In brief

Newly-appointed IEEE Fellows Xiaoli Li (right) and Kui Yao reflect on their scientific achievements in electronical and electronic engineering, and how teamwork has enabled their successes.

© A*STAR Research

Electrifying advances in smart devices

21 Mar 2024

From artificial intelligence to smart materials and sensors, A*STAR scientists Xiaoli Li and Kui Yao have spearheaded scientific breakthroughs that have led to their recognition as Fellows of the Institute of Electrical and Electronics Engineers (IEEE).

Since Benjamin Franklin’s kite, advances in our understanding of electricity—and the ways we can harness it—have brought the stuff of science fiction into modern-day life. Sensors made from piezoelectric materials can generate electricity from movement, ultraviolet (UV) radiation and heat; in turn, they give rise to ultrasound probes, Internet of Things (IoT) devices, and UV and thermal detectors. Self-powered sensors are also possible thanks to ferroelectrics, a subgroup of piezoelectrics that often have stronger properties and, in principle, need no external power supply.

The integration of artificial intelligence (AI) to analyse sensor data collected over time is also helping to create the next generation of smart devices and useful tools. AI unravels patterns and enhances decision-making for machines that optimise manufacturing processes, monitor equipment health and improve healthcare diagnostics.

A*STAR Research speaks to Xiaoli Li from the Institute for Infocomm Research (I2R) and Kui Yao from the Institute of Materials Research and Engineering (IMRE), who were both recently conferred Fellows of the Institute of Electrical and Electronics Engineers (IEEE) for their trailblazing work in their respective fields. The researchers discuss their research journeys, their proudest scientific achievements, and the collaborations made all of it possible.

What sparked your interest in your field?

XL: I was initially drawn to AI due to its transformative potential in solving complex problems. In 1976, AI successfully proved the Four-Colour Theorem; by the 90s, AI was winning games like checkers and chess. These breakthroughs have showcased AI’s computational prowess and ability to intelligently tackle complex problems.

I developed a deep-seated curiosity about AI, especially how it gains knowledge or creates models from data through machine learning (ML). Witnessing AI’s impact on various industries fuelled my passion.

KY: My research focuses on ferroelectrics and piezoelectrics—from materials to device applications. Since my PhD studies, I’ve found ferroelectrics attractive due to their extraordinary material properties and the rich physics underlying them.

Though there are billion-dollar markets for their commercial applications, ferroelectrics have greater unexplored potential in advanced sensors, transducers, memories and intelligent systems. There’s ample room for improvement even in our fundamental understandings of ferroelectric-related phenomena.

Tell us about the scientific contributions you’re most proud of.

XL: They include the development of advanced ML models for processing time-series sensor data, which has potential applications across diverse domains from manufacturing to healthcare.

In 2011, I led an I2R team that worked with Boeing to build a data-driven model for portable machine fault detection. As existing signal processing approaches to learn time and frequent domain features had limited accuracy, we began to explore AI, focusing on timeseries sensor data analysis. We pioneered a novel deep learning (DL) method that could automatically extract representative features from high-dimensional sensor signals, which generated much better outcomes than previous approaches. The resulting paper from this work has received over 1,000 citations.

I hope these models will see wider real-world applications, ranging from equipment health diagnosis to human activity recognition.

KY: I’m pleased with what we’ve done to advance ferroelectric film materials, including coatings for innovative piezoelectric sensors; and our exploratory work on bulk photovoltaics in ferroelectric films, which has led to robust ferroelectric UV sensors.

After I joined IMRE, our work on such films started with research projects jointly supported by the institute and industry. My initial efforts involved basic, yet vital lab-based engineering work to determine, reliably and efficiently, a film’s piezoelectric and bulk photovoltaic properties. Eventually, I became the main inventor for several piezoelectric and ferroelectric sensors and UV dosimeters produced in collaboration with my colleagues and our industry licensees.

I hope to see wider applications for the ferroelectric materials and devices we’ve developed, particularly as advanced electromechanical sensors and ultrasonic transducers.

What partnerships were key to your scientific successes?

XL: I’m glad to have had opportunities to work with world-class scientists through international conference networks. I’ve also had the privilege of leading key initiatives in partnership with aerospace, semiconductor and manufacturing companies, which not only provided problem statements driven by real-world use cases, but helped translate theoretical advancements into practical, domain-specific solutions.

KY: I appreciate the guidance, leadership and unwavering support of my mentors, international peers, as well as A*STAR and IMRE’s management and colleagues. The contributions of my team members, students and collaborators have been invaluable to our shared achievements.

What are some interesting projects you’re working on?

XL: I currently spearhead my AI Singapore project, “Self-Aware Continuous Learning Models” (SACoLM), which aims to create AI models capable of lifelong learning: a crucial skill for dynamic decision-making. Unlike traditional DL models that assume known distributions, SACoLM addresses the challenge of adapting to unknowns. Our models maintain awareness of representations; identify novel distributions and estimate their complexity; and adapt while retaining previous learning.

Our approach has highly relevant engineering applications, such as defect detection, fault diagnosis and engineering asset prognosis, as it uses IoT sensor data to manage systems with drifting characteristics.

I also work on sustainable AI, which aims to reduce the carbon emissions and huge computing power consumption associated with developing and deploying large-scale AI models.

KY: My team is currently developing ferroelectric materials with superior piezoelectric properties, and producing advanced electromechanical sensors and transducers for ultrasonic structural health monitoring and healthcare. Another goal is to create eco-friendly, lead-free alternatives to lead-based piezoelectric materials that are toxic but are currently widely used. To boost our impact, we’re striving to enhance our collaborations with A*STAR’s Institute of Microelectronics (IME) and industry partners.

What does your elevation to IEEE Fellow mean to you?

XL: It’s a significant honour that recognises our team’s collective efforts in advancing ML and AI, and bolsters my dedication to advancing knowledge in these domains. The IEEE Fellowship offers a platform to engage with a global community of experts and cultivate an environment conducive to innovative thought and idea exchange. Regardless of the title, I remain committed to working with our A*STAR colleagues to fulfil our mission of translating research into real-world impacts.

KY: I appreciate IEEE’s recognition of my technical contributions to ferroelectrics and their sensor applications; I’d like to share it with my A*STAR colleagues, students and collaborators in this area, who have made it possible together. This recognition increases the visibility of our research and strengthens my resolve to continue our efforts therewith.

What do you think makes good science?

XL: A delicate balance between rigorous methodology and a creative, open-minded approach; a commitment to pushing the boundaries of knowledge, coupled with a genuine curiosity to explore the unknown; and transparency, reproducibility, and a dedication to ethical practices. It’s not only about answering existing questions, but asking the right questions, challenging assumptions, and inspiring others to join the pursuit of knowledge.

KY: Curiosity, passion, collaboration, perseverance and courage to overcome various difficulties on the way.

What advice would you give other scientists looking to create impact?

XL: Delve into research topics grounded in real-world problems with tangible applications. It’s also crucial to collaborate with expert stakeholders to bridge gaps between theory and practice.

In the ever-evolving AI landscape, it’s imperative to continually learn and adapt. While striving to discover, we should also commit to making tools and insights accessible to the broader community. Ultimately, our work’s true measure is its ability to address real-world challenges and improve society.

KY: I’m still learning how to work more effectively and impactfully. Align your interests with your job; work hard with vision and strategy; appreciate and respect those who have supported you; and collaborate with the right people to achieve your goals.

Want to stay up to date with breakthroughs from A*STAR? Follow us on Twitter and LinkedIn!

This article was made for A*STAR Research by Wildtype Media Group