Suggestions
Yubin Park, PhD
Chief Builder at mimilabs | LinkedIn Top Voice | Advisor to Astrana Health | Ph.D., Machine Learning and Health Data
Professional Background
Yubin Park is a distinguished expert in the fields of machine learning and data mining, boasting an impressive career trajectory that emphasizes strategic innovation and leadership in technology. With an extensive background that encompasses critical roles in both academia and industry, Yubin has made significant contributions to healthcare technology and analytics. His professional journey is marked by a series of prominent leadership positions, including his role as Chief Technology Officer and Co-founder at Accordion Health, Inc., where he was pivotal in shaping the company's vision and technological strategy. Yubin also held the position of CEO at Accordion Health, guiding the organization through essential developments in analytics and data-driven healthcare solutions.
In his tenure as Managing Director of AI Solutions at Evolent Health, Yubin led initiatives that harnessed the power of artificial intelligence to enhance patient care and streamline healthcare operations. His commitment to leveraging data for better decision-making underscores his dedication to improving healthcare delivery systems. As Chief Data and Analytics Officer of Astrana Health, he pioneered innovative data solutions that addressed complex health challenges, demonstrating his exceptional ability to translate data into actionable insights.
Yubin's experience also includes advisory roles at organizations like Astrana Health and WaveBridge, where he provided strategic guidance and expertise to bolster data-driven initiatives. As an Adjunct Professor at Emory University, he shared his vast knowledge with the next generation of data scientists, fostering academic growth and encouraging research in machine learning applications in healthcare.
Education and Achievements
Yubin Park's educational foundation is deeply rooted in advanced studies in machine learning and data mining. He earned his Ph.D. and M.S. in these fields from The University of Texas at Austin, one of the leading institutions known for its commitment to technological advancement and research excellence. His academic journey began with a Bachelor of Science in Electrical Engineering from the esteemed Korea Advanced Institute of Science and Technology (KAIST). This rigorous educational background equipped him with a robust understanding of both theoretical and practical applications of engineering principles and data science.
Throughout his career, Yubin has been recognized for his profound insights into data analytics and health technologies, positioning him as a thought leader in the intersection of AI and healthcare. His approach to problem-solving is informed by a philosophy that emphasizes the cumulative effects of small, consistent actions—a principle that resonates with his favorite quote by Vincent van Gogh: "Great things are not done by impulse, but by a series of small things brought together."
Achievements
Yubin Park's achievements reflect his commitment to innovation and excellence in the tech industry, particularly within healthcare analytics. As a Chief Builder and Moderator at mimilabs, he facilitated collaborative efforts among experts to drive advancements in machine learning technologies. His strategic leadership has led to numerous successful integrations of AI solutions that have enhanced healthcare delivery and patient outcomes.
He has played an instrumental role in launching and scaling various startups within the health tech space, demonstrating his entrepreneurial spirit and dedication to developing pioneering solutions. His influence extends into the academic realm where, as an Adjunct Professor, he has inspired countless students to harness the potential of data analytics and AI in addressing critical healthcare challenges.
Through these roles and achievements, Yubin Park exemplifies the harmonious blend of technical expertise and visionary leadership essential for driving impactful change in today's healthcare landscape.
Highlights
Google's answer to hallucination: public datasets
While OpenAI's o1 model launch dominated headlines, the same day, Google quietly made a fascinating announcement [1]: "DataGemma: Using real-world data to address AI hallucinations," by Prem Ramaswami and James Manyika.
DataGemma combines Gemma, Google's open LLM model, with Data Commons, a comprehensive source of global statistical data. This new model enhances traditional LLMs with factual information using RIG (Retrieval Interleaved Generation) and RAG (Retrieval-Augmented Generation). Though still in the research phase, the results are already promising, according to the paper [2] by Prashanth Radhakrishnan, Jennifer Chen, Bo Xu, Prem Ramaswami, Hannah Pho, James Manyika, and R. Guha.
The most intriguing aspect is Google's implicit definition of "fact." By leveraging Data Commons, Google suggests that truth is rooted in public datasets from authoritative sources. The blog states:
"[Data Commons] sources this public information from trusted organizations like the United Nations (UN), the World Health Organization (WHO), Centers for Disease Control and Prevention (CDC) and U.S. Census Bureau."
Shifting the focus to healthcare, we're inundated with AI-related posts on LinkedIn every day—no single day passes without seeing someone talking about how LLMs will change healthcare; also, it turns out I may be adding more noise to it ;). We're experiencing an AI information overload. But where do we find the ground truth? How should we reference our data?
Ironically, as we drown in data of uncertain veracity, the importance of reliable public datasets will only grow. Not because we lack data but because we have too much data with questionable authenticity.
What are your thoughts on establishing a foundation of truth for AI in healthcare?
[1] https://lnkd.in/eA4N4EPF [2] https://lnkd.in/e9t_wUcX #AI #Healthcare #DataTruth #AIEthics
Unveiling Doctor Insights: 10 Questions You Can Answer with Public Healthcare Data! 🏥💻
As a data enthusiast, I've been diving deep into the ocean of publicly available healthcare datasets in the mimilabs data lakehouse. It's fascinating how much we can learn about healthcare providers using just public data and some SQL magic!
Here are 10 intriguing questions you can answer about doctors using these datasets:
📊 What's the average risk score of a doctor's patient population? 🩺 Which procedures does the doctor perform most frequently? 💊 What medications does the doctor commonly prescribe? 🦾 Which medical equipment does the doctor typically order? 🏥 In which insurance provider directories is the doctor listed? 🗺️ Where are the doctor's office locations according to provider directories? 💰 What are the negotiated prices for specific services? 💻 Which Electronic Health Record (EHR) system does the doctor use? 🏆 How did the doctor perform in the Merit-based Incentive Payment System (MIPS)? 🤝 What level of financial relationships does the doctor have with pharmaceutical and medical device companies?
Important notes:
These insights come from aggregated, non-identifiable public data. Proper context is crucial when interpreting this data, especially for metrics like risk scores or performance measures.
This information aims to promote transparency, not to imply any wrongdoing.
💡 For the data geeks out there, I'd be happy to share some sample SQL scripts in a future post if there's interest!
Now, I'm curious:
What other questions do you think we could answer with public healthcare datasets?
Are there specific datasets you've found particularly insightful for understanding healthcare provider performance or patient care?
How do you think this kind of data analysis could benefit patients or healthcare administrators?
Let's discuss! Your insights could help shape the future of data-driven healthcare decision-making. #HealthcareData #DataAnalytics #PublicHealth #SQL
