Shadab Khan
@skhanshadab87
Followers
211
Following
1K
Statuses
442
Team Lead - Healthcare AI at M42.
Abu Dhabi, UAE
Joined September 2009
Excited to share Med42, by @M42Health - our open-access clinical LLM. It scores well on several benchmarks (USMLE - 72%, 0-shot), and available on @huggingface. Beyond benchmarks, we're on course to evaluate it on safety + real-world use cases. Please reach out for collaboration!
Today @M42Health, we are releasing Med42: the best open-source clinical LLM to date, closing the gap with GPT-4 and MedPaLM-2 ! 🔥🩺 It is available on @huggingface 🤗:
1
4
9
RT @BrandesNadav: New preprint claims that most existing DNA language models perform just as well with random weights, suggesting that pret…
0
65
0
RT @BiologyAIDaily: Genomic Foundationless Models: Pretraining Does Not Promise Performance 1. This study challenges the paradigm of pretr…
0
15
0
RT @kirill_vish: 🧬 Genomic Foundation Models (GFMs) rely on costly pretraining just like models in NLP & vision. However, is unsupervised…
0
7
0
RT @OpenlifesciAI: 🚨 Medical AI Research alert! 🚨 Can we evaluate LLM's medical capabilities across multiple modalities and specialties?…
0
9
0
@paimadhu UAE! Great flight connectivity all over the world, straightforward visa process, availability of large convention centers!
0
0
0
Sila is amazing to work with!
Postdoc position available @Harvard on medical image analysis using deep learning! Join us to develop cutting-edge tech that will revolutionize patient care. #DeepLearning #MedicalImaging #Postdoc #Harvard P.S. Help us spread the word!
0
0
1
We also are hosting at all levels for MLE and applied scientists! Please DM me of your interested!
0
0
0
@Michael_D_Moor @_akhaliq Interesting work! I particularly appreciated the effort on deduplicating for eval and finding leakage! Quick que: I noted the number (4721) and categories of the books (fig 8) but the source was missing. Would you mind mentioning the source? Will MTB released at some point?
0
0
0
RT @MaartenvSmeden: NEW PAPER Our review of guidance documents for AI based prediction models in healthcare is out. Guidance *not just* fo…
0
92
0
@BrianPogue19 @DartmouthMP Good to see the familiar faces and a familiar room! Happy holidays Brian!
0
0
0
RT @DrHughHarvey: Clinical validation for CE marking, a thread 🧵[1/5]... Many startups get confused about the requirements for clinical va…
0
16
0
Best thing I read today!
Totally agree! Anyone screening applications and any applicant thinking their CV is not representative of their skills/potentials, I think you might want to read the story of my own PhD application in this thread: 1/
0
0
5
RT @devendratweetin: Late to the party but happy to share our #NeurIPS2021 work:led by @matej_zecevic where we cons…
0
8
0
Huazhu is n outstanding researcher and amazing human being - highly recommended!
[We're hiring!!] We have several Research Scientist (AI for Healthcare) Positions available at IHPC, A*STAR, Singapore. If you are interested in working with us, please send your CV to me. Also welcome to share it with your colleagues! Thx. JD link:
1
0
4
RT @lpachter: It's time to stop making t-SNE & UMAP plots. In a new preprint w/ Tara Chari we show that while they display some correlation…
0
1K
0
@lpachter @neurobongo Intriguing results! Some readers of this paper might benefit from a reference to this 2016 @distillpub article (in addition to [3,6] ) which discussed misleading nature of tsne embeddings and provides an accessible (+interactive) introduction.
0
0
4
@karpathy At loss plateau, the model could have overfit. When you drop the LR by 10X, would you take the model at the end of the plateau, or would you select the model with best val performance (from anywhere on the training curve). "ReduceLROnPlateau" in PT does the former. Thoughts?
0
0
1
This from someone who was previously a lecturer at Harvard (and whose workshop lectures that I attended were extremely good). Advise to students - when in doubt, find and talk to good mentors.
We trained two cohorts of incredible students in India. Our best students there were totally on par with the best I trained at Harvard. And they were NOT all from eminent Indian technical Universities like the IITs.
1
0
2
RT @kdpsinghlab: Thank you for folks who have shared or commented on our paper. I know the paper is being used by some to dunk on Epic. Rat…
0
165
0