/virtual/2024...
2. BTS: Building Timeseries Dataset
🎥 neurips.cc/virtual/2024...
3. Resolution-Agnostic Transformer-based Climate Downscaling
🎥 neurips.cc/virtual/2024...
#NeurIPS2024
1. ViLCo-Bench: VIdeo Language COntinual learning Benchmark
🎥 neurips.cc/virtual/2024...
2. BTS: Building Timeseries Dataset
🎥 neurips.cc/virtual/2024...
3. Resolution-Agnostic Transformer-based Climate Downscaling
🎥 neurips.cc/virtual/2024...
#NeurIPS2024
December 9, 2024 at 9:35 PM
Check out our 3 NeurIPS 📌 in thread.
1. ViLCo-Bench: VIdeo Language COntinual learning Benchmark
🎥 neurips.cc/virtual/2024...
2. BTS: Building Timeseries Dataset
🎥 neurips.cc/virtual/2024...
3. Resolution-Agnostic Transformer-based Climate Downscaling
🎥 neurips.cc/virtual/2024...
#NeurIPS2024
1. ViLCo-Bench: VIdeo Language COntinual learning Benchmark
🎥 neurips.cc/virtual/2024...
2. BTS: Building Timeseries Dataset
🎥 neurips.cc/virtual/2024...
3. Resolution-Agnostic Transformer-based Climate Downscaling
🎥 neurips.cc/virtual/2024...
#NeurIPS2024
Reposted by Scientist
Superpixel Tokenization for Vision Transformers: Preserving Semantic Integrity in Visual Tokens
arxiv: arxiv.org/abs/2412.046...
code: github.com/jangsoohyuk/...
Use superpixels instead of grid-based patches to tokenize images for ViTs.
arxiv: arxiv.org/abs/2412.046...
code: github.com/jangsoohyuk/...
Use superpixels instead of grid-based patches to tokenize images for ViTs.
December 9, 2024 at 11:56 PM
Superpixel Tokenization for Vision Transformers: Preserving Semantic Integrity in Visual Tokens
arxiv: arxiv.org/abs/2412.046...
code: github.com/jangsoohyuk/...
Use superpixels instead of grid-based patches to tokenize images for ViTs.
arxiv: arxiv.org/abs/2412.046...
code: github.com/jangsoohyuk/...
Use superpixels instead of grid-based patches to tokenize images for ViTs.
Reposted by Scientist
Project page: trellis3d.github.io
Paper: arxiv.org/pdf/2412.01506
Source code: github.com/Microsoft/TR...
Demo: huggingface.co/spaces/Jeffr...
Paper: arxiv.org/pdf/2412.01506
Source code: github.com/Microsoft/TR...
Demo: huggingface.co/spaces/Jeffr...
TRELLIS: Structured 3D Latents for Scalable and Versatile 3D Generation
trellis3d.github.io
December 9, 2024 at 10:07 AM
Project page: trellis3d.github.io
Paper: arxiv.org/pdf/2412.01506
Source code: github.com/Microsoft/TR...
Demo: huggingface.co/spaces/Jeffr...
Paper: arxiv.org/pdf/2412.01506
Source code: github.com/Microsoft/TR...
Demo: huggingface.co/spaces/Jeffr...
Reposted by Scientist
🧠 Brain stimulation helps partially paralysed patients walk again
Deep brain stimulation has been shown in a new @natureportfolio.bsky.social study to improve walking in spinal cord injury patients, with early trials demonstrating lasting recovery.
🔗 www.nature.com/articles/s41...
#SciComm 🧪
Deep brain stimulation has been shown in a new @natureportfolio.bsky.social study to improve walking in spinal cord injury patients, with early trials demonstrating lasting recovery.
🔗 www.nature.com/articles/s41...
#SciComm 🧪
Hypothalamic deep brain stimulation augments walking after spinal cord injury - Nature Medicine
Whole-brain anatomical and activity surveys identify the lateral hypothalamus as a key driver of recovery from spinal cord injury, leading to a deep brain stimulation therapy that augments the recover...
www.nature.com
December 9, 2024 at 11:45 AM
🧠 Brain stimulation helps partially paralysed patients walk again
Deep brain stimulation has been shown in a new @natureportfolio.bsky.social study to improve walking in spinal cord injury patients, with early trials demonstrating lasting recovery.
🔗 www.nature.com/articles/s41...
#SciComm 🧪
Deep brain stimulation has been shown in a new @natureportfolio.bsky.social study to improve walking in spinal cord injury patients, with early trials demonstrating lasting recovery.
🔗 www.nature.com/articles/s41...
#SciComm 🧪
