We seek exceptional postdoctoral candidates hosted at UCSD Computer Science & Engineering , jointly advised by by Prof. Rose Yu , Taylor Berg-Kirkpatrick , Yian Ma and Duncan Watson-Parris to start as soon as possible.
Current methods in AI for scientific discovery are ad-hoc. Separate models need to be developed for different scientific experiments and fine-tuned for different prediction tasks. Furthermore, these methods are still limited to uni-modal data, low-dimensional experiments, and lack physical or scientific reasoning.
The goal of the research is to build new foundation models for scientific discovery that can be the basis of an autonomous scientist. This effort envisions an autonomous scientist possessing the ability to characterize its own uncertainty and skepticism, and use them as drivers to systematically acquire and refine its scientific knowledge bases, in a way that human scientist partners can trust.
I am also recruiting postdocs through CSE Fellows Program and Schmidt AI in Science Postdocs Program.
CSE Fellows program has a rolling deadline. Schmidt AI program requires both a Science mentor and a AI mentor.
We seek exceptional postdoctoral candidates hosted at UCSD Computer Science & Engineering by Rose Yu , in collaboration with General Atomics to start as soon as possible.
Simulating complex partial differential equations (PDEs) is a fundamental task in science and engineering. However, it often suffers from high-computational cost and cannot generalize to PDEs with different boundary and initial conditions. Deep learning has emerged as a powerful tool to streamline and accelerate scientific simulation, e.g. [Wang et al. 2020]. But challenges still exist in ensuring the physical consistency of the solution and in handling complex geometries.
The goal of the research is to develop novel hybrid deep learning algorithms and models for solving complex PDEs, especially for gyrokinetic simulations of plasma turbulence in magnetic fusion. The project will explore interdisciplinary techniques in scientific computing, deep learning, and turbulence modeling to help speed up the modeling and simulation progress in fusion and plasma science, with the ultimate goal of realizing of a fusion pilot plant on a decadal timescale.
We seek exceptional postdoctoral candidates to be hosted at UCSD/HDSI, Columbia, or UCI. These positions have an initial term of one year with the possibility of extension. The starting date is flexible, but no later than Fall 2022.
The past two decades have witnessed natural disasters and extreme weather events that affect millions of lives. At the same time, the data volume from high-resolution climate models, satellite, in-situ and ground-based measurements have substantially increased to petabyte scales. These new and readily accessible datasets create the previously missing pipeline for scientific machine learning (ML), which in turn can improve our understanding and ability to predict extreme climate events.
This project will develop deep latent variable models (LVMs) to discover hidden physical structures in high-dimensional, spatiotemporal data of extreme climate events such as droughts or heatwaves, see Project Website for details.We seek exceptional postdoctoral candidates hosted at UCSD Computer Science & Engineering by Prof. Rose Yu , in collaboration with Robin Walters and Jan-Willem Van Der Meernt. The initial term of these positions is 1 year with possibility of extension. The target start date is early 2022 but is flexible.
In the past decade, deep learning has had transformative impacts across society. However, progress has often relied on heuristic methods with limited theoretical understanding. Symmetry plays a key role in human reasoning. Representation theory, which is the mathematical study of symmetry, offers new tools to broaden research into why deep learning works.
The goal of the research is to understand the role of representation theory in enabling efficient optimization and improved generalization of deep learning even in domains with approximate or unknown symmetry. This project pursues three lines of research that will broaden the impact of representation theory in deep learning beyond strict inductive biases. The first is the trade-off between the degree of symmetry in the model and the degree of symmetry in the domain. This line of research will study networks that combine equivariant and non-equivariant features. The second line of research will examine learning symmetry directly from data to improve generalization in domains without known symmetries. The third aim is to develop a theoretical basis for deep learning using quiver representations. This perspective reveals the symmetry of the structure of deep-learning models themselves, through their parameter spaces, even when the domains have no obvious symmetry.We seek exceptional postdoctoral candidates hosted at UCSD Computer Science & Engineering by Prof. Rose Yu . The initial term of these positions is one year with the possiblity for renewal. The target start date is early 2021 but is flexible.
The recent breakthroughs in Artificial Intelligence (AI) and Machine Learning (ML) demonstrate new promises for decision making in complex spatiotemporal environments. Unfortunately, a key challenge to deploy ML in real-world is the large amount of labeled training data required to train complex deep learning models. We will develop sample efficient physics-guided machine learning to learn from spatiotemporal data such as trajectories and videos. The goal is to learn complex spatiotemporal dynamics in various real-world scenarios, including autonomous vehicle tracking and navigation, multi-agent team behavior modeling, and atmospheric turbulence simulation.