Healthcare policy and IT, with special focus on eldercare, care coordination, end of life. Analytics, AI, NLP, corpus linguistics. I write books now and then.
Joined January 2008
- Tweets 8,127
- Following 262
- Followers 2,218
- Likes 7,272
29 Photos and videos
Les Morgan retweeted
21 Feb 2025
We’ve updated our Global-MMLU paper to include Global-MMLU Lite ✨—a faster, more efficient eval set. Plus, we’ve added to the eval harness!
It covers a balanced subset of languages with an equal mix of Culturally Sensitive (CS) & Culturally Agnostic (CA) questions per language
5 Dec 2024
Is MMLU Western-centric? 🤔
As part of our cross-institutional work:
🥢 We conduct a large-scale cultural bias study on MMLU
🔍 Examine how cultural sensitivity impacts multilingual evaluations
🌍 Release Global-MMLU: a benchmark with MMLU translations in 42 languages
2
5
21
7,931
Les Morgan retweeted
23 Dec 2024
From @JAMANetworkOpen:
#AI-generated emergency department handoff notes, while promising for enhancing efficiency in patient transitions, still need further validation and integration with clinician oversight to ensure quality and patient safety.
ja.ma/4gQk3ff
2
5
768
Les Morgan retweeted
29 Nov 2024
New study shows LLMs outperform neuroscience experts at predicting experimental results in advance of experiments (86% vs 63% accuracy).
They use a fine-tuned Mistral 7B but other models worked too. Suggests LLMs can integrate scientific knowledge at scale to support research
18
118
669
53,091
Les Morgan retweeted
27 Nov 2024
Years of effort to get people to “have the conversation” @SWHPN @CAPCpalliative @CTACorg @ADVault_Inc @DrJoanneLynn @lesmorgan @endwellproject #hpm The Most Important Conversation to Have Before You Die nytimes.com/2024/11/27/well/…
1
2
40
Les Morgan retweeted
23 Nov 2024
A few weeks ago, with @LiamGMcCoy and @arjunmanrai published "LLMs and the Degradation of the EHR" (nejm.org/doi/full/10.1056/NE…), arguing that our headlong push to insert AI-generated text into the chart would have unforeseen (well, okay, foreseen to some of us!) consequences ...
1
6
18
3,518
Les Morgan retweeted
20 Nov 2024
BRIEFING | JAHF is holding a virtual briefing on 12/5 detailing #policy options the incoming administration should consider to better support #FamilyCaregivers. We invite you all to participate in this dialogue featuring @NA4Caregiving. Register: bit.ly/3NRGvbo
3
3
174
Les Morgan retweeted
20 Nov 2024
Every other Twitter-like has stalled out as a competitor to X, but it looks like the cascade to one of them is real. There is even meaty AI talk there now.
Alternatives are good, I guess, but a lot of good ideas are born from brokering across diverse networks, which harder now.
50
45
533
76,284
Les Morgan retweeted
19 Nov 2024
#Xodus: Bluesky Hits 20M Users as People Continue to Flee X hollywoodreporter.com/busine…
11
22
50
16,012
Les Morgan retweeted
If an individual relies on incorrect provider directory information and gets out of network services they expected to be in network, 1) cost sharing must be limited to in network rates indicates Cassie Thompson of @CMSgov at @WEDIOnline #NSA #ProviderDirectory webinar
1
40
Les Morgan retweeted
19 Nov 2024
1/ Introducing ᴏᴘᴇɴꜱᴄʜᴏʟᴀʀ: a retrieval-augmented LM to help scientists synthesize knowledge 📚 @uwnlp @allen_ai
With open models & 45M-paper datastores, it outperforms proprietary systems & match human experts.
Try out our demo!
We also introduce ꜱᴄʜᴏʟᴀʀQᴀʙᴇɴᴄʜ, a new large-scale multi-domain benchmark for scientific research synthesis, covering CS, Bio and Physics.
39
287
1,056
249,194
Les Morgan retweeted
19 Nov 2024
This #HospiceandPalliativeCareMonth, bring #advancecareplanning to your community! Host an event with @_ProjectTalk to engage underserved communities in #acp conversations.
We're in year 4 of our 5-year research study. Don't miss this opportunity! Apply: projecttalktrial.org
3
2
86
Editorial @David_Ouyang, MD, and Joseph Hogan, ScD (@jwhogan42): We Need More Randomized Clinical Trials of AI nejm.ai/4hfETpq
#ArtificialIntelligence
ALT Page 1 of the editorial "We Need More Randomized Clinical Trials of AI."
1
43
110
13,705
Les Morgan retweeted
18 Nov 2024
Health systems band together to test and publicly rank top AI models | @healthcaredive 11/14/2024
#AI #HealthAI #Healthcare #LLM #HealthTech #TechHartford
healthcaredive.com/news/heal…
1
30
Les Morgan retweeted
18 Nov 2024
Join NIA’s webinar, “Leveraging Large Datasets and Large Language Models to Improve Health Equity,” on Nov 22! Experts will share insights on best practices, opportunities, and ethical considerations in #healthcare #data and #AI.
Register today: go.nia.nih.gov/3CfHdfX
2
489
Les Morgan retweeted
18 Nov 2024
🔥 Excited to share our TrialGPT work published in @NatureComms today.
🧪 Patient recruitment is challenging for clinical trials.
🤖 We introduce TrialGPT, an end-to-end framework for zero-shot patient-to-trial matching with large language models.
nature.com/articles/s41467-0…
7
17
66
3,986
Les Morgan retweeted
18 Nov 2024
Researchers from NLM and @theNCI developed an AI algorithm that could identify relevant clinical trials based on eligibility criteria. The tool can help make it easier for both clinicians & patients to connect with the right clinical trial opportunities. loom.ly/MADnvSM
ALT Photo of Dr. Lu presenting the framework for TrialGPT
2
5
10
1,698
another day, another million new users on bluesky
blew right past 18M people today!
bsky.app/download
212
623
2,720
371,112
Here's a classic misleading AI headline from @nytimes: Chatbots Defeated Doctors at Diagnosing Illness
First, the context is clinical vignettes not real patients.
Second, the participants in the AI augmented human condition were not trained to use LLMs and leave a big open Q:
1
1
5
888
Les Morgan retweeted
17 Nov 2024
No one likes to be lectured about their social media habits. How about a pragmatic case then:
There has been a massive migration to Bluesky. Engagement there is now vastly higher (also: no Musk, no advertisement, no bots).
If you value my perspective, you’ll find it there:
2
18
50
4,837