DeepCast: Towards Personalized QoE for Edge-Assisted Crowdcast With Deep Reinforcement Learning

Wang, Fangxin; Zhang, Cong; Feng, Wang; Liu, Jiangchuan; Zhu, Yifei; Pang, Haitian; Sun, Lifeng

doi:10.1109/tnet.2020.2979966

Cited by 32 publications

(16 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These studies mostly come from two areas of research: computer science and engineering. Their focus is more on providing solutions to problems, as such, they particularly rely on experiments [ 73 , 74 , 75 ], the development of applications, and applied engineering [ 76 , 77 , 78 , 79 ]. Therefore, they do not need a strong grounding in well-established theories, although some do occasionally appear; for example, the Lyapunov optimization theory [ 16 ], the psychophysiological approach [ 80 ], Blommaert’s framework of critical discourse analysis, and Computer-mediated discourse analysis (CMDA) [ 81 ].…”

Section: Resultsmentioning

confidence: 99%

Exploring the Emerging Domain of Research on Video Game Live Streaming in Web of Science: State of the Art, Changes and Trends

Cabeza-Ramírez

García

Fernández

2021

IJERPH

View full text Add to dashboard Cite

In recent years, interest in video game live streaming services has increased as a new communication instrument, social network, source of leisure, and entertainment platform for millions of users. The rise in this type of service has been accompanied by an increase in research on these platforms. As an emerging domain of research focused on this novel phenomenon takes shape, it is necessary to delve into its nature and antecedents. The main objective of this research is to provide a comprehensive reference that allows future analyses to be addressed with greater rigor and theoretical depth. In this work, we developed a meta-review of the literature supported by a bibliometric performance and network analysis (BPNA). We used the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analysis) protocol to obtain a representative sample of 111 published documents since 2012 and indexed in the Web of Science. Additionally, we exposed the main research topics developed to date, which allowed us to detect future research challenges and trends. The findings revealed four specializations or subdomains: studies focused on the transmitter or streamer; the receiver or the audience; the channel or platform; and the transmission process. These four specializations add to the accumulated knowledge through the development of six core themes that emerge: motivations, behaviors, monetization of activities, quality of experience, use of social networks and media, and gender issues.

show abstract

Section: Resultsmentioning

confidence: 99%

Exploring the Emerging Domain of Research on Video Game Live Streaming in Web of Science: State of the Art, Changes and Trends

Cabeza-Ramírez

García

Fernández

2021

IJERPH

View full text Add to dashboard Cite

show abstract

“…Wang et al [81], [82] optimize towards personalized Quality of Experience (QoE) in the use-case of crowdsourced livecast (crowdcast) edge content distribution. Crowdcast is characterized by highly diverse viewer side content watching preferences and environments.…”

Section: ) Edge-layer Content Schedulingmentioning

confidence: 99%

“…The authors argue that existing edge resource allocation approaches use predefined rules or model-based heuristics to make scheduling decisions. Integrating users' personalized QoE-demands results in a N P-complete problem [81]. Wang et al [81] developed DeepCast, a reinforcement learning edge-assisted framework combined with a deep neural network for performing scheduling policy based on real-time network information to accommodate personalized QoE.…”

Section: ) Edge-layer Content Schedulingmentioning

confidence: 99%

“…Integrating users' personalized QoE-demands results in a N P-complete problem [81]. Wang et al [81] developed DeepCast, a reinforcement learning edge-assisted framework combined with a deep neural network for performing scheduling policy based on real-time network information to accommodate personalized QoE. The model learns suitable strategies for scheduling and transcoding using an A3C network model [96].…”

Section: ) Edge-layer Content Schedulingmentioning

confidence: 99%

“…In DeepCast, the edge-server receives the high-bitrate content from the content distribution network, which, if needed, is transcoded (downsampled) at the edge layer before delivering it to the end-user. For evaluation purposes, Wang et al [81] perform trace-driven experiments by integrating three real-world datasets comprising viewing, location, and bandwidth information, respectively. DeepCast aims at minimizing a penalty function that considers QoE, computation cost (from transcoding operations at the edge), and system bandwidth cost.…”

Section: ) Edge-layer Content Schedulingmentioning

confidence: 99%

See 2 more Smart Citations

Learning Combinatorial Optimization on Graphs: A Survey With Applications to Networking

et al. 2020

View full text Add to dashboard Cite

Existing approaches to solving combinatorial optimization problems on graphs suffer from the need to engineer each problem algorithmically, with practical problems recurring in many instances. The practical side of theoretical computer science, such as computational complexity, then needs to be addressed. Relevant developments in machine learning research on graphs are surveyed for this purpose. We organize and compare the structures involved with learning to solve combinatorial optimization problems, with a special eye on the telecommunications domain and its continuous development of live and research networks.

show abstract

Optimizing Uplink Bandwidth Utilization for Crowdsourced Livecast

Zhang

et al. 2022

Lecture Notes in Computer Science

View full text Add to dashboard Cite

DeepCast: Towards Personalized QoE for Edge-Assisted Crowdcast With Deep Reinforcement Learning

Cited by 32 publications

References 18 publications

Exploring the Emerging Domain of Research on Video Game Live Streaming in Web of Science: State of the Art, Changes and Trends

Exploring the Emerging Domain of Research on Video Game Live Streaming in Web of Science: State of the Art, Changes and Trends

Learning Combinatorial Optimization on Graphs: A Survey With Applications to Networking

Optimizing Uplink Bandwidth Utilization for Crowdsourced Livecast

Contact Info

Product

Resources

About