You can tell when someone has released an open-source large language model because it suddenly becomes impossible to get a GPU from AWS. It’s similar to what happened when the Mistral team released the 8x22B.
You can tell when someone has released an open-source large language model because it suddenly becomes impossible to get a GPU from AWS. It’s similar to what happened when the Mistral team released the 8x22B.
unpopular opinion: many AI models are just learning correlation.
The fact ppl often use random.seed(42) leads to GPT thinking 42 is the answer to a random number.
unpopular opinion: many AI models are just learning correlation.
The fact ppl often use random.seed(42) leads to GPT thinking 42 is the answer to a random number.
In a time when ML claims to predict anything from everything, working with Tinyi on method rooted in biology’s basics and statistics is a breath of fresh air. And, it is the collaboration with MSK clinical experts @ciacobu@CentralParkWMD Shigaki & Aki made it sound and feasible!
In a time when ML claims to predict anything from everything, working with Tinyi on method rooted in biology’s basics and statistics is a breath of fresh air. And, it is the collaboration with MSK clinical experts @ciacobu@CentralParkWMD Shigaki & Aki made it sound and feasible!
The 2017 solar eclipse: my first year phd @ the rooftop of MSKCC.
The 2024 solar eclipse: first year after phd from home.
Eclipses remained the same. What truly mattered? The people I was with ❤️.
Tried to solve a dataframe issue with 'the smartest' AI Claude3 Opus for 30min and it failed. The first google result from Stack Overflow solved the problem in 10s.
Tried to solve a dataframe issue with 'the smartest' AI Claude3 Opus for 30min and it failed. The first google result from Stack Overflow solved the problem in 10s.
Question for the computational biology community: Leiden algorithm has been great for scRNA-seq data, but it's not working well on very large datasets (like 10 million cells). It either creates too many clusters (1000+) or groups them incorrectly. Any suggestions?
This is a great piece of writing. Large Language Models work well because they can find answers in public data. But AI in science is different because often the needed answers and data aren't available. Sometimes, even the questions in science are not well structured.
This is a great piece of writing. Large Language Models work well because they can find answers in public data. But AI in science is different because often the needed answers and data aren't available. Sometimes, even the questions in science are not well structured.
There should be a group for people who work with single-cell data over a 10M cells. Thing changes at that scale and needs new practice (cpu->gpu, knn et al parameters change...).
359 Followers 3K FollowingPh.D Candidate at University at Buffalo @UBuffalo | Research Scientist Intern @Yahoo | Ex. Research Scientist Intern at Adobe Research @Adobe
142 Followers 1K FollowingPostdoc in Jiayang Li Lab, Institute of Genetics and Developmental Biology, CAS. Focus on the Multi-omics of dandelions & rice🌱🌾Recruiting Top-Tier Talents👇
170 Followers 3K FollowingFinal year Ph.D. student working on Out-of-Distribution Generalization and Causality of Large Pre-trained Models, and Graph Neural Networks.
909 Followers 2K FollowingInvestigating the #immunology of infection and vaccination with #singlecell #multiomics.
Curating single-cell news and conversations on https://t.co/RUQfOpNXNf
3K Followers 3K FollowingStaff Research Scientist at Google DeepMind. Former adjunct assistant prof at @NYU_Courant. PhD at @mldcmu. ML for Bio/Chem (Prev. NLP).
All opinions my own.
81 Followers 1K Followingl am for signs and wonders.
https://t.co/IyQoZenxmw.,https://t.co/FyX2nYPeoU. Microbiology.
I love research, you can consultant me on microbiology,enzyme, environmental mon
259 Followers 887 FollowingAssociate Director Oncology @RecursionPharma ($RXRX) | ex-@clevelandclinic | Boise State Alum | Heart for Folks with Disabilities | Comments are my own
26 Followers 425 FollowingAssociate professor at Pontificia Universidad Católica del Perú (PUCP). Principal researcher at the Artificial Intelligence Group (IA-PUCP)
467 Followers 855 FollowingA little bit of slope makes up for a lot of y-intercept. Formerly @wise, @revolutapp wealth & trading. Building @go_lightyear.
3K Followers 2K FollowingHorticulture Research is an open access international journal devoted to publishing original research results and reviews on the best research in the field.
22K Followers 4K FollowingScientific editor at Nature Methods. Ph.D. Also at @[email protected] 🔬 All tweets and comments represent my independent views. (she/her)
49 Followers 117 FollowingPhD student at Boeva Lab, @ETH_en | Mostly working on ML applied to cancer research | Previously @DKFZ, DKRZ, @uni_tue, @HKUniversity
259 Followers 887 FollowingAssociate Director Oncology @RecursionPharma ($RXRX) | ex-@clevelandclinic | Boise State Alum | Heart for Folks with Disabilities | Comments are my own
163 Followers 139 FollowingAssistant professor at Dartmouth College. PhD@Yale, PostDoc@UChicago. Disease genetics, genomics, statistical genetics, data science. We are hiring! She/her.
7K Followers 1K FollowingRay and Stephanie Lane Professor of Computational Biology at CMU School of Computer Science @CMUCompBio @SCSatCMU @CarnegieMellon
9K Followers 1K FollowingHarold Hodgkinson Professor of Biomedical Engineering and Pathology, Yale University https://t.co/kLH8WPg2c2 https://t.co/kbDKoBh9Ak https://t.co/MduFkomdDj
31K Followers 933 FollowingI’m Alex Graveley, creator of GitHub Copilot, AI Tinkerers, Dropbox Paper, MobileCoin, and Hackpad.
Building @ai_minion
Hiring https://t.co/nsHar8OLPC
3K Followers 802 FollowingReader @EdinburghUni. Wrote @QuPath & https://t.co/eem8QmCaK8
Views my own, unless I've changed my mind. Then perhaps not even mine.
77K Followers 15K FollowingI science & try to make science open. Tweets do not represent my employer's views, but are always correct. My conflicts of interest https://t.co/bZ7I9oSJe0.
10K Followers 59 Following@Stanford professor. Chan-Zuckerberg investigator. Sloan Fellow. AI for biotech + health. Making AI more trustworthy, reliable and human compatible.
3K Followers 4K FollowingDirector CDH @PrincetonDH; English Prof. @prosodyarchive. https://t.co/pAil6htB3E @mmvty.bsky.social I HATE ALDEN GLOBAL CAPITAL b/c I love journalism!
1K Followers 948 FollowingAssistant Professor & Director of Physician-Scientist Development @UTSWNeurology, co-founder of @AJIAorg, co-founder of @BrainDESTROKE