Managing bias in data sets
Moderator: Daniel Engels, professor of practice, SMU dept. of computer science and engineering
Panelists: Pradyot Prasoon, director of emerging technologies & technology strategies, American Heart Assoc.
Hettie Tabor, director of masters of science in business analytics, SMU
Dwayne Paschell, data scientist - machine learning and AI, IDEXX
Kalyana Chakravarthy Bedhu, head of AI and data innovation, Ericsson Group Analytics
Brooks Fitzsimmons, AVP of data insights, AT&T
The significant challenge of bias in data sets and particularly data sets used to train AI is no secret. We know that AI, machine learning and deep learning can produce not only inaccurate but potentially disastrous results if training sets are left unchecked. Bias isn't just a bad data set but can also be the algorithm itself. Optimizing the AI algorithm and constructing data sets to avoid bias go hand in hand. AI holds significant power to improve the way we live and work, but only if AI systems are developed and trained responsibly, and produce outcomes we can trust. In this session we will discuss:
Pradyot Prasoon is the director of emerging technologies and technology strategy at the American Heart Association. He leads the association's emerging business and health strategies to accelerate health impact at the intersection of science, data, analytics, and technology. He works on developing strategic partnerships and relationships both internally and externally with a panel of volunteers, partners, and cross-functional collaborators in helping create a marketplce of healthcare solutions for advancing AHA's mission.
Hettie Tabor is the director of the master of science in business analytics program at SMU Cox. She is a seasoned senior executive with over 26 years of experience with Accenture in information technology including 21 years of practical SAP implementation experience and 20 years of analytics experience where she ran Accenture's SAP Business Analytics Global Practice prior to joining SMU five years ago.
Dwayne Paschell is a data scientist - machine learning and AI at IDEXX. His background is in time series signal enhancement and perceptual processing of sensory data by the human brain, i.e., biological neural networks. This data is similar to other times series domains such as financial times series and custom and patient data over time. He uses this data to make decisions, optimize processes, and improve products and services. He engineers features for both human and machine learners and teaches them both to become experts in data interpretation. This helps people to make important data-driven decisions, prove their intuitions, predict the future, and optimize outcomes and processes through insightful application of data science. His focus is in the healthcare, financial services, energy, and marketing industries.
Kalyana Chakravarthy Bedhu is head of AI and data innovation at Ericsson Group Analytics. He is a leader in the expansion of data sciences at Ericsson, establishing a data science lab and building competence in solving practical data science problems. His team serves Ericsson's enterprise business intellignce by defining, building, prototyping and industrializing cross-department data science use cases leading to meaningful business solutions. He is an award-winning author, the recipient of multiple patents, a frequent speaker, and the developer of data science courses. His team won the CiMi.CON award for market intelligence products and Top Performance - Innovation of the Year Award at Ericsson in 2018.