This column highlights the experiences and perspectives of leaders and volunteers within the IEEE Photonics Society, offering members a closer look at the individuals shaping the community. Through Q&A-style interviews, readers gain insight into their career journeys, contributions to the field, and thoughts on the future of photonics, the broader industry, and professional development.
In this edition, we feature Shaobo Han, Ph.D., a Senior Researcher in the Optical Networking and Sensing Department at NEC Laboratories America in Princeton, New Jersey. He earned his Ph.D. in Electrical and Computer Engineering and an M.S. in Statistical Science from Duke University, where his work focused on probabilistic modeling, transfer learning, and structured variational inference. He also holds an M.Eng. in Signal and Information Processing from the University of Chinese Academy of Sciences.
Han additionally serves as a photonics liaison and Associate Editor for IEEE Transactions on Big Data, a peer-reviewed IEEE journal dedicated to advancing research and real-world applications in big data. The journal spans areas such as data analytics, machine learning, large-scale data systems, and data security, bringing together contributions from both academic and industry communities.
What is your current position/area of expertise/research study?
I am a Senior Researcher in the Optical Networking and Sensing Department at NEC Laboratories America in Princeton, New Jersey. My work sits at the intersection of machine learning, signal processing, and physical sensing. Specifically, I develop AI methods that turn fiber-optic communication cables into large-scale sensing systems capable of monitoring infrastructure and the surrounding environment in real time.
By applying machine learning to analyze tiny changes in the light scattered within optical fibers, we can automatically detect and classify events such as vehicle traffic, construction activities, cable damage threats, gunshots, and even undersea earthquakes. Within our team, this line of work has led to multiple world-first and industry-first technology field trials, as well as multiple NEC commercial products.
I am deeply engaged in the photonics and optical communications community, publishing in journals and conferences including the Journal of Lightwave Technology, OFC, Optics Express, and the Proceedings of the IEEE. In parallel, my machine learning research focuses on making large-scale AI models more efficient and adaptable, including parameter-efficient fine-tuning (PEFT) for large language models and knowledge distillation for specializing multimodal foundation models. This work has been published at major machine learning and signal processing venues, including NeurIPS, ICLR, CVPR, and ICASSP.
What inspired you to accept the position of photonics liaison and Associate Editor to TBD?
I was drawn to this role because it reflects a convergence that I experience firsthand in my daily research: the growing interdependence between photonics systems and large-scale data analytics. Data serves as the interface between physical sensing and digital intelligence—and ultimately the gateway to real-world impact. Photonic systems are rapidly becoming large-scale data generators, creating opportunities for data science research in areas such as anomaly detection in optical networks, multimodal sensor fusion, spatiotemporal modeling, physics-informed machine learning, and real-time streaming data processing and management.
Many photonics researchers are developing innovative sensing, communication, and computing systems that produce or process enormous volumes of data, such as detecting anomalies in high-dimensional signals, extracting patterns from noisy physical measurements, and performing inference under uncertainty. Yet, they may not always be aware of IEEE Transactions on Big Data (TBD) as a potential venue for their work. Conversely, the data science community could benefit greatly from the novel problem formulations, new data modalities, and real-world challenges that photonics researchers bring. Photonics technologies are generating a ‘deluge’ of data that often has unique properties (extremely large, spatially distributed, temporally continuous, physically structured), which provide fertile ground for new research in data science, machine learning, statistical inference, and scalable computing. I wanted to help strengthen this connection.
What specific assets do you bring to the table as an editorial board member?
I bring a somewhat uncommon combination of expertise in both core machine learning research and real-world physical sensing applications. My doctoral training at Duke University focused on probabilistic machine learning and Bayesian statistics—the mathematical foundations that underpin many modern data-driven methods. At the same time, my career at NEC Laboratories America has been devoted to algorithmic innovations in distributed fiber-optic sensing, optical networking, and optical computing, while also drawing on challenges from real-world sensing systems to inspire new machine learning methods. This work has given me extensive hands-on experience with the practical challenges of working with massive, noisy, and heterogeneous data in high-stakes real-world applications.
This dual perspective helps me appreciate work that spans the full pipeline—from novel data acquisition and curation to scalable machine learning models, data management platforms, and end-to-end system design. It has also given me some perspective on both the algorithmic innovation valued by the data science community and the practical relevance important to photonics and sensing practitioners.
How do you hope to contribute to the relevance of TBD in the rapidly advancing field of photonics?
IEEE Transactions on Big Data (TBD) complements venues such as the Journal of Lightwave Technology. While IEEE Photonics Society journals focus primarily on the physical layer and hardware performance, TBD places greater emphasis on data analytics and algorithms, helping bridge photonics research with real-world applications. It provides a natural venue for photonics work that involves large-scale data challenges—an area that will likely grow significantly in the coming years.
I hope to encourage more submissions from photonics researchers working on data-intensive problems, such as distributed fiber sensing for infrastructure monitoring, machine learning for optical network optimization, photonic computing, AI-driven spectroscopy, and LiDAR-based perception systems. These areas generate massive datasets that demand innovative big data solutions, yet the work is often published only in optics-specific venues. TBD can serve as a complementary home that highlights the data science dimension of this research and exposes photonics challenges to the broader data science and computer science communities, fostering new collaborations.
In many traditional optics venues, AI is often used as an off-the-shelf tool alongside established signal-processing techniques. While these approaches are valuable, emerging photonic systems often pose data challenges that conventional methods cannot readily address. In IEEE Transactions on Big Data, I hope to encourage work that treats these domain-specific challenges as opportunities to advance the algorithms themselves. For example, new AI methods are needed to bridge the gap between controlled laboratory prototypes and large-scale deployments, where models must adapt to real-world complexity that is difficult to express through physical equations alone. By rethinking these problems more fundamentally, researchers can move beyond simply applying existing tools and redefine what is possible at the intersection of photonics and data science.
Why Photonics? What Was Your “Photonics Moment?” (Personal background story, etc.)
My path into photonics was an unexpected journey. My background is in machine learning, so I originally worked mostly in the digital world of algorithms and data. What drew me toward photonics was the idea of sensing—the possibility of using physical signals to understand the world around us.
I became fascinated by the idea that you could shine a laser pulse down a standard telecom fiber and, by analyzing the backscattered light, effectively turn every meter of that fiber into a microphone. The dot-com era left the world with an enormous amount of “dark fiber”. Much of that capacity was originally built for communication, but today the same fibers can be “lit up” in new ways to help us observe and understand the physical environment. In a sense, the fiber becomes a nervous system, transmitting signals about what is happening along the infrastructure at the speed of light, while AI acts as the brain that interprets those signals and turns them into insight for real-time situational awareness.
That realization was my “photonics moment.” Seeing how advances in photonics and AI could translate into tangible sensing capabilities through the physics of light in fiber was a powerful experience, and I have been deeply committed to this area ever since.
Would you like to add anything else, e.g., about other professional or personal interests?
More broadly, I’m excited about the growing convergence of photonics, AI, and data science, where advances in light-based technologies and modern data analytics are beginning to reinforce and accelerate each other. Photonics technologies are generating increasingly large and complex datasets—from optical networks and spectroscopy to imaging and sensing systems—while advances in machine learning and large-scale data analytics are opening new ways to analyze complex data and optimize these systems. At the same time, photonic systems themselves are becoming critical infrastructure for the AI era. Optical interconnects power modern AI data centers by enabling the high-bandwidth, energy-efficient communication needed to move massive amounts of data between GPUs and computing clusters.
I believe there is a tremendous opportunity in bringing these communities closer together. When these perspectives meet, they can unlock new capabilities and address challenges that neither field could solve alone. I also enjoy mentoring students and collaborating across disciplines, because many of the most exciting discoveries emerge at the boundaries between research communities.
