You are here

Bullard Spotlight: Comparative investigation of heterogeneity across the natural sciences, social sciences, and humanities

January 3, 2024
Printer-friendly version
Image shows Bullard Fellow Zhanshan Ma with Harvard Forest Senior Research Fellow in Ecology Emeritus Aaron Ellison

By Zhanshan (Sam) Ma

During his Bullard Fellowship, Zhanshan (Sam) Ma - CAS 100-Talents endowed professor and PI of Computational Biology and Medical Ecology Lab, Chinese Academy of Sciences (CAS) - is working with Harvard Forest Senior Research Fellow in Ecology Emeritus Aaron Ellison to review and synthesize the literature on heterogeneity published across the sciences, social sciences, and humanities.

Heterogeneity is often conflated with diversity, and their differences are not always explicated clearly. Shavit and Ellison (2021) argue that heterogeneity is an essential concept for noticing differences well beyond the ecological sciences, and “is fundamental for building our most basic categories, social systems, models, and their causal explanations.” 

Since the start of Sam's Fellowship in July of 2023, he and Aaron have identified more than 1,200 papers on “heterogeneity” spanning more than two dozen disciplines—including biochemistry, biomedicine and clinical medicine, computational science and engineering, ecology and evolution, economics and management science, geology and planetary science, natural resources (including forests) and agriculture, physics and material science, psychology and psychiatry, sociology and political science, statistics and mathematics—each with more than 400 citations, on average.  

Sam’s initial findings include:

  • Many tipping points (both “black swan” and “grey rhino” events) such as biodiversity loss and climate changes are associated with ecological heterogeneity.

  • Many difficulties in early diagnosis of neurodegenerative diseases including Alzheimer’s, Parkinson’s, and autism spectrum disorder relate to multidimensional heterogeneities in genomics, epigenetics, brain imaging, and phenotypic behaviors. Failures in treating cancers can be traced to their inherent heterogeneity.

  • Geological heterogeneity and its evolution can have far reaching influences on the fault strength and stability, which in turn affect the occurrence and scale of earthquakes.

  • Heterogeneity can be a better metric than diversity in guiding policy design for achieving social justice. This has been demonstrated by James Heckman, who in 2000, shared the Nobel Prize in Economics “for his work on the microeconometrics of diversity and heterogeneity and for establishing a sound causal basis for public policy evaluation" (Center for the Economics of Hman Development, University of Chicago).

  • Heterogeneity influences performance and innovation for business and team management, international trade, financial and banking systems, and overall economic growth and sustainable development.

  • The apparent success of deep-learning-based AI is rooted in its capacity to harness heterogeneous information flowing through heterogeneous neural networks. Pushing heterogeneity into cores is likely to be an avenue to design powerful next-generation AI chips because energy consumption of multiple cores should be more efficient in handling heterogenous computational tasks.

  • Statisticians believe that there is not a single approach to dealing with heterogeneities across different studies on the same problem.

  • Significant gaps in dealing with heterogenies exist among different disciplines and strategic-level thinking and design, and cross-disciplinary approaches will be crucial for dealing with the heterogeneities inherent in some of the pressing challenges humans face today and in the foreseeable future.

There are two goals for Sam’s Bullard research, which will extend through September of 2024. First, to write a review paper on heterogeneity focusing on the classification and synthesis of existing approaches for measuring heterogeneity across more than a dozen disciplines. Second, to develop novel methods for assessing and interpreting heterogeneities using network science, graph theory, and deep learning. The overall aim is to develop a new and more general cross-disciplinary framework for measuring heterogeneity and predicting its effects on complex systems.  

Above image shows Bullard Fellow Zhanshan (Sam) Ma with Harvard Forest Senior Research Fellow in Ecology Emeritus Aaron Ellison.

Learn about Sam Ma & Aaron Ellison's presentation at MIT in January 2024!

Content Tags: