Rachel Wities
Healthcare NLP researcher @Microsoft
Tel Aviv, Israel
Actions
Rachel Wities is an NLP Researcher at Microsoft Healthcare and Life Sciences group, and a Co-Founder of MeDS - Medical Data Science Israel community.
Rachel is a public speaker addressing Healthcare NLP challenges, and believes that understanding doctors and their needs is the key to successfully implementing AI Healthcare algorithms.
Holds an M.Sc from BIU NLP lab, and previously worked as a researcher at Zebra Medical and at PayPal.
Rachel is devoted to promoting women speaking opportunities at work and outside it: during the week she is one of the leaders of Microsoft Women Speakers program, and on Shabbat she is part of a feminist Partnership Minyan in her Jewish Orthodox community.
Loves her family, God and Oxford Comma jokes.
Links
Area of Expertise
Topics
Choosing the Third of Two Options: Hybrid AI Systems in NLP
So, you have an AI product, based on models you've trained with many resources and much thought (and perhaps some rule-based systems - don't be ashamed, everyone has those) and overall, you're pretty happy about it.
Then one day, comes the revolution of Large Language Models. The possibilities are endless! The sky opens up before you! What a time to be a data scientist!
But what will you do with the existing AI system, the invested and accurate one you've built up until now? Continue to develop it and ignore the LLMs? Discard it and start from scratch? Keep it as a backup for a rainy day?
In the last year, I've heard this dilemma repeatedly from data scientists and AI leaders. In Microsoft Healthcare and Life Sciences group we have experienced it as well, with our flagship Text Analytics for Health service.
In this lecture, I will resolve this dilemma by introducing the Hybrid AI approach and explaining how to build Hybrid AI systems that utilize both the strengths of LLMs and the advantages of your existing ML-based system. I will demonstrate three basic designs of such hybrid systems that we tried in our Data Science team, and shortly analyze their respective advantages and challenges.
At the end of the talk, you will no longer have to choose between LLMs and your current system - you will choose to enjoy the best of two worlds.
Beyond Medical Knowledge: Bridging the Gap between LLMs and Real-life Clinical Challenges
Medical professionals struggle with the burden of processing more and more data, mostly in the form of free text - good thing they have LLMs, which did so nicely on medical licensing examinations, to help them, right?
Wrong.
In order to create systems that integrate into real-life medical professional workflows, LLMs need to have much more than medical knowledge. We need to be able to evaluate their performance, add guardrails, explainability, and human-in-the-loop feedback to the system. All these things require not only a deep understanding of medicine, but also a deep understanding of doctors.
In this lecture, I will discuss these challenges based on my experience as a data scientist in Microsoft Healthcare group, and suggest promising ways to solve them. I will show how to use medical ontologies to achieve grounding; how to use semantic structuring for smart precision and recall medical free text evaluation; and how to use various augmentation techniques in the medical domain.
This talk is intended for data scientists interested in the medical data science domain, as well as for data scientists coping with similar problems in other domains.
Dr. GPT - LLMs in Healthcare
ChatGPT passed the US medical licensing exam with flying colors - but does it really understand medicine?
In this session I will talk about LLMs performance in different medical-related tasks, like medical notes generation, summarization and simplification, medical NER and medical coding, I will talk about the importance of evaluation and grounding, and how we can improve LL's performance in these tasks.
Explainability is Not Enough - What Doctors Want From AI
In this talk I will share from my experience as an NLP Researcher in a medical company how to make AI products that doctors actually trust and use - and contrary to general belief, it’s not only about explainability…
Doctor-in-the-loop: Interactive Machine Learning in Healthcare AI
Working in a Healthcare startup, one of my most frustrating experiences was to ask doctors to do tedious work of data annotation or result verification. Surely there’s a better way, I told myself, to exploit the knowledge and expertise of doctors, than to turn them into a labeling conveyor belt!
Well, it turns out there is.
Human-in-the-loop ML refers to human-machine interaction in data annotation and model training. In Zebra Medical we used Human-in-the-loop techniques to compensate for lack of tagged data and to better exploit clinical expert knowledge. In this lecture I will show how to make data annotation quicker and smarter by turning it into an interactive process, and how an interactive process of experts and models writing rules together can improve your model performance without additional training.
This talk is intended for AI researchers interested in better ways to exploit the knowledge and experience of domain experts, and for people interested in the challenges of AI in the Healthcare domain.
Please note that Sessionize is not responsible for the accuracy or validity of the data provided by speakers. If you suspect this profile to be fake or spam, please let us know.
Jump to top