Roch Guerin

Roch Guerin

Harold B. and Adelaide G. Welge Professor, Washington University in St. Louis

Conferences

Optimizing Edge Offloading Decisions for Object Detection.

Published:

The paper addresses the problem of object detection in an edge computing setting, where a local weak detector is available but is supplemented by a shared stronger edge detector accessible across a network and available when local decisions are uncertain. Constraints exist that limit how many images can be offloaded to the strong detector. To assist in deciding which images to offload, the paper introduces a metric and a method for estimating it, which enable efficient online offloading decisions. The code for the system described in the paper is available on GitHub and a slightly extended version is available on arXiv

Recommended citation: J. Qiu, R. Wang, B. Hu, R. Guerin, and C. Lu, "Optimizing Edge Offloading Decisions for Object Detection." 2024 ACM/IEEE Symposium on Edge Computing (SEC 2024), Rome, Italy, December 2024. https://doi.org/10.1109/SEC62691.2024.00021

Progressive Neural Compression for Adaptive Image Offloading Under Timing Constraints

Published:

The paper leverages stochastic taildrop to train a rateless encoder that prioritizes transmissions of features of greater importance for inference tasks, e.g., classification. This ensures robust performance even in the presence of severe bandwidth fluctuations.

Recommended citation: R. Wang, H. Liu, J. Qiu, M. Xu, R. Guerin, and C. Lu, "Progressive Neural Compression for Adaptive Image Offloading Under Timing Constraints." 2023 IEEE Real-Time Systems Symposium (RTSS), December 2023, Taipei, Taiwan https://doi.ieeecomputersociety.org/10.1109/RTSS59052.2023.00020

Adaptive Edge Offloading for Image Classification Under Rate Limit

Published:

The paper investigates an edge computing scenario where weak and strong image classifiers located in local devices and an edge server, respectively, collaborate to make the most accurate image classification decisions possible, under the constraint that the number of images that can be offloaded to the strong classifier in the edge server is rate limited using a token bucket mechanism. The paper relies on a reinforcement learning approach to realize a simple policy that maximizes classification accuracy under general image arrival patterns and arbitrary sequences of classification decisions. The code for the system described in the paper is available on GitHub and an extended version of the EMSOFT paper is accessible on arXiv here

Recommended citation: J. Qiu, R. Wang, A. Chakrabarti, R. Guerin, and C. Lu, "Adaptive Edge Offloading for Image Classification Under Rate Limit." IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, 2022. The paper was presented at the ACM International Conference on Embedded Software (EMSOFT), October 2022, Hybrid+Shanghai+Phoenix. https://doi.org/10.1109/TCAD.2022.3197533

Real-Time Edge Classification: Optimal Offloading under Token Bucket Constraints

Published:

This paper designs and validates efficient policies for distributing classification decisions between local devices and more powerful and accurate edge servers

Recommended citation: A. Chakrabarti, R. Guerin, C. Lu, and J. Liu, "Real-Time Edge Classification: Optimal Offloading under Token Bucket Constraints." Proc. The Sixth ACM/IEEE Symposium on Edge Computing (SEC), December 2021, San Jose CA https://arxiv.org/abs/2010.13737

Impact of Distributed Rate Limiting on Load Distribution in a Latency-sensitive Messaging Service

Published:

This paper investigates the trade-off between the better response time that load balancing affords from accessing more resources and the resulting increase in access delays when the resulting rate control function is also distributed

Recommended citation: C. Li, J. Liu, C. Lu, R. Guerin, and C.D. Gill, "Impact of Distributed Rate Limiting on Load Distribution in a Latency-sensitive Messaging Service." Proc. IEEE CLOUD 2021, online virtual congress https://arxiv.org/abs/2101.05865

Minimizing network bandwidth under latency constraints: The single node case

Published:

This paper explores the trade-off between scheduler complexity and the amount of bandwidth required to meet latency constraints in a single node setting

Recommended citation: J. Song, R. Guerin, and H. Sariowan, "Minimizing network bandwidth under latency constraints: The single node case." Proc. 2021 International Teletraffic Congress (ITC 33), Avignon, France, August 2021 https://ieeexplore.ieee.org/document/9625625

Multipath and Rate Stability

Published:

This paper explores if and when the use of multipath solutions leads to more stable end-to-end transmission rates

Recommended citation: J. Liu and R. Guerin, "Multipath and Rate Stability." Proc. IEEE Globecom 2016 - CQRM: Communication QoS, Reliability & Modeling Symposium, Washington, D.C., December 2016 https://openscholarship.wustl.edu/cse_research/1166/

Why didnt my (great!) protocol get adopted?

Published:

This paper applies statistical analysis to features of Internet protocols to identify those most likely to contribute to their success or failure

Recommended citation: M. Nikkhah, C. Dovrolis, and R. Guerin, "Why didnt my (great!) protocol get adopted?" Proc. ACM HotNets, Philadelphia, PA, November 2015 http://dl.acm.org/authorize?N20965

Choice-based pricing for user-provided connectivity

Published:

This paper proposes a pricing policy that seeks to realize an effective compromise between pricing complexity and maximizing system profit in a setting where users contribute resources towards building overall network connectivity

Recommended citation: M.H. Afrasiabi and R. Guerin, "Choice-based pricing for user-provided connectivity." Proc. NetEcon 2015, Portland, OR, June 2015 http://dl.acm.org/authorize?N20954