Dr. Eulanca Y. Liu’s current research goals in the Savjani laboratory focus on the integration of advanced imaging and artificial intelligence to enhance the management of brain metastases and radiation necrosis (RN). Her first aim is to develop a comprehensive map of brain metastases by leveraging an expansive dataset of clinical and imaging records from over 3,000 patients and 10,000 radiographic images. A convolutional neural network (CNN) will be adapted to normalize thousands of MRI images for population-level analysis, improving lesion localization and enabling personalized treatment recommendations. Her second aim involves automating the detection of radiation necrosis using generative AI and large language models to address the challenges of manual diagnosis. This project will additionally evaluate RN response to treatments and predict its occurrence based on individual patient profiles.

About Dr. Liu

Dr. Liu is a Radiation Oncology resident physician and ABR Holman Pathway postdoctoral scholar at UCLA, working in the lab of Dr. Ricky Savjani. She completed her BS in Chemical Biology at UC Berkeley, with a minor in Global Poverty and Practice, and obtained her MD and PhD in Neurosciences from UC San Diego as part of the NIH-funded Medical Scientist Training Program (MSTP). Her PhD work focused on testing noninvasive functional MRI techniques to evaluate neural activity through quantification of cerebral blood flow and oxygen metabolism. 

Dr. Barbara Lopez-Doriga is currently working on finding the best airfoil design that most-effectively mitigates the effect of gust interactions using machine learning. In simple terms, a plane experiencing “turbulence” represents an instance of gust encounters. Traditionally, the configuration of the airfoil (including the shape, angle of attack or size) is found through an iterative process in which an experienced aerodynamicist manually adjusts all the parameters involved until a threshold is met (usually in terms of maximal lift and minimal drag). This is generally time-consuming, and even after achieving a candidate configuration, it is not possible to infer meaningful physical relationships between the aerodynamic performance and the geometry of the airfoil. Barbara’s goal is to achieve the optimal design, considering the effect of gusty conditions, using an autoencoder. This is achieved in two steps: first, a collection of high-fidelity simulations that includes several angles of attack, airfoil geometries and gusts of different strength, is generated; second, a reduced-order model of of the turbulent dynamics is achieved in a reduced-order nonlinear space. In this manner, the properties of the airfoil can be directly expressed in terms of the aerodynamic performance. This will allow her to not only find the optimal airfoil geometry, but also the most effective control strategy to mitigate the effect of the impinging gust.

About Dr. Lopez-Doriga

Dr. Lopez-Doriga recently started her position as a postdoctoral scholar in Professor Kunihiko (Sam) Taira’s lab in the department of Mechanical and Aerospace Engineering at UCLA. Barbara was born in Madrid, Spain, where she received a MS and a BS in Mechanical Engineering from the Polytechnic University of Madrid (UPM). She then received a ME in Mechanical and Aerospace Engineering from Illinois Tech (Chicago, IL), and a PhD in Mechanical and Aerospace Engineering from Illinois Tech as well. Since joining Taira’s lab, her interest lies in the development of a data-driven machine-learning-based framework to find the optimal airfoil design and control systems for gust mitigation.

Dr. Claire Schollaert is an environmental health scientist who specializes in exposure assessment, leveraging multidisciplinary modeling approaches to better understand how air quality and climate-related hazards impact the health of vulnerable communities. Claire’s research is currently focused on examining the impacts of future climate change and proposed forest management scenarios on air quality and public health across the western United States. A member of the Western Fire and Forest Resilience Collaborative, Claire is part of an interdisciplinary team that aims to bridge forest and fire ecology models, climate projections, air pollution models, and exposure and health analysis to better incorporate public health considerations into future forest management decision making.

About Dr. Schollaert

Dr. Schollaert is a postdoctoral scholar in Dr. Miriam Marlier’s lab in the Environmental Health Sciences department in the UCLA Fielding School of Public Health. Claire completed her PhD in the Department of Environmental and Occupational Health Sciences at the University of Washington, where her research focused on examining the smoke exposure tradeoffs of wildfire and forest management strategies among vulnerable communities. Claire received an MPH from the Boston University School of Public Health, and a BA from Bowdoin College in Biology and Anthropology. When Claire isn’t thinking about wildfire smoke, you can find her hiking, camping, and skiing with her dog Louie.

Dr. Hesam Soleimani’s research centers on data-driven urban infrastructure monitoring to promote resiliency and sustainability in urban settings. This topic can be narrowed down to individual infrastructure, such as bridges and buildings health monitoring, to city- and region-scaled risk analysis (simulation) and post-hazard assessment through remote sensing data. In dealing with urban settings (urban textures) and infrastructure topologies, generalizability is the main hurdle in data-driven decision-making. Accordingly, improving the generalizability of such designs is the main essence of Dr. Soleimani’s research. Given the huge leap forward in self-supervised learning and attention mechanisms, every day, another “All” or “Everything” model emerges, such as Segment anything, Depth anything, and Large Language Models. These models, on a general scale, offer excessive generalizability, and how to integrate them into the aforementioned downstream tasks is Dr. Soleimani’s current interest, in addition to devising novel machine learning strategies to further improve generalizability, for instance, domain-knowledge-informed transfer learning. 

About Dr. Soleimani

Dr. Soleimani is a postdoctoral researcher in the Civil Engineering department at the Samueli School of Engineering. He holds a Ph.D. in Civil Engineering and an MSc in Computer Science, both obtained from Virginia Tech and his research has always been at the intersection of these two areas.

Dr. Tsukasa Yoshinaga aims to identify potential contributing factors to each person’s unique voice quality by using his computational model. Due to potential risks in human subject studies, direct observation of the human larynx is limited. In this study, he develops a computational model of muscular control of the larynx and fluid-structure interaction between the airflow and soft tissues in phonation. Magnetic resonance imaging data are used to develop a realistic geometry of the human larynx and vocal tract. The simulations are conducted to investigate physiological parameters contributing to the production of different voice qualities in normal speakers, trained singers, and patients. These findings will clarify how humans control their unique voices.

About Dr. Yoshinaga

Dr. Yoshinaga is a visiting scholar working with Professor Zhaoyan Zhang in the Department of Head and Neck Surgery at UCLA. He received a Ph.D. in Mechanical Engineering from Osaka University, where he is currently an Assistant Professor. His research focuses on sound generation mechanisms in human speech as well as musical instruments using numerical simulations.

Dr. Mo Zhou conducts advanced studies in the field of stochastic optimal control, mean field control and games, deep learning, and reinforcement learning. He specializes in developing efficient numerical algorithms and performing rigorous theoretical analysis of control problems. Currently, his research focuses on the intersection of mean field games and generative models, which have become a cornerstone of modern AI. These models enable machines to generate data that closely replicate the real world, including images, text, and complex simulations. Dr. Zhou is currently working on an innovative technique to efficiently compute a critical component in generative models, known as the score function. This new technique involves applying a dynamic or ordinary differential equation instead of a traditional deep learning algorithm or a stochastic dynamic. This technique could enhance the performance of generative models.

About Dr. Zhou

Dr. Zhou is an Assistant Adjunct Professor in the Department of Mathematics at UCLA, mentored by Professor Stanley Osher. He earned his bachelor’s degree in mathematics from Tsinghua University in 2018 and his Ph.D. in mathematics from Duke University in 2023, advised by Professor Jianfeng Lu. During his Ph.D. studies, Dr. Zhou developed advanced deep learning algorithms to overcome the curse of dimensionality and address traditional scientific computing challenges, including eigenvalue problems and optimal control problems. Currently, his research focuses on mean field control and games.