Roger Ng, MD

Cardiologist | Board Certified in Echocardiography, Cardiovascular Diseases, Internal Medicine, Nuclear Cardiology, Cardiac CT | AI/ML Developer

[Patrick Dunn @yourheartscore]

Great session on digital tools in cardiac rehabilitation with @JRGolbus @SethShayMartin @DoctorMarvelMD @AnaBaracCardio @DrLopezJimenez @drMSupervia Dan Forman, Erica Schorr, Alexis Beatty, Asif Ali, Jedrek Wosic

[mTECH @mTECH_hopkins]

Opening remarks by @DoctorMarvelMD at #AHA23 presenting “Creating a Scalable, Sustainable Business Model with Remote Patient Monitoring and Digital Therapeutics”
@AHAScience @American_Heart @CiccaroneCenter @hopkinsheart @HopkinsMedicine @CorrieHealth @heartmaryland

[Patrick Dunn @yourheartscore]

Great session on health tech research @SethShayMartin @ErinMichos @JRGolbus @NinoIsakadze @AJBloodMD

[Roger Ng, MD @rkchee]

@yourheartscore @glfceo @DrDave01 @dcianfloneMD @StopHeartDz @Cleerlyhealth @AccurKardia The team at NOBELMIND and I are deeply honored to be opened by Dr. Michael V. McConnell, MD, Professor of Medicine, #Stanford #Cardiology #stanfordmedicine at #AHA23 Sharing how we use Large Language Models in being there for the right patient, the right time, with right care

[Patrick Dunn @yourheartscore]

Great session on Ai @glfceo @DrDave01 @rkchee @dcianfloneMD @StopHeartDz @Cleerlyhealth @AccurKardia

[Patrick Dunn @yourheartscore]

getting ready for the Health Innovation Pavilion and Health Tech Competition at #AHA23 in Philadelphia #CHTI

[Seth Martin @SethShayMartin]

Look forward to connecting with many friends and colleagues at #AHA23! Check our our team's presentations below!

[AHA Science @AHAScience]

#AHA23 packing checklist:

✔️Sunday, November 12th – Wear Red Day❤️
✔️Monday, November 13th – Sneaker Day👟

We can’t wait to see your ‘fit checks!

[Robert Harrington @HeartBobH]

Heading to #AHA23 Enjoying the ease of east coast @Amtrak travel. Beats flying! Looking forward to some great science and seeing friends and colleagues. @AHAScience @American_Heart @dramitkhera @Joseph_C_Wu

[American Heart Association @American_Heart]

Increase your knowledge of maternal cardiovascular risks before, during and after pregnancy with this free accredited learning module. Register for the accredited course at https://t.co/rukJQR6R8I

[Stephanie Cohen, MD @StephCohenMD]

Finished!

[Lior Alexander @LiorOnAI]

This is mind blowing technology. Generative AI will completely change how films are made.

From: @Flawlessai

[Roger Ng, MD @rkchee]

Discovering that breakthrough in AI is like https://t.co/dS4mifLO5W

[Roger Ng, MD @rkchee]

@elonmusk What happens if you tried just empty space

[AK @_akhaliq]

Fine-Grained Human Feedback Gives Better Rewards for Language Model Training

paper page: https://t.co/ac8ZcSPrKB

use fine-grained human feedback (e.g., which sentence is false, which sub-sentence is irrelevant) as an explicit training signal. We introduce Fine-Grained RLHF, a framework that enables training and learning from reward functions that are fine-grained in two respects: (1) density, providing a reward after every segment (e.g., a sentence) is generated; and (2) incorporating multiple reward models associated with different feedback types (e.g., factual incorrectness, irrelevance, and information incompleteness). We conduct experiments on detoxification and long-form question answering to illustrate how learning with such reward functions leads to improved performance, supported by both automatic and human evaluation. Additionally, we show that LM behaviors can be customized using different combinations of fine-grained reward models.

[Roger Ng, MD @rkchee]

Contributing to one of the most high quality-projects to advance AI in the world is like getting a chance to make history with you. Thank you @neurosp1ke @ykilcher

[Roger Ng, MD @rkchee]

Thank you for all ur amazing podcasts @lexfridman so fortunate to be included and engaged in your circle and the people that drive AI. A striking observation I realized is LLM behaves differently than all of DL. Its obj is not stationary but changed by each input

[Roger Ng, MD @rkchee]

@Kronenberg211 @ylecun Very clever. Assume intelligence prefers more intelligence, we train a model recognizing what intelligence looks like (GANS 0 vs 1 for intelligence). @ilyasut Can we then distill the evolutionary objectives from interesting properties of unsupervised learning (work backwards)

[Yann LeCun @ylecun]

In many animal species, evolution encodes objectives.
It is up to the individuals to figure out the behaviors that optimize these objectives.

Encoding objectives rather than behaviors is much simpler, more efficient, and more adaptive way for evolution to specify complex behaviors.

the same way it is much easier for human engineers to specify a loss function and rely on optimization for a system to perform perception or control than it is to design a perception or control system from scratch.

[David E. Albert, M.D @DrDave01]

A lot of exciting news today. I am proud to announce FDA clearance for the use of our Kardia 6L to monitor QT enabling a new service for instant QTc assessment anywhere & anytime. This innovation will boost patient safety. https://t.co/Nt6zu8ZeOa