Lindvall Lab retweetledi
Few rigorous studies of large language models have been done in cancer care. Off-the-shelf models developed on general patient populations may need significant tuning. nature.com/articles/s4174…
English
Lindvall Lab
74 posts
Lindvall Lab
@lindvalllab
Physician-Investigator, Psychosocial Oncology and Palliative Care @DanaFarber
Boston, MA Katılım Nisan 2018
32 Takip Edilen254 Takipçiler
English
Lindvall Lab retweetledi
BioClinical ModernBERT is out!
Built on the largest, most diverse biomedical/clinical dataset to date
‼️Delivers SOTA across the board
Thrilled to be part of this effort led by @tsounack
Thomas Sounack@tsounack
Very excited to share the release of BioClinical ModernBERT! Highlights: - biggest and most diverse biomedical and clinical dataset for an encoder - 8192 context - fastest throughput with a variety of inputs - sota results across several tasks - base and large sizes (1/8)
English
ArXiv: arxiv.org/abs/2506.10896
HF Collection: huggingface.co/collections/th…
(3/3)
Français
🧠 Long context (8192 tokens)
📚 Trained on 53.5B tokens, the largest biomedical + clinical corpus ever used for an encoder
📈 SOTA on biomedical and clinical tasks
⚡ Fastest inference & fine-tuning
🔓 Released in base & large sizes with training checkpoints
(2/3)
English
We’re proud to announce BioClinical ModernBERT, led by our team at the @lindvalllab and collaborators!
(1/3)
English
Lindvall Lab retweetledi
Clinical encoders are joining the ModernBERT family ☺️
Thomas Sounack@tsounack
Very excited to share the release of BioClinical ModernBERT! Highlights: - biggest and most diverse biomedical and clinical dataset for an encoder - 8192 context - fastest throughput with a variety of inputs - sota results across several tasks - base and large sizes (1/8)
English
Lindvall Lab retweetledi
You can just continue pre-train things ✨
Happy to announce the release of BioClinical ModernBERT, a ModernBERT model whose pre-training has been continued on medical data
The result: SOTA performance on various medical tasks with long context support and ModernBERT efficiency
Thomas Sounack@tsounack
Very excited to share the release of BioClinical ModernBERT! Highlights: - biggest and most diverse biomedical and clinical dataset for an encoder - 8192 context - fastest throughput with a variety of inputs - sota results across several tasks - base and large sizes (1/8)
English
Lindvall Lab retweetledi
🚀Announcing BioClinical ModernBERT, a SOTA encoder for healthcare AI, developed by Thomas Sounack @tsounack for Dana-Farber Cancer Institute in collaboration with @Harvard, @LightOnIO, @MIT, @mcgillu, @AlbanyMed, @MSFTResearch.
Seamless continued pre-training enables SOTA performance on various clinical benchmarks, with the usual long context support and efficiency of ModernBERT
👉 Discover BioClinical ModernBERT now: lighton.ai/lighton-blogs/…
Kudos to @tsounack, @joshp_davis, @DurieuxBrigitte, @antoine_chaffin, @tompollard, @lehmer16, @alistairewj, @MattBMcDermott ,@TristanNaumann, @lindvalllab on this major leap for biomedical NLP!
English
Lindvall Lab retweetledi
Very excited to share the release of BioClinical ModernBERT!
Highlights:
- biggest and most diverse biomedical and clinical dataset for an encoder
- 8192 context
- fastest throughput with a variety of inputs
- sota results across several tasks
- base and large sizes
(1/8)
English
Excited to be representing @danafarber at #ASCO25 to share our findings on "Using large language models to assess adherence to ASCO patient-oncologist communication standards"
Learn more at Poster #125 today at 1:30 pm!
English
Lindvall Lab retweetledi
A locally deployable, open-source LLM-Anonymizer can remove personal identifiers with high accuracy, offering a scalable and accessible solution for secure medical data processing. Learn more: nejm.ai/4iFTASJ
English
Lindvall Lab retweetledi
Feasibility Study for Using #LargeLanguageModels to Identify Goals-of-Care Documentation at Scale in Patients With Advanced Cancer: brnw.ch/21wS5cT
Authored by @lindvalllab et al.
English
Lindvall Lab retweetledi
Editorial by @AdamRodmanMD and colleagues: When It Comes to Benchmarks, Humans Are the Only Way nejm.ai/4hMrAM9
#AIinMedicine
English
Great work! Our lab is actively exploring how LLMs can enhance palliative care. We recently demonstrated that LLMs can capture ACP domains directly from the EHR
jpsmjournal.com/article/S0885-…
Ravi B. Parikh@ravi_b_parikh
@NEJM @ASCO @oncologyCOA @JCO_ASCO @ramsedhom @realbowtiedoc @NCCN @miteshspatel @kevin_volpp @CassSunstein @CAPCpalliative 🌍 Implications: Algorithm-driven PC referrals scale in comm onc. Need better PC eligibility algorithms (esp incorporating symptom/psychosocial distress). Key area for #LLMs! @lindvalllab @kenlkehl @dbittermanmd @layerhealth Defaults can improve care beyond #oncology. (9/)
English
Lindvall Lab retweetledi
Perspective by E. Pierson (@2plus2make5) et al.: Using Large Language Models to Promote Health Equity nejm.ai/40gbwvn
#AIinMedicine
English
🚨Preprint🚨
Open-source LLMs outperform proprietary models in extracting clinical note sections!
📄 HPI, Interval History, Assessment/Plan
🏆 Llama 3.1 8B: F1=0.92 (internal), F1=0.85 (external)
✅ Cost-effective, private, accessible.
#AI #LLM
🔗 arxiv.org/abs/2501.14105
English
Lindvall Lab retweetledi
Read this week’s #EAPCblog by Charlotta Lindvall to find out more about AI and contributing to the upcoming special edition of Palliative Medicine on Digital Health in Palliative Care! bit.ly/3Wkllal @lindvalllab @amaranwosu @COstgathe @cewalshe @PalliativeMedJ
English
Non-Hispanic white patients experience less asymmetry (73%) than other racial/ethnic groups (82%). Bridging this gap is essential #HealthEquity. #OncologyResearch @PECthejournal
secure-web.cisco.com/1HDUB-Qk0YW9Jh…
2/2
English
🚨 New findings 🚨 : 77% of illness understanding discussions between oncologists and advanced cancer patients are clinician-dominated. 1/2
English