🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Told to Alena from #Odessa by a #Kherson resident "In my apt we got shelled, a woman in the 1st apt lost her legs. Another man's leg burst; another had his artery injured; there was no one to stop the bleeding none to save them." I don't know how to live with it, but I have to.

Morning in Kyiv is when you go to the balcony with your coffee, and a russian cruise missile passes just over your head.

Yesterday more than 100 rockets launched at #Ukrainian cities, #UkraineTrustChain team on the ground sent us this reply. "Yep, still waiting for lights to come back on so that I can finish your report. Can't get to the bank info. If the power comes back on, will finish today."

Pavel's #UkraineTrustChain team they delivered and distributed 3,500 packages of humanitarian aid last week. Look at the crowds of people that showed up to get aid in a small town town in #Poltava region. #StandWithUkraine ukrainetrustchain.org/donate

Quite the story from #Ukraine: Ksenia's team was in a small village doing repairs and got stuck. A helpful tractor driver pulled them out using a downed power line as a tow cord. #UkraineTrustChain

Please join our Problem Summarization task at the BioNLP Workshop in ACL this summer in Toronto. We want to start designing for computerized diagnostic decision support, and augmentation with NLP is the future!lnkd.in/gs9iHJuG #nlp Dmitriy Dligach yanjungao Tim Miller

We are hosting an EHR summarization task in #BioNLP workshop at #ACL2023 ! Welcome to sign up for the task here: docs.google.com/forms/d/12DpfT…

Excited to share the first set of clinical NLP tasks to help systems design for clinical decision support for bedside care. We need more than just knowledge representation in our models: sciencedirect.com/science/articl… UW-Madison Critical Care Medicine Data Science Lab yanjungao Tim Miller Dmitriy Dligach

Proud to present: DR.BENCH, the first clinical NLP benchmark for Diagnostic Reasoning, a critical cognitive process for clinical diagnostic decision-making. A pathway to next-gen AI for clinical diagnostic decision-support 🩺🖥️

#ChatGPT has taken #NLProc by storm 🤖. We investigated time, data, and performance trade-offs of ChatGPT family of models in 2 health NLP tasks beyond QA: health advice classification and causal reasoning. ⚕️ [1/6] arxiv.org/abs/2304.02496 @ShaaanChen Brigham and Women’s Radiation Oncology CHIP Informatics

⏲️Time investment for prompt development limits practical value compared to smaller models. ChatGPT models required 100x more time to achieve best performance. Even for zero-shot settings, inference time is longer and yields worse performance than BoW and BioBERT models. [3/6]

When I hear that LLMs are "just doing next word prediction" this Stephen King quote comes to mind: A radio talk-show host asked me how I wrote. My reply—“One word at a time”—seemingly left him without a reply. I think he was trying to decide whether or not I was joking. I wasn’t.

Please spread the word: Our UW ICU Data Science Lab is hiring a new NLP-focused Postdoctoral Research Associate to help our growing lab! Apply this July! linkedin.com/jobs/view/3653…

We are hiring! Do you have NLP expertise with application of large language models/foundation models in clinical care? Then apply to our unique, multidisciplinary team: Check out this job at University of Wisconsin-Madison: linkedin.com/jobs/view/3653…. UW-Madison Critical Care Medicine Data Science Lab UW School of Computer, Data & Information Sciences

Great opportunity for an nlp expert to work on real clinical problems and collaborate with a multi-disciplinary team (including myself)!

Those fired linguists can finally get their jobs back. Every time I tweak the prompt of an LLM, the performance goes up by one point.

Who is more creative - people or AI? We find that GPT-4 is by far more creative than humans on standard tasks of divergent thinking #creativity #chatgpt4 #ai psyarxiv.com/6py2m/

#LoyolaChicago #ComputerScience has an open full-time, tenure-track Asst Prof position for Fall 2024 in #Algorithms, #ProgrammingLanguages, #FormalMethods, #TheoryOfComputation, and #SoftwareEngineering. For full consideration, apply by Jan 2, 2024, via careers.luc.edu/postings/26849.

We have a NEW PAPER in Nature Medicine on reporting recommendations for addressing the unique challenges of #largelanguagemodels (LLMs) in biomedical applications nature.com/articles/s4159… #StatsTwitter #MedTwitter #artificialintelligence #generativeAI #transparency

This is a great opportunity to work with a brilliant PI…don’t miss your chance