NYC Gazette

A team of researchers led by Dr. Honggang Wang, chair of the graduate Department of Computer Science and Engineering at the Katz School, presented significant advancements in data science at the 2024 Joint Statistical Meetings conference held in Portland, Oregon. The focus was on deep learning and federated learning applications in medical research.

Dr. Wang emphasized the importance of integrating advanced mathematical tools like the Choquet integral into data science for healthcare improvements. "Our presentations underscored the importance of integrating advanced mathematical tools like the Choquet integral into data science, particularly in the healthcare sector," said Dr. Wang. He added that their research could transform patient care approaches in an era dominated by big data and decentralized research.

The conference centered around "Statistics and Data Science: Informing Policy and Countering Misinformation." It gathered statisticians and data scientists to discuss developments in statistical learning and data science.

In a presentation titled “A New Choquet Activation Function Based Deep Neural Network for Drug Interaction Detection,” Matthew Fried, a Ph.D. candidate at Katz School, along with Dr. Wang, introduced a method to improve neural networks' detection of drug interactions using a new activation function based on the Choquet integral.

Dr. Wang collaborated with Semyon Lomasov from Stanford University to develop an algorithm using the Choquet integral for handling complex medical data without prior statistical assumptions. Their approach accurately modeled drug interactions' synergistic and antagonistic effects, outperforming conventional techniques.

Another session featured Lomasov and Dr. Wang presenting “Federated Choquet Regression with Categorical Variables for Outcome Prediction in Longitudinal Trial Data.” This addressed privacy concerns when aggregating clinical site data by developing a federated regression algorithm based on the Choquet integral.

Their algorithm showed strong performance but highlighted dependency on aggregation methods for effectiveness in decentralized scenarios. This marks a pioneering application of Choquet-based regression within federated learning contexts.

"As the need for more sophisticated data analysis methods continues to grow, our work sets the stage for future research that could lead to more accurate predictions and better health outcomes," stated Dr. Wang.