Research News | Duke Kunshan University

Research News

Recent research highlights from Duke Kunshan University’s centers and programs, plus a few other places. Use the filter to narrow your search.

Stress testing with influencing factors to accelerate data race software failures

Software failures caused by data race bugs remain major concerns in parallel and distributed systems. To address the issue, iAPSE’s Kishor Trivedi and fellow researchers propose in IEEE Transactions on Reliability stress testing with influencing factors, in which the system runs under certain workloads for a long time with controlled stress conditions, and explore mathematical relationship models between data races’ statistical characteristics of time to failure (TTF) or mean TTF and the influencing factors.

Women still missing from cardiovascular trials

Cardiovascular disease is the leading cause of death among women worldwide. So why don’t more women participate in trials? Lijing Yan, an expert in non-communicable diseases, and global health master’s student Xurui Jin, explore the key factors behind the under-representation of women in cardiovascular trials in Circulation, an American Health Association journal.

First ancient DNA from West Africa illuminates deep human past

An international research team including archaeologist Scott MacEachern has produced the first whole-genome ancient human DNA sequences from West and Central Africa. Scientists recovered the data from two pairs of children buried at an iconic archaeological site in Cameroon from between 3,000 and 8,000 years ago. The findings, published by Nature, provide fresh clues in the search to identify the populations that first spoke Bantu languages, the most widespread and diverse group of languages in Africa today.

The continuity of BCI-mediated action and conventional action

In response to an article on technologies controlled via brain-computer interfaces (BCIs), Ege Duman, undergraduate Class of 22, and Daniel Lim, co-director of the Planetary Ethics and Artificial Intelligence Lab, share philosophical and scientific reasons against a qualitative distinction between conventional action and BCI-mediated action that results in a distinction in moral responsibility. Their paper was published in the American Journal of Bioethics: Neuroscience.

How VLSI CAD led the way in machine learning technologies

VLSI computer-aided design (VLSI CAD) has been at the forefront of incorporating cutting-edge algorithms in the software tools and methodologies that engineers use to weave the digital fabric of our world. “Machine Learning in VLSI Computer-Aided Design,” a book co-authored by iAPSE director Xin Li, shows VLSI CAD has also led the way in making good use of machine-learning technologies to automate the design, verification and implementation of the most advanced chips.

The role urgency plays in Nanjing property presales

Writing in the Journal of Cultural Economy about residential property presales at the urban fringe of Nanjing, China, cultural anthropologist Menqi Wang argues for an analytical focus on urgency as a temporal quality that creates the local housing market and facilitates urban accumulation. She examines how urgency, grounded in a linear imagination of urban development, motivates home presales for developers, estate agents and homebuyers.

Linking early life risk factors to frailty in old age

Unfavorable socioeconomic status and a worse health condition in childhood and adolescence may increase the risk of late-life frailty among Chinese older adults. Global health professor Chenkai Wu and fellow researchers identified an association by analyzing data from the China Health and Retirement Longitudinal Study. The findings were published in Age and Ageing.

Efficient rare failure analysis via correlated Bayesian inference

In a paper published by IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, electrical and computer engineering professor Xin Li and colleagues propose a correlated Bayesian inference (CBI) method to estimate the system-level failure rates for large-scale circuit systems over multiple process corners. Numerical experiments show that compared with the state-of-the-art algorithm, the method can achieve around 10 times runtime reduction without surrendering accuracy.

End-to-end language identification using attention-based CNN-BLSTM

In a paper accepted at the 2019 International Conference on Acoustics, Speech and Signal Processing held in Brighton, U.K., electrical and computer engineering professor Ming Li and fellow researchers present an end-to-end language identification framework, the attention-based convolutional neural network-bidirectional long-short term memory (CNN-BLSTM), which is performed on the utterance level.

Mapping child deaths to help prevent them

Worldwide progress in reducing child deaths is among the greatest success stories in global health. Yet progress has not been universal, with data showing differences across geographical regions and age groups. To help policymakers better understand where, and at what ages, child survival remains most tenuous, environmental health scientist John Ji helped map 123 million child deaths between 2000 and 2017, with the results published in Nature.