Raj Ghugare will present his General Exam "Normalizing Flows are Capable Models for RL" on Thursday, April 30, 2026 at 3:00 PM in CS 105.
Raj Ghugare will present his General Exam "Normalizing Flows are Capable Models for RL" on Thursday, April 30, 2026 at 3:00 PM in CS 105. Committee Members: Benjamin Eysenbach (advisor), Karthik Narasimhan, Brenden Lake Abstract: Modern reinforcement learning (RL) algorithms have found success by using powerful probabilistic models, such as transformers, energy-based models, and diffusion/flow-based models. To this end, RL researchers often choose to pay the price of accommodating these models into their algorithms – diffusion models are expressive, but are computationally intensive due to their reliance on solving differential equations, while autoregressive transformer models are scalable but typically require learning discrete representations. Normalizing flows (NFs), by contrast, seem to provide an appealing alternative, as they enable likelihoods and sampling without solving differential equations or autoregressive architectures. However, their potential in RL has received limited attention, partly due to the prevailing belief that normalizing flows lack sufficient expressivity. We show that this is not the case. Building on recent work in NFs, we propose a single NF architecture which integrates seamlessly into RL algorithms, serving as a policy, Q-function, and occupancy measure. Our approach leads to much simpler algorithms, which often achieve higher performance in imitation learning, offline, goal conditioned RL and unsupervised RL. Reading List: https://docs.google.com/document/d/1cMCRPwSPhpgrE_Qnku0INV3z7dVFFrJeThvF0NZh... Everyone is invited to attend the talk, and those faculty wishing to remain for the oral exam following are welcome to do so.
Yunxiang Chi will present his General Exam "Measurement, Modeling, and Systems for LEO Satellite Networks" on Weds 5/6/2026 at 2:30pm in Friend 005. Committee: Kyle Jamieson(adviser), Jialin Ding, Felix Heide Zoom link: https://docs.google.com/document/d/1-7nowbkiXLynAjWp5jPQlMSkQf0A9Qh7IqCp4zBF... Abstract: LEO satellite networks are emerging as a new systems platform for world-wide connectivity, sensing, and mobile applications. Their operation depends on rapidly changing link conditions driven by satellite geometry, environment, atmosphere, and protocol across multiple layers of the stack. These characteristics make LEO networks both challenging to study and promising as a domain for optimization and application-driven system design. My current research focuses on understanding how real-world geometry and atmospheric conditions affect LEO satellite link quality under a range of operating scenarios. To do so, I am building a measurement-driven testbed that combines live satellite signal observations with satellite geometry and fine-grained atmospheric data. This testbed is used to characterize how factors such as motion, aiming, tropospheric conditions, and ionospheric conditions influence received signal behavior and communication performance. These measurements are then incorporated into a digital twin for accurate temporal and spatial prediction of link quality and network capacity. Our work paves the way for the design of more robust communication and networking protocols and mechanisms for dynamic LEO satellite environments. Reading List: [ https://docs.google.com/document/d/1-7nowbkiXLynAjWp5jPQlMSkQf0A9Qh7IqCp4zBF... | https://docs.google.com/document/d/1-7nowbkiXLynAjWp5jPQlMSkQf0A9Qh7IqCp4zBF... ] Everyone is invited to attend the talk, and those faculty wishing to remain for the oral exam following are welcome to do so.
participants (2)
-
CS Grad Department -
Gradinfo