Industry executives and experts share their predictions for 2023. Read them in this 15th annual VMblog.com series exclusive.
Synthetic data will accelerate AI innovation
Dr. Rana el Kaliouby,
Deputy CEO, Smart Eye, and former CEO and co-founder, Affectiva
Despite the prevalence of AI solutions in-market today, a
significant challenge persists. The successful development of AI relies on
getting the right amount and diversity of data to train machine learning-based
algorithms. Acquiring such data is time-consuming and
expensive, but the risk of developing AI based on insufficient or
unrepresentative data is the potential for biased systems that do not function
the way they are intended to for all users that will interact with the
technology.
However, in 2023, we expect that synthetic data will be a
game-changer in solving this challenge, accelerating the development and
deployment of ethical AI.
Synthetic Data as a
Differentiator
As mentioned, one of the significant challenges in
developing AI is ensuring sufficient quantities of diverse, high-quality
training data. These algorithms require massive amounts of data representative
of the different people interacting with the technology and the contexts in
which it will be used. It is difficult, time-consuming, and costly to acquire
this breadth and depth of data.
Data synthesis enables AI companies to rapidly augment
their existing datasets and simulate scenarios that are difficult to generate
in the real world. This means a company can generate a dataset based on
whichever conditions or qualities they require-such as appearance features or
environmental conditions -without exhausting extensive resources in the
process. And even more importantly, data can be diversified and scalable from
the start. This capability can set AI companies apart, providing them with a
competitive edge to develop and deploy more accurate and effective systems.
Synthetic Data in Action
There is a multitude of use cases where synthetic data can
make a tangible impact. For example, in automotive, synthetic data tools can
use a source image of a driver to create synthetic variations that use varying
lighting conditions or head movements. It can even simulate a driver falling
asleep behind the wheel - data that is rare and very dangerous to capture in
real life. From there, datasets can validate intelligent safety technology.
Once a driver's cognitive or emotional state is analyzed, the car can trigger
and alert an emergency feature and prevent a potential accident. Driver
monitoring is a complicated AI use case and requires a diverse pool of data to
draw from. If safety software is unable to identify dangerous driver behavior
because it does not recognize people or environmental conditions accurately,
that bias can lead to critical safety concerns down the road.
Looking ahead
It can be difficult to know what kind of data you need to
capture from the beginning of a project. AI models are constantly evolving, and
the technology we use to build them out needs to be equally adaptable.
Deploying synthetic data tools will be key to not only solving these complex
challenges of data collection moving forward but also to combat algorithmic
bias by ensuring datasets are truly diverse.
##
ABOUT THE AUTHOR
Rana
el Kaliouby is an Egyptian-American scientist, entrepreneur, investor and an AI
thought leader on a mission to humanize technology before it dehumanizes us.
She is the Deputy CEO at Smart Eye and formerly, Co-Founder and CEO of
Affectiva, an MIT spin-out. Rana is also General Partner at AI Operators Fund,
an early stage AI-focused venture fund, as well as an executive fellow at the
Harvard Business School where she teaches about AI and startups. Her
bestselling memoir, Girl Decoded: A Scientist's Quest to Reclaim Our Humanity
by Bringing Emotional Intelligence to Technology follows her journey, growing
up in the Middle East and moving to the United States to become an entrepreneur
and Emotion AI pioneer. Rana has a track record of translating technology
innovations into products that address massive market needs including in
automotive, health and robotics. Rana is a Trustee at the Boston Museum of
Science, as well as the American University in Cairo. She is a member of the
Young Presidents' Organization (YPO) where she serves on YPO's New England
board and is a World Economic Forum Young Global Leader. She is co-chair of the
Fortune Brainstorm AI conferences. A TED speaker and co-host of a PBS NOVA
series on AI, Rana has been recognized on Entrepreneur's 100 Women of
Influence, Fortune's 40 Under 40 list, Forbes' Top 50 Women in Tech and
Newsweek's top Disruptors. She holds a Ph.D. from the University of Cambridge
and a Post Doctorate from MIT.