Josh Haimson

Co-founder @Inductive_Bio via @FlatironHealth @MIT_CSAIL | occasional takes on ai/ml, life sciences, music, & nerd puns... but mostly retweets

[Josh Haimson @Just_Joshing]

I’m excited to share that @inductive_bio has raised a $25M Series A to transform small molecule drug discovery with AI.

Grateful to @obviousvc, @a16zBioHealth, @Lux, @S32, @CharacterCap, @AminoCollective & our amazing team.

https://t.co/pgCWJ96PNH

[Josh Haimson @Just_Joshing]

Really interesting and suprising results from some of our experiments with #LLMs for #Chemistry

[Inductive Bio @inductive_bio]

LLMs’ impact in generative chemistry has been limited to date. We show that a slight change in how you ask the question, simply by switching the input from SMILES to IUPAC, can reduce invalid molecules by >90%.
Learn more: https://t.co/wLVx7rZh0Y
#GenerativeChemistry #AI #ML

[Jay Rughani @JayRughani]

High value AI application: Predicting properties of new drug candidates.

Congrats to the Nested Tx and Inductive Bio teams on this successful collaboration showing how new AI models can predict ADME properties and accelerate small molecule drug discovery.

[Josh Haimson @Just_Joshing]

With all of the excitement around #AlphaFold3, our team at @inductive_bio has been diving deep to understand where these advances might be useful for
#AI enabled small-molecule drug discovery in practice. Check out our latest blog to learn more: https://t.co/MhKfXYoD0W

[Josh Haimson @Just_Joshing]

I’m excited to start sharing more publicly what I’ve been working on for the last couple of years! At Inductive, we're focused on delivering better therapies to patients, faster, by accelerating compound optimization with fit-for-purpose datasets and ML

#ML #AI #DrugDiscovery

[Inductive Bio @inductive_bio]

We are thrilled to be coming out of stealth! See our press release (https://t.co/TdqeBLgFzS), along with some coverage from STAT in their Readout Newsletter (https://t.co/HA4BoPhYtF).

[Flatiron Health @flatironhealth]

Extracting meaningful clinical detail from EHRs for millions of patients with cancer is challenging.

@FlatironHealth uses #NLP & #ML to extract key information from unstructured documents in the curation of high quality #RWD.

Read more on our approach: https://t.co/kaHpx1pEhN

[Ben Birnbaum @benbirnbaum]

Finally got around to reading this upcoming @iclr_conf paper by @WenhaoGao1 @rociomer3 and @cwcoley: https://t.co/8IhLvZjMyV. I really liked this pragmatic yet principled approach to generative chemistry, and I highly recommend the full paper. Here’s my own summary:

[Josh Haimson @Just_Joshing]

This list is awesome! Very cool to see the ways in which #ML and #AI is being used to discover new drugs

[Ben Birnbaum @benbirnbaum]

Been diving into machine learning for small-molecule drug discovery. I've assembled some helpful papers, data sets, tools, blogs, Twitter accounts, and more on Github. Contributions welcome! https://t.co/U1JjfBtSBa

[Earth, Wind & Fire @EarthWindFire]

What are y’all doing tomorrow?
Asking for a friend.

[Natasha Jaques @natashajaques]

I couldn't resist. With contributions from @maxhkw

[François Chollet @fchollet]

People who say $GME is overvalued don't understand the value proposition. It's no longer a stock, it's a store of value. It's the new digital gold. And for now it has only 3% of the $BTC market cap, still plenty of room to grow. My price target is over $9,000

[Kexin Huang @KexinHuang5]

How to extract a clinical variable out of the entire patient’s history notes of > 200K words, using an expensive BERT-style model? 

Happy to share SnipBERT done @flatironhealth in NeurIPS ML4H! 

- up to >20% prediction gain
- clues for model interpretation

Details ⬇️

[Flatiron Health @flatironhealth]

Proud to partner with @huntsmancancer on this exciting research on how #MachineLearning plus EHR data can predict the likelihood of clinical outcomes, presented at #ASCOQLTY20 @ASCO #RWD
https://t.co/65r4EWcPCB

[Josh Haimson @Just_Joshing]

@zachweinberg Doesn't mean we ignore observational studies, but if RCTs point the other way we need to listen. The study people are now pointing to ironically concludes with: "Prospective trials are needed to examine this impact." https://t.co/zRboFXNgE2

[Josh Haimson @Just_Joshing]

@zachweinberg Also RCTs seem to show no effect for solo HCQ tx- ORCHID closed after 470 pts concluding: "the data from this study indicate that this drug provided no additional benefit compared to placebo control for the treatment of COVID-19 in hospitalized patients" https://t.co/CDrkB9k70I

[Josh Haimson @Just_Joshing]

@zachweinberg Fair point

[Josh Haimson @Just_Joshing]

@zachweinberg Wouldn't market dynamics just push rich people away from the MTA though? Price ⬆️ demand ⬇️ substitute seeking ⬆️

[Josh Haimson @Just_Joshing]

@peteyreplies @zoom_us Wonky answer, but this doesn't violate HIPAA. HIPAA security rules recommend encryption in transit and at rest, not end-to-end and the HIPAA privacy rules allow business associates (i.e. zoom) to view protected health information in service of treatment, payment or operations.

[Josh Haimson @Just_Joshing]

@KyleJudah @AledadeACO @Farzad_MD was trying to coordinate for their network of primary care providers https://t.co/MEPv3sDpo2

[Farzad Mostashari @Farzad_MD]

1/ At @AledadeACO we partner w independent practices on the frontlines of this epidemic, caring for over 500,000 seniors most at risk from #COVID19

We asked docs in 100+ primary care practices "what worries you most about COVID-19 today?"

Their answers should worry us