Lin Integrates VR With Plant Digital Twin in Brookhaven Lab Study

Jasmin Lin integrated virtual reality (VR) with a plant digital twin during her first Science Undergraduate Laboratory Internships (SULI) project at Brookhaven National Laboratory, a connection between plant biology and artificial intelligence. Lin’s interests expanded as she transitioned to exploring humanoid robots and embodied AI within Brookhaven’s Computing and Data Sciences Directorate, demonstrating adaptability throughout her internships. Lin explained that she created a pipeline connecting VR interaction with a 3D model of a plant, allowing users to access original images through a simple point-and-click interface. This work, along with her investigations into reinforcement learning for robot locomotion, a project that informed further exploration, contributes to the DOE’s Genesis Mission to accelerate AI innovation and discovery, showcasing the broad impact of her research.

VR Integration with Plant Digital Twins for Computational Biology

A virtual greenhouse is emerging at the intersection of artificial intelligence and plant biology, connecting AI and robotics work with plant biology. Lin’s project centered on creating a dynamic, real-time digital model of a plant, a digital twin, and linking it to a VR interface. This allowed her to visually explore the plant’s structure in VR and trace the origins of individual visual elements back to the original images used in its creation. By pointing and clicking within the VR environment, Lin could access the source imagery for each part of the 3D model, a feature designed to enhance data traceability and understanding. This project was part of her first SULI term and proved pivotal in her broader exploration of AI.

The system’s development wasn’t merely about visualization; it represented Lin’s first practical application of AI within the Brookhaven’s Artificial Intelligence group, and she found AI to be more reliable over time, with fewer irrelevant responses. Lin’s subsequent shift from computational biology to humanoid robots and embodied AI further highlights the interdisciplinary nature of her research. She expanded on this by using the SKRL reinforcement learning library to investigate humanoid robot locomotion, a fundamental component of coordinated humanoid movement, informing her further exploration into embodied AI. This transition wasn’t arbitrary; Lin found a synergy between her biological background and her newfound coding skills. “Artificial intelligence is very adaptive, and I use it to understand the research I’m doing,” she stated, emphasizing AI’s efficiency in accelerating research processes.

Lin’s experience underscores a growing trend: the use of immersive technologies to enhance scientific inquiry, allowing researchers to interact with complex data in intuitive and insightful ways, and ultimately, to accelerate the pace of discovery. The ability to visualize and manipulate plant models in VR offers a powerful tool for understanding plant physiology, genetics, and responses to environmental factors, potentially leading to advancements in agriculture and environmental science.

SKRL Library Enables Humanoid Robot Locomotion Research

Jasmin Lin’s trajectory from computational biology to the intricacies of humanoid robotics exemplifies a convergence of scientific fields to accelerate innovation in artificial intelligence. This initial project, undertaken during her first SULI term, established a foundation for her later explorations in embodied AI. Lin’s subsequent shift towards robotics wasn’t arbitrary; it was facilitated by the Scientific Embodied Agents Lab (SEAL) at Brookhaven, a dedicated testbed for embodied AI initiatives. This work, building on her initial VR experience, allowed her to explore how reinforcement learning policies function and the diverse approaches to robot training. She used the SKRL reinforcement learning library to investigate humanoid robot locomotion, a fundamental component of coordinated humanoid movement, informing her further exploration into embodied AI. The project’s success culminated in a presentation at the New York Scientific Data Summit, a significant milestone for the young researcher.

The implications of Lin’s work extend beyond academic curiosity, directly contributing to the Department of Energy’s Genesis Mission, an initiative designed to accelerate AI innovation and discovery. She is applying her skills to practical challenges within Brookhaven’s user facilities, specifically the National Synchrotron Light Source II (NSLS-II). The goal is to develop robotic solutions that can perform maintenance tasks in hazardous environments, preventing the need to shut down the beam and improving operational efficiency. Lin is connecting vision-language-action policies, AI models that enable robots to perceive, understand, and act, with a physical robot, training it to autonomously manipulate objects. “We’re hoping to branch into more complex tests,” she stated, indicating the ambitious scope of the project. Lin’s experience underscores the power of interdisciplinary research and the potential of AI to address real-world problems, even those seemingly unrelated to its origins.

Vision-Language-Action Policies Reduce NSLS-II Downtime

Lin’s work, initially sparked by a Science Undergraduate Laboratory Internship (SULI), now focuses on leveraging robots equipped with vision-language-action policies to perform maintenance tasks within the NSLS-II accelerator tunnel, potentially preventing the need to shut down the X-ray beam during repairs. This approach builds upon her earlier exploration of virtual reality and digital twins during her first SULI term. She explains that her initial foray into AI involved a dynamic model mirroring a physical plant’s structure. This experience, though seemingly distant from robotics, proved crucial in developing her understanding of AI pipelines. The ultimate goal is to deploy these robots for real-world tasks within the NSLS-II facility. Currently, any maintenance or repair work requiring human access to the accelerator tunnel necessitates a complete shutdown of the X-ray beam. Lin’s research aims to circumvent this limitation, allowing robots to perform these tasks without interrupting valuable research time.

AI Accelerates Research & Data Analysis in Bioinformatics

Jasmin Lin’s journey from biology to embodied artificial intelligence exemplifies how AI is rapidly reshaping research methodologies, particularly in fields demanding complex data analysis and real-time modeling. She found earlier AI models to have irrelevant responses, an observation as she began exploring the technology, but her user experience has steadily improved, and she found AI to be increasingly reliable. Lin’s transition to humanoid robots and embodied AI during her first SULI term, and continued through her Supplemental Undergraduate Research Program (SURP), showcased a willingness to embrace interdisciplinary approaches. Lin’s experience underscores a broader trend: AI is no longer simply a tool for data crunching, but an integral component of the scientific process itself, accelerating the pace of discovery and enabling solutions to complex challenges.