Lisa JIN

Some great statistics buried in the case studies on how digital technologies and AI are improving the patient and workforce experience, as well as the health outcomes of the communities they serve!! #digitalhealth #healthcare #health

9

JCHI's Christopher Longhurst emphasizes the importance of #ethicalAI development and deployment, guided by principles that ensure fairness, safety, and effectiveness during a recent convo with Tal Behar of PMWC - Precision Medicine World Conference. See more below!

9

🌟Women in Tech: if you are one, I am here with you. I am eager to hear your story - feel free to share in the comment section.🌟 In my undergrad engineering class of 40 students, there were only six women. Fast forward to my Ph.D. program, I was one of three women in a program of 20 . At my first corporate job, I was one of just two women on a floor of 35 engineers. This is not just my story, it's the story of a lot of women who came before me and along with me. And then, it gets even lonelier as you climb the corporate ladder into leadership roles. But despite the numbers, the drive to excel has always been there. That hunger to push boundaries, achieve success, and pave the way for others. Over the course of my career, I’ve been fortunate to: Earn a fellowship award for grad school - completed my Ph.D in <4 years! Win the Employee Recognition Award at IBM - and I got a Super Man cape (!?!) for it and some equity award (got rid of the cape 6 years later :) and didn't wait for equity to vest)! Become a trusted leader in the AI/ML space - Went from Lead DS to ML Manager to AI/ML Director in 3 years at Centene Corporation, a fortune 25 company! It’s never just about the accolades; it's about representing those who continue to be the minority in rooms filled with opportunity, pushing forward and claiming our space. ✨ Let’s make sure your voice is heard. ✨ #WomenInTech #TechLeadership #AI #ML #Coaching #CareerGrowth

40

2 Comments

This week, Bara Badwan is representing us on the West Coast at The Alliance for Artificial Intelligence in Healthcare (AAIH) and sharing our work on how our patented #AI solution identifies promising early-stage #biopharma assets and companies. You can read more about that here: https://lnkd.in/eTJReqfj. #pharmaceutical #portfoliostrategy

16

1 Comment

Excited to be back in San Francisco for #AMIA2024, reconnecting with old friends and meeting new ones! Being here is nostalgic, as the Bay Area is where my journey in the U.S. began back in 2011. San Francisco feels like the home court of #AI, and the energy at this year’s symposium has been inspiring, particularly in exploring AI’s transformative impact on #healthcare. I’m grateful to share my work on using #GenAI to identify suicidal phenotypes in clinical notes, contributing to the advancing role of augmented intelligence in mental healthcare. Reflecting on some key takeaways: 1. #Informatics in #SafetyNetHospitals: There’s a growing presence of informatics in safety-net hospitals, yet creating a true #learninghealthsystem for underserved populations remains a work in progress. Building leadership and community networks within this unique care setting is essential. 2. Combining Expert Knowledge and Data-Driven Methods: Addressing the complex challenge of #suicide requires both expert insight and data-driven approaches. Extracting more social determinants of health (#SDoH) and personal narratives can help restore a fuller picture of the #mentalhealth perspective. 3. Limitations of #AmbientDocumentation: While ambient documentation holds promise, it’s still far from matching the depth and nuance of clinician-created records. Technology can aid documentation, but it cannot replace the sophisticated cognitive processes clinicians bring to their notes and orders. Language processing should go beyond words to address unspoken, cognitive layers of patient care. A heartfelt thanks to my mentors for their invaluable support and guidance on this journey! #hongfangliu #minghuang #huaxu #uthealth

39

2 Comments

Yay! Another patent in the books, who me? 💅🏻 Joking aside, innovation in healthcare is a constant process of incremental progress made by teams of experts (with the occasional big win). What an immense privilege to work with some of the best in the field on this incremental, yet important, piece of machine learning technology.

95

9 Comments

🚀 natHacks 2024 – From Cardio to Focus Timers! This past November, my incredible teammates Ali Senol, Willy Zuo, Arno Xu, and Sahi Advi Islam and I teamed up at NeurAlbertaTech's natHacks 2024 neurotech hackathon to explore an ambitious idea: leveraging EEG brainwave data to enhance productivity and performance. We initially set out to analyze beta wave activity during cardio workouts to optimize exercise sessions. The concept was to monitor brain activity in real time and recommend changes in exercise intensity—or even suggest a break—if irregular beta waves indicated a drop in focus or performance. However, we quickly realized that head movements during cardio introduced excessive noise into the Muse 2’s EEG signal, making the data unreliable. By Day 2, we pivoted to a more feasible application: creating an EEG brainwave-based Pomodoro focus timer. My idea was to move beyond the traditional 25-minute study session followed by a 5-minute break. Instead, we aimed to analyze brainwaves to assess the user’s focus levels in real time. If the user’s beta waves, which reflect concentration, consistently dropped below a certain threshold for a sustained period, the system would recommend taking a break. This approach allows for personalized study intervals tailored to the user’s actual focus levels, potentially making it far more effective than a rigid timer. Using the BrainFlow Python library, we successfully streamed EEG data into a file, applied Fast Fourier Transforms (FFTs) for signal analysis, and began building a front end with Flask to dynamically visualize focus levels. I implemented the EEG data streaming process to ensure accurate capture and storage of signals and collaborated with my teammates to figure out the FFT calculations for signal processing. Although we didn’t fully complete the project, the experience was incredibly rewarding. I gained valuable skills in signal processing, tackled real-world BCI challenges, and learned how to adapt and innovate under time pressure. A big thank you to my amazing teammates and to natHacks for providing this incredible opportunity to dive into cutting-edge technology. Excited to continue developing this project in the future! Feel free to check out the code in my GitHub repo here: https://lnkd.in/g2-qWaBC #natHacks2024 #BrainFlow #EEGAnalysis #SignalProcessing #HackathonExperience #Teamwork #PivotingInTech

4

Watch this session and dive into the 2nd #WiDSDatathon challenge using MATLAB!

10

A few short days ago, I wrapped up my research position at UC Davis Health with Professor Farzad Fereidouni. My role at the lab entailed a specialization in machine learning and computer vision and I'm excited to share what I was able to accomplish with everyone. I worked on two projects during my time at the lab. The first one was something I called "Kernel Generation for Image Similarity" (https://lnkd.in/ghsY73ts). It was to process an image such that it would look like it was H&E stained. I learned a lot about the basics of histology and digital histology through this process. The most novel part of this approach was using Wiener deconvolution to generate a kernel that would be used to enhance image similarity (typically using SSIM, MSE or a combination of both). A big part of this process was optimization. I learned how to use Python libraries such as Cupy and other tools to GPU accelerate a lot of these image-processing optimization problems. My second project and my most meaningful contribution was called "FastSAM-needle-biopsy" (https://lnkd.in/gXXydq-X). The objective of this project was to use the FastSAM image segmentation network to accurately generate image segmentation masks of live tissue such that a microscope would only scan the relevant parts of an image. Here is step by step walkthrough of the image processing pipeline: 1) VB.NET (Darius) application calls a DLL (compiled from csharp.NET) 2) DLL calls FastSAM image segmentation code using Python.NET 3) FastSAM segments image and returns image segmentation bitmask and as well as the fastest path to visit every bit (Traveling Salesman Problem) 4) Python returns values are returned to DLL using JSON as a reliable data transfer solution  5) DLL returns an object that can be parsed by VB.NET FastSAM is a faster version of Meta Research's Segment Anything Model and I used it to greatly speed up development time on this project. It was a big challenge to integrate such cutting-edge, machine-learning driven, image processing pipelines into a legacy codebase that was written in Visual Basic .NET but thanks to tools such as DLLs, JSON, and C#, I was able to figure it out! I am very grateful for the time I spent in the lab. I feel like I developed some great programming skills, especially in computer vision, that will prove useful for the rest of my career. I'd like to give special thanks to Professor Fereidouni, Ruby Mascareno, Willy Ju, Dena Sayrafi, Ruben Gnanaruban, Inyola, Antara Mallick, and Jayla. The link is here if you'd like to see my lab report going over everything I did in more technical detail and with a few pictures! https://lnkd.in/gRpZVmjG #computervision #machinelearning #AI #histology

67

10 Comments

In a recent class at Santa Clara University Leavey School of Business, we explored the evolving role of AI in our lives and careers. Our exercise focused on evaluating AI-generated responses and their broader implications, including testing the AI model's ability to assist with completing tasks end-to-end. During the exercise, we observed that while AI responses often share similar contexts, variations arise depending on the specificity of the questions and the scope of the training data. For instance, when models were asked to generate a PowerPoint presentation from a document, it effectively summarized the content and highlighted key points. However, when asked to create a succinct presentation, some critical details were omitted—highlighting how precise input and iterative prompts can significantly enhance AI's utility. While working with AI, we can refine results through a series of prompts to achieve the desired outcome. However, there are tasks where models perform exceptionally well and others where they require more guidance. Our discussion emphasized AI's transformative role in reshaping industries. Beyond automating tasks, AI is driving innovation and creativity, creating opportunities that demand adaptability and expertise in leveraging AI for meaningful outcomes. Technology doesn't eliminate jobs—it transforms them. As future professionals, we must: 1. Continuously upskill in AI technologies 2. Develop skills that complement AI capabilities 3. Embrace lifelong learning as a core professional strategy By understanding AI's potential and limitations, we can position ourselves to contribute more strategically and creatively in an AI-augmented workplace. I'm grateful to be part of a classroom environment fostered by Professor Sriram Sundararajan, that fosters such thought-provoking conversations about the future. #AI #ArtificialIntelligence #FutureOfWork #LifelongLearning #Innovation #SantaClaraUniversity #SCULeavey #SCU #SantaClaraUniversityLeaveySchoolOfBusiness

22

1 Comment

🎉 Excited to share my first, first-author paper, "EHRMamba: Towards Generalizable and Scalable Foundation Models for Electronic Health Records," co-authored with my amazing colleagues Mahshid Alinoori, Arash Afkanpour, and Amrit Krishnan at the Vector Institute! 🚀 🏥 Electronic Health Records (EHR) capture detailed patient medical histories and hold immense potential for revolutionizing personalized medicine and clinical decision-making. However, deploying EHR models in real-world settings is challenging due to high computational demands and limited hospital resources. Traditional transformer models, with their quadratic computational costs, are impractical for processing extensive EHR sequences. Additionally, maintaining separate downstream models for each clinical task complicates deployability. 🐍 To overcome these limitations, we introduce EHRMamba, a robust foundation model built on the Mamba architecture. EHRMamba can process sequences up to four times longer than previous models due to its linear computational cost. We also introduce a novel approach to Multitask Prompted Finetuning (MPF) for EHR data, which enables EHRMamba to simultaneously learn multiple clinical tasks in a single finetuning phase, significantly enhancing deployment and cross-task generalization. Our evaluations on the MIMIC-IV dataset demonstrate that EHRMamba advances state-of-the-art performance across 6 major clinical tasks and excels in EHR forecasting, marking a significant leap forward in the field. 👨‍🍳 Alongside EHRMamba, we open-source Odyssey, a toolkit designed to support the development and deployment of EHR foundation models, with an emphasis on data standardization and interpretability. 🔥 Discover more about our work and explore EHRMamba and Odyssey: https://lnkd.in/eDYspsuA #EHRMamba #Odyssey #ElectronicHealthRecords #Healthcare #MachineLearning #AIInHealthcare #VectorInstitute

166

13 Comments

Thrilled to hear about this new resource! Having seen firsthand how challenging it can be to find a suitable mentor, especially as a first generation student, I know the profound impact a good mentor can have. If you are in search of a mentor or interested in providing mentorship, check out TISP!

3

2 Comments

❔ How is generative AI having a positive impact on healthcare? 💡 The doctors at UC San Diego Health are finding out. Although still in the experimental stage, Gen AI is showing potential to improve patient communications and prevent physician burn out. Doctors often receive over 200 queries a day with time only for brief, factual responses. With the assistance of AI, they can now transform these messages to include compassion, empathy, and personal engagement. This initial study identified promise in relieving physician stress and improving patient satisfaction. https://lnkd.in/eq3vGZBq #genai #AIinmedicine

2

AI tech from Viz.ai halved the stroke-transfer time at a UC Davis Health partner hospital. This reduction in door-in-door-out time is critical as timely stroke care is essential. UC Davis Health also utilizes this new AI technology. Learn more: https://lnkd.in/gys2xPug

1