Dr. Tingting Yuan
Dr. Tingting Yuan
tingting.yuan@cs.uni-goettingen.de
Computer Networks Group,
Institute of Computer Science,
Georg-August-University of Göttingen
Göttingen, Lower Saxony, Germany
Address: Room 3.109, Goldschmidtstr. 7, 37077 Goettingen, Germany
Interests
Communication and Networking
Neural-based Video analytics
UAV-assistant networking
Multi-agent reinforcement learning
Communication-aware AI
AI for networking and computing
Upcoming Events
Recent News
28-30 Feb 2024 - Attend Cover 2nd-year plenary meeting in London.
6-8 Feb 2024 - Attend CODECO's 5th plenary meeting in Madrid, Spain.
2-6 Oct 2023 - Attend Mobicom'23 and host MobiArch'23.
Aug-Sept 2023 - Visit CUHK, PolyU, Tsinghua University, Fudan University, BUPT, UESTC, SUN YAT-SEN UNIVERSITY, and Southeast University.
Give a talk on "AI meets networking"4-5 July 2023 - Attend CODECO's 3rd plenary meeting in Greece.
19 May. 2023 - I am invited to give a talk on AccDecoder in Infocomm 23, New York.
21 Mar. 2023 - I am invited to give a talk on AccDecoder and DACOM at the Dept. of Software Engineering, Sun Yat-sen University.
16-17 Jan. 2023 -I will attend CODECO kick-off meeting in München.
18 Jan. 2023 -I will visit DLR in München.
22 Nov. 2022 -06.Jan. 2023 -I visit University of Essex, King's College London, University of Birmingham, etc.
16 Dec. 2022 - I am invited to give a talk on Neural-enhanced Video Analytics at the Dept. of Computer Science, King's College London, hosted by Dr. Yali Du.
15 Dec. 2022 - I visit Birmingham City Council for COSAFE project.
14 Dec. 2022 - I visit the Dept. of Computer Science, University of Birmingham, hosted by Dr. Shan He.
09 Dec. 2022 - I am invited to give a talk on Neural-enhanced Video Analytics at the Dept. of Engineering, King's College London, hosted by Dr. Deng Yansha.
08 Dec. 2022 - I am invited to give a talk on cooperative multi-agent learning and applications at the Dept. of Computer Science, University of Essex, hosted by Prof. Jianhua He.
2 Dec. 2022 - One paper on “AccDecoder: Accelerated Decoding for Neural-enhanced Video Analytics” has been accepted by the 42nd IEEE International Conference on Computer Communications (INFOCOM 2023).
16 Nov. 2022 - One paper on “DACOM: Learning Delay-Aware Communication for Multi-Agent Reinforcement Learning” has been accepted by Proc. 37th AAAI Conference on Artificial Intelligence (AAAI 2023). http://arxiv.org/abs/2212.01619
23 Nov 2022 - Visiting Cooperative AI Research Lab, KCL.
01 Nov. 2021- Got Humboldt Postdoctoral Researcher and working at the University of Göttingen.
Publications
Selected Publications
DACOM: Learning Delay-Aware Communication for
Multi-Agent Reinforcement Learning
AAAI'23
Communication is supposed to improve multi-agent collaboration and overall performance in cooperative Multi-agent reinforcement learning (MARL). However, such improvements are prevalently limited in practice since most existing communication schemes ignore communication overheads (e.g., communication delays). In this paper, we demonstrate that ignoring communication delays has detrimental effects on collaborations, especially in delay-sensitive tasks such as autonomous driving. To mitigate this impact, we design a delay-aware multi-agent communication model (DACOM) to adapt communication to delays. Specifically, DACOM introduces a component, TimeNet, that is responsible for adjusting the waiting time of an agent to receive messages from other agents such that the uncertainty associated with delay can be addressed. Our experiments reveal that DACOM has a non-negligible performance improvement over other mechanisms by making a better trade-off between the benefits of communication and the costs of waiting for messages.
AccDecoder: Accelerated Decoding for
Neural-enhanced Video Analytics
IEEE INFOCOM' 23
The quality of the video stream is key to neural network-based video analytics. However, low-quality video is inevitably collected by existing surveillance systems because of poor-quality cameras or over-compressed/pruned video streaming protocols, e.g., as a result of upstream bandwidth limit. To address this issue, existing studies use quality enhancers (e.g., neural super-resolution) to improve the quality of videos (e.g., resolution) and eventually ensure inference accuracy. Nevertheless, directly applying quality enhancers does not work in practice because it will introduce unacceptable latency. In this paper, we present AccDecoder, a novel accelerated decoder for real-time and neural-enhanced video analytics. AccDecoder can select a few frames adaptively via Deep Reinforcement Learning (DRL) to enhance the quality by neural super-resolution and then up-scale the unselected frames that reference them, which leads to a 6-21% accuracy improvement. AccDecoder provides efficient inference capability via filtering important frames using DRL for DNN-based inference and reusing the results for the other frames via extracting the reference relationship among frames and blocks, which results in a latency reduction of 20-80% than baselines.
BiSwift: Bandwidth Orchestrator for Multi-Stream
Video Analytics on Edge
IEEE INFOCOM' 24
High-definition (HD) cameras for surveillance and road traffic have experienced tremendous growth, demanding intensive computation resources for real-time analytics. Recently, offloading frames from the front-end device to the back-end edge server has shown great promise. In multi-stream competitive environments, efficient bandwidth management and proper scheduling are crucial to ensure both high inference accuracy and high throughput. To achieve this goal, we propose BiSwift, a bi-level framework that scales the concurrent real-time video analytics by a novel adaptive hybrid codec integrated with multi-level pipelines, and a global bandwidth controller for multiple video streams. The lower-level front-back-end collaborative mechanism (called adaptive hybrid codec) locally optimizes the accuracy and accelerates end-to-end video analytics for a single stream. The upper-level scheduler aims to accuracy fairness among multiple streams via the global bandwidth controller. The evaluation of BiSwift shows that BiSwift is able to real-time object detection on 9 streams with an edge device only equipped with an NVIDIA RTX3070 (8G) GPU. BiSwift improves 10-21% accuracy and presents 1.2-9× throughput compared with the state-of-the-art video analytics pipelines.