Pouya Parsa

Pouya Parsa, Kawon Han, Seongjin Choi

Frontier autonomous-driving models -- especially vision-language-action (VLA) models, whose forward pass approaches $\sim$60~TFLOPs -- are outgrowing economical onboard deployment, since peak hardware sits idle most of the day. Cloud inference can instead share GPUs across active vehicles, but the vehicle must upload through a capacity-limited uplink, reach a GPU without queueing, and return a decision within the closed-loop budget. This paper asks: can the cloud drive? We answer with an analytical framework coupling communication limits, a roofline GPU service model, stochastic latency, and utilization-aware cost across three model classes, three offloading strategies, and three communication generations, applied to New York City. Separating a reactive 100~ms budget from a 300~ms deliberative tier (presuming an onboard reactive fallback), we find three \emph{nested} binding regimes. Communication binds first in dense cells: 5G fails early, 5G-Advanced is the practical threshold for feature-level offloading, and 6G adds headroom. Compute binds next under the reactive budget: near-term VLA is latency-infeasible regardless of bandwidth, because autoregressive FP16 decode is memory-bandwidth-bound (~114 ms on 2025 hardware). Its floor clears 100 ms around 2027; 6G then admits feature-level VLA by ~2028, 5G-Advanced only at light loading and not the dense corridor, and the deliberative tier from 2026. Cost binds last: once admissible, utilization-pooled cloud GPUs undercut onboard hardware for VLA, whose baseline (up to \$8,500 per vehicle-year) is expensive and idle; feature-level offloading (S2) is where the VLA cost crossover concentrates. Latency decides which model is admissible in which year; cost decides whether it is economical.

Pouya Parsa, Raoof Zare Moayedi, Mohammad Bornosi et al.

The reinforcement learning algorithms that focus on how to compute the gradient and choose next actions, are effectively improved the performance of the agents. However, these algorithms are environment-agnostic. This means that the algorithms did not use the knowledge that has been captured by trajectory. This poses that the algorithms should sample many trajectories to train the model. By considering the essence of environment and how much the agent learn from each scenario in that environment, the strategy of the learning procedure can be changed. The strategy retrieves more informative trajectories, so the agent can learn with fewer trajectory sample. We propose Where2Start algorithm that selects the initial state so that the agent has more instability in vicinity of that state. We show that this kind of selection decreases number of trajectories that should be sampled that the agent reach to acceptable reward. Our experiments shows that Where2Start can improve sample efficiency up to 8 times. Also Where2Start can combined with most of state-of-the-art algorithms and improve that robustness and sample efficiency significantly.

Wenya Xie, Shengming Zhou, Zelin Li et al.

LLM agents increasingly act as personal assistants that must remember a user's profile over months: who they are (attributes), what they routinely do (habits), and what they prefer (preferences), and keep it updated as jobs, routines, and tastes drift. Existing benchmarks evaluate this "memory" ability through short, simplified interactions, missing three core properties of real behavior: the profile is heterogeneous, with attributes, habits, and preferences evolving on different timelines; changes are driven by external context such as seasons and life events; and evidence is rarely stated explicitly, instead scattered across many small actions in different apps that a memory system must infer from. We introduce DynamicMem, a synthetic benchmark that constructs 15 months of activity per user, providing long-term multi-app data that real users' privacy keeps out of reach. It provides user-consistent trajectories averaging 2.2M tokens and 1,772 grounded events per user across 16 applications such as e-commerce, fitness, and social platforms. The profile evolves over this period and is never given explicitly: each attribute, habit, or preference must be inferred from small signals scattered across apps. We evaluate at five quarterly checkpoints to track how systems scale as history grows. Benchmarking five representative systems exposes problems a single accuracy score hides: (i) profile reconstruction degrades with history length while service-task accuracy stays flat, despite both drawing on the same memory; (ii) no system both keeps facts that stay true and replaces facts that change, with errors clustering on preferences and on naming the exact referent; and (iii) over 93% of failures trace to what the memory retrieves, not to the model writing the answer, so the largest room for improvement lies in memory itself. Code: https://wenyaxie023.github.io/DynamicMem/

Pouya Parsa, Keya Li, Kara M. Kockelman et al.

Automatic license plate recognition (ALPR) and vehicle make and model recognition underpin intelligent transportation systems, supporting law enforcement, toll collection, and post-incident investigation. Applying these methods to videos captured by handheld smartphones or non-static vehicle-mounted cameras presents unique challenges compared to fixed installations, including frequent camera motion, varying viewpoints, occlusions, and unknown road geometry. Traditional ALPR solutions, dependent on specialized hardware and handcrafted OCR pipelines, often degrade under these conditions. Recent advances in large vision-language models (VLMs) enable direct recognition of textual and semantic attributes from arbitrary imagery. This study evaluates the potential of VLMs for ALPR and makes and models recognition using monocular videos captured with handheld smartphones and non-static mounted cameras. The proposed license plate recognition pipeline filters to sharp frames, then sends a multimodal prompt to a VLM using several prompt strategies. Make and model recognition pipeline runs the same VLM with a revised prompt and an optional self-reflection module. In the self-reflection module, the model contrasts the query image with a reference from a 134-class dataset, correcting mismatches. Experiments on a smartphone dataset collected on the campus of the University of Texas at Austin, achieve top-1 accuracies of 91.67% for ALPR and 66.67% for make and model recognition. On the public UFPR-ALPR dataset, the approach attains 83.05% and 61.07%, respectively. The self-reflection module further improves results by 5.72% on average for make and model recognition. These findings demonstrate that VLMs provide a cost-effective solution for scalable, in-motion traffic video analysis.