We are excited to announce at
a new strategic partnership with Cleveland
Clinic's Hwang Lab. xLab and Hwang Lab will collaborate to work on
training workshops, seminars, and client projects combining design thinking and
digital business competencies of xLab with data analytics and AI competencies
of Hwang Lab.
Our next xLab Quarterly Rountable Meeting for members will feature a
presentation by Dr. Hwang, followed by an in-depth discussion on how companies
can build their AI strategy and roadmap. In preparation of our quarterly
meeting, we sat down with Dr. Tae-Hyun Hwang for a short interview.
Question: Tell us a bit about your lab at Cleveland Clinic
Hwang: Our lab is translational machine learning AI lab. Our
members are from backgrounds in computer science, electrical engineering,
financial engineering, medical degree from 18 years old and up.
We utilize large, complex, sparse (and noisy) data from
image to genomic, electronic health records, wearable devices to anything that
we can feed to the algorithms.
We are focused on developing algorithms that can be readily
useful in the clinical setting to ultimately help patients with lethal disease.
Question: In your mind, what is AI? In today's world, if it is an
exciting application of computers and data, people tend to call it AI. But I
worry that if everything is AI, then nothing is AI. What are some of the
defining characteristics of AI that are different from other computing
technologies?
Hwang: This is a tough and complicated question. At least, from my
point of view, AI is an algorithm that can learn patterns from the data. The
patterns related to recognizing your face from a selfie you took, predicting
the weather forecast, or what would be your traffic for your commute.
Question: Where is the field
of AI going? What in your mind are the most exciting technological developments
in AI?
Hwang: I think everyone would admit that there is still a hype about
AI, and no longer any doubt about
whether AI is driving and impacting many industries or even lives.
The field of AI, and many people, are starting to think
about building more rigorous, accurate and ethical AI models (no longer simply
claiming that AI can solve any problems) and how to educate people to
responsibly use AI.. In particular, people are more cautious and careful to
claim the capability and feasibility of the AI system they are developing.
The most exciting technologies would be the AI in
healthcare, of course. Sorry I am biased to the healthcare industry since I
work in this area, but isn't it amazing that the AI system could help care for your
loved ones?
Question: What are the 2-3 most significant applications of AI in
business today?
Hwang: There are too many applications. Internet of Things (IoT)
that you use every day, your phone, speaker, voice command, etc. Anomaly
detection in finance, such as fraud detection, identity theft, and fake news in
financial media outlets. The personalized
advertising industry based on an average consumer spending a significant amount
of time with all devices. Also, using virtual reality in the fashion and
leisure industry.. Lastly, the healthcare industry with better patient care,
wellbeing management with a wearable device and drug development!
Question: What are the 2-3 mistakes or misconceptions that managers
have about AI?
Hwang: The biggest misconception is we can build the AI system with
our data; I mean any data we have. The key to developing rigorous and reliable
AI systems are mainly relying on the quality of the data, not the quantity.
Zillions of data points from your data are not going to help to build better
the AI system. Likewise, even relatively small datasets could help to build the
AI system.
Once you have the AI system, you consistently feed new data
to make it better and more rigorous. That means you need feedback from your
employees on evaluating which areas the AI systems are doing well and
performing poorly. To do that, you need to educate your employees on how to use
the AI system, how to build a better AI system, etc.
Bio
Tae Hyun Hwang received his PhD in Computer Science (Machine
Learning, Data Mining and Computational Biology) at the University of Minnesota
Twin-Cites at 2011. He and his research group lead machine learning and AI
research at Cleveland Clinic. Prior his appointment at the Cleveland Clinic, he
was a Research Associate in the Department of Computational Biology and
Bioinformatics at Genentech and was a tenure-track faculty at the University of
Texas Southwestern Medical Center where he led a team of computational
scientists for cancer research. He currently serves as a bioinformatics core
director for NASA Specialized Centers of
Research (NSCOR) as well as committees of various Machine Learning, Data
Mining, and AI conferences.
His lab is working on developing novel machine learning and
AI algorithms that are readily applicable in the clinical setting to help
patients with a lethal disease