
AI Engineering @bmsnews, previous management consultant and systems biologist - lover of digital biomedical research, views always my own
Philadelphia, PA
Joined May 2012
- Tweets1.3K
- Following580
- Followers348
- Likes750
my computer sciences career arc has really come full circle - taught myself html 28 years ago and here am I now generating all reports and presentations in this format #AI
Haven’t touched anything @meta related in probably ten years but exploring this new muse agent…
Excited to be part of a panel today discussing, "From bad #agents to game-changing #foundationmodels in #AI #drugdiscovery - digitalinsights.qiagen.com/a…
All I want from #AI is copy paste to always match style and formatting
Definitely the most scientifically aware out of the box foundational model I've tested with translational medicine based questions #LLMs #AI
Anthropic is so back. Two things I like the most about Claude-3's release:
1. Domain expert benchmarks. I'm much less interested in the saturated MMLU & HumanEval. Claude specifically picks Finance, Medicine, and Philosophy as expert domains and report performance. I recommend all LLM model cards to follow this, so that the different downstream applications know what to expect.
2. Refusal rate analysis. LLMs' overly cautious answers to innocent questions are becoming a pandemic. Anthropic is typically on the ultra safe end of the spectrum, but they recognize the problem and highlight their efforts on it. Bravo!
I love that Claude dials up heat in the arena that GPT and Gemini dominate. Though keep in mind that GPT-4V, the high water mark that everyone desperately tries to beat, finished training in 2022. It's the calm before the storm.
Artificial intelligence (#AI) and machine learning (ML) continue to push the boundaries of what is possible in marketing and sales.
Now, generative AI promises to disrupt the way B2B and B2C players think about customer experience, and growth. Learn more: mck.co/45Jwcwu
The OG human gain/loss of function mutation #genetics
Wahoo for #opendata that is well structured, >1000 #drug #target pairs - direct link to data here: zenodo.org/record/7378746
Novartis researchers have shared the Secondary Pharmacology Database, a new resource that elucidates associations between drug bioactivity and potential adverse drug reactions (ADRs). Thank you, team, for this useful tool! nature.com/articles/s41467-0…
Huge, if reproducible #molecularbiology101
😲 Time to revise textbooks 🤔
Title: "Transcription factors interact with RNA to regulate genes"
"at least half of TFs also bind RNA, doing so through a previously unrecognized domain"
cell.com/molecular-cell/pdf/…
The first person I ever hired I met on @Twitter (hi @DanRyder3) - can my network help me again for this new open role? careers.bms.com/jobs/R157028… #biopharma #data #digital
We like the frameworks #deeplearning
Applications of Transformers
New survey paper highlighting major applications of Transformers for deep learning tasks. Includes a comprehensive list of Transformer models.
arxiv.org/abs/2306.07303
Great listen, lots of #AI #ML use cases from across the biopharma pipeline discussed (personal favorite liver fat prediction via @uk_biobank data🤖)
On today's Cross Validated podcast, we have Kim Branson, Senior Vice President and Global Head of AI and ML at GSK. GSK is one of the 10 largest pharmaceutical companies globally. @altcap
In this episode, we discuss:
- Applying ML to genetic data, clinical trials and more + gen AI to processes
- Building on open source technologies and reusable components in models
- Data and data generation as a critical asset
- Regulatory framework on AI
Subscribe for the next episode: t.ly/6imFA
Don't come across this gene every day, here is an informative 2013 paper including Figure 1 with a visualization of the USP1 complex as a regulator of some cellular processes ncbi.nlm.nih.gov/pmc/article…
ISM3091, Insilico's novel synthetic lethality inhibitor targeting USP1, receives FDA IND approval. Designed by AI, it has shown +results in preclinical trials across multiple tumor cell lines and in vivo models & could tackle resistance to PARP inhibitors. lnkd.in/gBZzgP3t
that whole citizen #datascientist promise actually coming to fruition (yes I've been efficiently developing python code 🤯)
With Windows Copilot, every Windows user can be a power user. blogs.windows.com/windowsdev…




















