Pith. sign in

REVIEW 3 major objections 7 minor 62 references

Learning Driven Elastic Task Multi-Connectivity Immersive Computing Systems

T0 review · 3 major / 7 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A centralized phasic policy-gradient agent that sees all users jointly cuts VR offloading latency by 28% and energy consumption by 78% compared with independent per-user agents.

desk verdict A competent centralized-offloading DRL paper whose headline 28% latency / 78% energy gains likely rest on an unimplemented shared-MEC constraint; treat the margins as provisional until the allocation mechanism is specified. read the letter →

arxiv 2505.13331 v1 pith:4QT433L7 submitted 2025-05-19 cs.NI cs.ETcs.MMcs.SYeess.SY

Babak Badnava , Jacob Chakareski , Morteza Hashemi This is my paper · ORCID
classification cs.NIcs.ETcs.MMcs.SYeess.SY
keywords elastictaskoffloadingmulti-connectivityVRsystemsedgecomputingcomputationalenergy-efficiencymulti-agentreinforcementlearningcentralizedversusdecentralizeddecision-makingphasicpolicygradient360-degreevideo
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper argues that when VR users with elastic computational tasks share an edge server and several wireless channels, offloading decisions are best made by a single centralized agent that sees every user's task and history, rather than by independent per-user agents. The authors formulate the problem as a constrained stochastic optimization that maximizes computational throughput per energy unit while respecting task deadlines, and solve it online with a centralized phasic policy gradient (CPPG) agent. On a trace-driven simulator with 30 users, real 4G/5G/WiGig throughput traces, and a 360-degree video rendering task set, they report that CPPG reduces response latency by 28% and energy consumption by 78% relative to the independent phasic policy gradient (IPPG) baseline, at the price of full observability. If the comparison holds, it means that coordination through shared edge resources, not just per-user channel selection, is where the largest gains in immersive edge computing lie.

What carries the argument

The load-bearing mechanism is the constrained computational energy-efficiency objective of Eq. (11), paired with the CPPG actor-critic architecture. Phasic policy gradient here means a PPO-style training loop with a separate auxiliary phase: after the usual policy and value update, the network is further trained on behavioral-cloning and value losses, which the paper argues reduces interference and lets the joint state shape each user's reward. The centralized actor takes a joint action $[u^1, \dots, u^K]$, while the critic emits per-user state values. The per-user reward folds energy efficiency and deadline compliance into one scalar, so each environment step is self-contained and no bootstrapping is needed. The shared-MEC constraint $\sum_k z(f^k_{\mathrm{mec}}) < Z_{\mathrm{mec}}$ is what makes the joint decision matter in principle: one user's offload changes the edge compute available to others, and CPPG is designed to capture that coupling.

What would settle it

Instrument the offloading simulator to record, for every task, the MEC frequency $f^k_{\mathrm{mec}}$ actually assigned to each user, and rerun the comparison with the shared constraint $\sum_k z(f^k_{\mathrm{mec}}) < Z_{\mathrm{mec}}$ replaced by independent per-user MEC caps. If CPPG still shows roughly 28% lower latency and 78% lower energy when no user's offloading affects another's compute allocation, the result is an observability artifact; if the gap disappears, shared-resource coordination is doing the work.

Watch

Extended reading notes

Core claim

The paper's central claim is that a centralized multi-agent reinforcement-learning agent, CPPG, can solve the constrained stochastic computational energy-efficiency maximization problem for multi-connectivity VR task offloading in real time. CPPG observes the full system state, namely each user's elastic task size and computational intensity plus past transmission times, response times, and energy consumption, and outputs a joint offloading action for all users, choosing local execution or one of the available channels (4G, 5G, or WiGig) to a shared MEC unit. The reward for each user is the task's computational throughput per joule plus a learned deadline-violation penalty, and because this reward is self-contained the discount factor can be set to zero. After training on real network traces and 360-degree video data, the authors report that CPPG outperforms the decentralized IPPG baseline by 28% in latency and 78% in energy consumption, and that the learned policy offloads larger tasks, favors 5G and WiGig over 4G, and shifts computation toward whichever side, headset or MEC, has more available speed.

Load-bearing premise

The headline comparison assumes that the users genuinely share one constrained edge-computing resource during the experiments; if each user effectively gets independent edge service, the reported CPPG advantage would come from central observability rather than from coordination.

Editorial extensions

If this is right

  • A single edge-side controller can serve many VR users with one inference pass: CPPG's decision time is $O(1)$ in the number of users, while decentralized IPPG scales as $O(K)$.
  • Per-user, partially observable offloading is materially worse on this workload: the reported numbers imply 28% higher latency and 78% higher energy consumption compared with centralized joint decisions.
  • The learned policy is task-size aware: larger elastic tasks are increasingly offloaded rather than computed locally, and higher-throughput channels (5G, WiGig) absorb the offloaded load while 4G is avoided.
  • System resource changes steer the policy in predictable directions: faster VR headsets pull computation local, faster MEC pulls it to the edge, and both moves improve response time.
  • Quality-of-experience targets are reachable: in the 30-user test setting the paper reports 12.24% of tasks missing the one-second deadline overall, or 0.4% per user.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the claimed 28%/78% edge comes mostly from full observability, a decentralized agent given a compact aggregate signal, such as the MEC's current load, could recover part of that gap; the paper does not test this.
  • Because the reward is a per-task scalar with no temporal dependency, the centralized problem is closer to contextual bandits than to full RL; comparing CPPG against a contextual bandit with the same joint state would isolate what the phasic policy-gradient machinery adds.
  • The shared-MEC constraint of Eq. (11d) is asserted, but the simulator's allocation of MEC frequency among users is not specified; if users do not actually contend for one MEC, the reported advantage is an information advantage by construction.
  • Task elasticity could be turned into an active control: instead of taking task size as given, an agent could select the video-quality layer jointly with the offloading action, extending energy efficiency to a rate-quality trade-off frontier.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 7 minor

Summary. The paper studies elastic task offloading in a multi-user, multi-connectivity edge-assisted VR system. It formulates a constrained stochastic optimization problem that maximizes the sum of computational energy-efficiency across users, subject to per-user deadline constraints and a shared MEC computation capacity constraint (Eq. 11d). To solve this online, the authors propose three learning-based methods: a centralized phasic policy gradient (CPPG) agent with full observability, a decentralized independent phasic policy gradient (IPPG) agent, and decentralized shared multi-armed bandit variants (DSMAB). The agents are trained in a custom gym-like simulator driven by real 4G/5G/WiGig throughput traces and a 360-degree video dataset. The main reported result is that CPPG reduces latency by 28% and energy consumption by 78% compared to IPPG. The paper also analyzes the effect of task size, scalability with the number of users, and sensitivity to headset and MEC computational speeds.

Significance. If the reported gains are real and stem from the mechanism the authors invoke (centralized coordination under a shared MEC resource), the paper would provide a useful benchmark for centralized versus decentralized decision-making in edge-assisted immersive computing. The strengths are the use of real network traces and a public 360-degree video dataset, the formulation of an elastic-task model with a deadline-aware reward, and the inclusion of several decentralized alternatives for comparison. However, the central interpretation depends on the simulator faithfully instantiating the shared MEC capacity constraint, which is not demonstrated in the manuscript. As written, the headline comparison may reflect differences in state information and architecture rather than coordination benefits from handling shared computational resources, so the significance of the result is conditional.

major comments (3)
  1. [Sections III-A, IV, and VI (Eqs. (2), (11b), (11d))] The shared MEC capacity constraint that motivates centralized decision-making is not shown to be implemented in the simulator. Eq. (2) and Eq. (11b) define the per-user MEC frequency f^mec_k, and Eq. (11d) imposes sum_k z(f^mec_k) < Zmec, but Section VI, which describes the simulator, never specifies how Zmec is divided among concurrently offloading users, what value f^mec_k takes in the experiments, or whether any queueing or service discipline is modeled at the MEC. Since the reward for user k in Eq. (17) depends on f^mec_k only through that user's response time T_r^k, if f^mec_k is a fixed constant independent of the other users' actions, then the per-user rewards are separable and the environment contains no multi-user coupling. In that case the advantage of CPPG over IPPG would not be evidence of successful coordination under shared resources but rather a consequence of giving the centralized agent a richer state representation and a joint action space; the claim in Section V-A that CPPG 'capture[s] the effect of shared computational resources on the reward each user receives' would then be unsupported by the experiments. The authors should describe the MEC resource allocation rule (e.g., equal splitting, proportional demand, or a contention model), confirm that Eq. (11d) is actually enforced in the simulator, and re-run the comparison in an environment with genuine coupling before attributing the 28%/78% margins to centralized coordination.
  2. [Section V-A, Eqs. (14) and (25), Algorithm 1] The deadline coefficient update contains a sign inconsistency. Eq. (14) defines L_{\lambda_k} = \lambda_k(T_d^k - T_r^k), so the gradient used in Eq. (25) is \nabla_{\lambda_k} L = T_d^k - T_r^k. When a task misses its deadline (T_r > T_d), this gradient is negative, and the update \lambda_k \leftarrow \lambda_k + \eta \nabla_{\lambda_k} L decreases \lambda_k. The text states that \lambda_k should increase when the response time exceeds the deadline, which is the opposite of what the formula implements. Standard dual ascent would use the constraint violation T_r - T_d with a positive step size. Please correct the sign (or the definition of L_{\lambda_k}) and justify the resulting update. This is load-bearing because the reward in Eq. (17) depends on \lambda_k, and the reported deadline violation probabilities in Section VI, Fig. 9a, are outputs of this procedure.
  3. [Section IV, Eqs. (11)-(14)] The passage from the constrained problem (11) to the Lagrangian form (12) and then to Eqs. (13)-(14) is not a valid duality argument as written. Minimizing \lambda_k(T_d^k - T_r^k(u_k^*)) over \lambda_k in Eq. (14) is degenerate: when the constraint is feasible the minimum is attained at \lambda_k = 0, and when it is infeasible the expression is unbounded below for \lambda_k \ge 0. The paper should either provide a correct derivation of the Lagrangian dual (e.g., with \lambda_k \ge 0 and the standard saddle-point formulation) or explicitly state that the deadline penalties are a heuristic mechanism rather than the exact dual of Eq. (11). Since the training algorithm and the reported constraint-satisfaction results are built on this formulation, the distinction is consequential for the correctness of the method.
minor comments (7)
  1. [Section I] There is a typo in 'elsatic tasks' in the first paragraph of the Introduction; it should read 'elastic tasks'.
  2. [Section V-C] 'refereed to' appears twice (for ILin-UCB and ILin-TS) and should be 'referred to'.
  3. [Section V-A, Eqs. (23) and (29)] 'L Jonit' is a typo for 'L Joint' in both equations.
  4. [Section VI, Table I] The DRL training hyperparameters (learning rates, batch sizes, optimizer, clip epsilon, discount factor) are not listed; without them the experiments are not reproducible. Please add a complete hyperparameter table.
  5. [Section VI, second paragraph] The state size for IPPG is stated as 15, but the composition of the 15 features (e.g., the history length used for transmission time, response time, and energy consumption) is not specified. Please define the state vector explicitly.
  6. [Section VI, Figure 8] The markers for the different methods are difficult to distinguish because several points cluster in the same region. Consider reporting a table of mean and standard deviation values, or labeling the points directly.
  7. [Abstract and Section VII] The 28% latency and 78% energy reductions are stated without the experimental conditions (e.g., K=30, specific traces, and testing episodes). Please add a qualifier or refer the reader to the exact setup in Section VI.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the claimed CPPG/IPPG gap is an empirical simulation result, not a fitted input or a self-cited theorem.

full rationale

The paper's central claim is that the centralized CPPG agent outperforms the decentralized IPPG agent on a trace-driven offloading simulator, reducing latency by 28% and energy consumption by 78%. These numbers are measured test-time outcomes of training two agents on the same Lagrangian objective and the same per-user reward function (Eqs. 12 and 17), rather than parameters fitted to match those target reductions. No load-bearing step in the derivation is defined in terms of its own conclusion: the shared-MEC constraint in Eq. (11d) is a stated modeling constraint, but the paper never claims to have derived an allocation rule, so the question of whether the simulator actually implements the coupling is a validation or correctness issue, not a circularity in the mathematical argument. The authors' self-citations supply empirical data (e.g., the 360-degree video dataset [61]) or background context, and are not used as an external authority to force the central comparison. No uniqueness theorem or ansatz is imported from prior work to justify the design choice. The lack of external baselines weakens the external validity of the reported margins, but it does not make the derivation circular in the sense of reducing by construction to its inputs.

Assumptions & free parameters 5 free parameters · 6 assumptions · 0 invented entities

The central claim rests on a suite of domain assumptions about computation time, CPU energy, and reward stationarity, plus a silent MEC allocation mechanism. There are no curve fits to targets, but several hand-chosen constants (lambda_k, beta) and unreported environment details (I_k distribution, f^mec_k allocation, DRL hyperparameters) shape the results. No new physical entities are introduced.

free parameters (5)
  • Initial deadline coefficient lambda_k = 16
    Set by hand in Table I; scales the deadline-violation penalty in the reward (Eq. 17) and is adapted during training via Eq. (25). The learned policy depends on this initial value and its update schedule.
  • Entropy weight beta = 0.01
    Balances exploration and exploitation in the dual-clip PPO loss (Eq. 19); chosen by hand in Table I and not ablated.
  • MEC per-user CPU frequency allocation f^mec_k = not specified
    Determines offloaded task computation time (Eqs. 2 and 11b) and the shared-capacity constraint (11d); the paper never specifies how the MEC divides its compute among K users, leaving a silent degree of freedom in the simulator.
  • Task computational intensity I_k = not specified
    The task model (Section III-A) defines I_k as a task feature, but Section VI never states the distribution or values used to sample computational intensity for the 360-video render tasks, so a main driver of computation time is not pinned down.
  • DRL training hyperparameters (learning rates, batch sizes, clip epsilon, optimizer) = not reported
    Algorithm 1 and Table I omit learning rates, batch sizes, the PPO clip parameter, replay buffer sizes, and optimizer details, which are needed to reproduce the training dynamics.
assumptions (6)
  • domain assumption Computation time scales linearly as Te = S * I / z(f) with no queuing, caching, or parallel processing effects (Eq. 3).
    The entire latency model rests on this linear scaling; a real shared MEC server would introduce queueing delays, which the model omits.
  • domain assumption CPU dynamic power dominates, with local energy E = kappa * S * I * f^2 and voltage approximately linear in frequency (Eq. 6, citing [50, 51]).
    A standard but approximate energy model; this local-energy term largely drives the offloading policy decisions reported in Section VI.
  • domain assumption Rewards have no temporal dependency, allowing gamma = 0 and reducing the RL problem to a contextual bandit (Section V-A, Eqs. 15-17).
    This is in tension with the state vector containing past transmission times, response times, and energy consumptions as features, and with MEC load varying over time; if temporal structure matters, gamma = 0 discards it.
  • standard math The Lagrangian penalty in Eqs. (12)-(14) with the sub-gradient update (25) drives deadline feasibility.
    Assumes the dual function is well behaved and the penalty coefficient update converges; no convergence analysis or constraint-satisfaction statistics are provided.
  • ad hoc to paper The MEC shared-capacity constraint (11d) is satisfied by an allocation of f^mec_k whose mechanism is never described.
    The coupling between users that the paper credits for CPPG's advantage depends on an allocation mechanism the paper does not define or validate.
  • domain assumption Expected throughput over the transmission interval (Eq. 4) adequately represents time-varying 4G/5G/WiGig channel conditions.
    Replaces the fast-varying rate process with an average, which can bias transmission-time estimates for short tasks and ignores rate variability within a transmission.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Learning Driven Elastic Task Multi-Connectivity Immersive Computing Systems." pith.science (2026). https://pith.science/paper/4QT433L7

@misc{pith2026250513331,
  author       = {Pith},
  title        = {Pith review of: Learning Driven Elastic Task Multi-Connectivity Immersive Computing Systems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4QT433L7}},
  note         = {Machine review of arXiv:2505.13331}
}
read the original abstract

In virtual reality (VR) environments, computational tasks exhibit an elastic nature, meaning they can dynamically adjust based on various user and system constraints. This elasticity is essential for maintaining immersive experiences; however, it also introduces challenges for communication and computing in VR systems. In this paper, we investigate elastic task offloading for multi-user edge-computing-enabled VR systems with multi-connectivity, aiming to maximize the computational energy-efficiency (computational throughput per unit of energy consumed). To balance the induced communication, computation, energy consumption, and quality of experience trade-offs due to the elasticity of VR tasks, we formulate a constrained stochastic computational energy-efficiency optimization problem that integrates the multi-connectivity/multi-user action space and the elastic nature of VR computational tasks. We formulate a centralized phasic policy gradient (CPPG) framework to solve the problem of interest online, using only prior elastic task offloading statistics (energy consumption, response time, and transmission time), and task information (i.e., task size and computational intensity), while observing the induced system performance (energy consumption and latency). We further extend our approach to decentralized learning by formulating an independent phasic policy gradient (IPPG) method and a decentralized shared multi-armed bandit (DSMAB) method. We train our methods with real-world 4G, 5G, and WiGig network traces and 360 video datasets to evaluate their performance in terms of response time, energy efficiency, scalability, and delivered quality of experience. We also provide a comprehensive analysis of task size and its effect on offloading policy and system performance. In particular, we show that CPPG reduces latency by 28% and energy consumption by 78% compared to IPPG.

Figures

Figures reproduced from arXiv: 2505.13331 by the authors.

Figure 1
Figure 1. Multi-user multi-connectivity edge-assisted AR/VR offloading [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Task Elasticity: Task computational and communication re [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. The CPPG agent follows an actor-critic model. The actor [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: The IPPG agent observes the state of each agent and makes [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: DSMAB agent observes contextual information of a user’s [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Empirical cumulative distribution function (CDF) and empir [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]
Figure 7
Figure 7. Figure 7: Network throughput cumulative distribution function (CDF) of [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]
Figure 8
Figure 8. Figure 8: Performance trade-offs between response time and energy [PITH_FULL_IMAGE:figures/full_fig_p010_8.png]
Figure 9
Figure 9. Figure 9: a) The trade-offs between viewport PSNR and task deadline violations as task size increases. (b) Examination of task size impact on [PITH_FULL_IMAGE:figures/full_fig_p011_9.png]
Figure 10
Figure 10. Figure 10: Scalability analysis of the proposed methods. Each bar and the vertical line represent the average and standard deviation of task [PITH_FULL_IMAGE:figures/full_fig_p012_10.png]
Figure 11
Figure 11. Figure 11: (a) As the VR headset’s computational speed increases, a higher percentage of tasks are computed locally, resulting in reduced [PITH_FULL_IMAGE:figures/full_fig_p012_11.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

62 extracted references · 57 canonical work pages

  1. [1]

    State of the Edge,

    SeaGate, “State of the Edge,” https://www.seagate. com/www-content/enterprise-storage/it-4-0/images/ Data-At-The-Edge-UP1.pdf, 2019, [Online]

    work page 2019

  2. [2]

    Towards Enabling Next Generation Societal Virtual Reality Applications for Vir- tual Human Teleportation,

    J. Chakareski, M. Khan, and M. Yuksel, “Towards Enabling Next Generation Societal Virtual Reality Applications for Vir- tual Human Teleportation,” IEEE Signal Processing Magazine , Sep. 2022

    work page 2022

  3. [3]

    Millimeter Wave and Free-Space-Optics for Future Dual- Connectivity 6DOF Mobile Multi-User VR Streaming,

    J. Chakareski, M. Khan, T. Ropitault, and S. Blandino, “Millimeter Wave and Free-Space-Optics for Future Dual- Connectivity 6DOF Mobile Multi-User VR Streaming,” ACM Trans. Multimedia Computing, Communications, and Applica- tions., vol. 19, no. 2, pp. 1–25, Feb. 2023

    work page 2023

  4. [4]

    Virtual Reality (VR) Market Size, Share & Industry Analysis,

    F. B. Insights, “Virtual Reality (VR) Market Size, Share & Industry Analysis,” https://www.fortunebusinessinsights.com/ industry-reports/virtual-reality-market-101378, [Online]

    work page

  5. [5]

    Ghazikor, K

    M. Ghazikor, K. Roach, K. Cheung, and M. Hashemi, “Channel-aware distributed transmission control and video 12 8 15 20 30 Number of Users 0.0 0.2 0.4 0.6 0.8Response Time (s) CPPG IPPG INN -Greedy INN-TS ILin-TS INN-UCB ILin-UCB (a) Response Time Scalability 8 15 20 30 Number of Users 0 1 2 3Energy Consumption (mW) CPPG IPPG INN -Greedy INN-TS ILin-TS INN...

    arXiv 2024

  6. [6]

    Phasic Policy Gradient,

    K. Cobbe, J. Hilton, O. Klimov, and J. Schulman, “Phasic Policy Gradient,” arXiv:2009.04416, 2020

    arXiv 2009

  7. [7]

    Openai gym,

    G. Brockman, V . Cheung, L. Pettersson, J. Schneider, J. Schulman, J. Tang, and W. Zaremba, “Openai gym,” arXiv:1606.01540, 2016

    arXiv 2016

  8. [8]

    Adaptive User-managed Service Placement for Mobile Edge Computing: An Online Learning Approach,

    T. Ouyang, R. Li, X. Chen, Z. Zhou, and X. Tang, “Adaptive User-managed Service Placement for Mobile Edge Computing: An Online Learning Approach,” in IEEE INFOCOM, 2019

    work page 2019

Show all 62 references
  1. [9]

    Joint scheduling of communication and computation resources in multiuser wireless application offloading,

    M. Molina, O. Mu ˜noz, A. Pascual-Iserte, and J. Vidal, “Joint scheduling of communication and computation resources in multiuser wireless application offloading,” in IEEE 25th PIMRC, 2014

    2014

  2. [10]

    Edge-Centric Bandit Learning for Task-Offloading Allocations in Multi-RAT Hetero- geneous Networks,

    B. Wu, T. Chen, K. Yang, and X. Wang, “Edge-Centric Bandit Learning for Task-Offloading Allocations in Multi-RAT Hetero- geneous Networks,” IEEE Tran. on Vehicular Technology, 2021

    2021

  3. [11]

    Decentralized Task Offloading in Edge Computing: A Multi-User Multi-Armed Bandit Ap- proach,

    X. Wang, J. Ye, and J. C. Lui, “Decentralized Task Offloading in Edge Computing: A Multi-User Multi-Armed Bandit Ap- proach,” in IEEE INFOCOM, 2022

    2022

  4. [12]

    Optimal dynamic spectrum allocation-assisted latency minimization for multiuser mobile edge computing,

    Y . Li, T. Wang, Y . Wu, and W. Jia, “Optimal dynamic spectrum allocation-assisted latency minimization for multiuser mobile edge computing,” Digital Communications and Networks, 2022

    2022

  5. [13]

    Fine-Grained Task Offloading for UA V via MEC-Enabled Networks,

    S. Huang, L. Li, Q. Pan, W. Zheng, and Z. Lu, “Fine-Grained Task Offloading for UA V via MEC-Enabled Networks,” inIEEE 30th International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC Workshops) , 2019

    2019

  6. [14]

    Learning-Based Queuing Delay-Aware Task Offloading in Collaborative Vehic- ular Networks,

    Z. Jia, Z. Zhou, X. Wang, and S. Mumtaz, “Learning-Based Queuing Delay-Aware Task Offloading in Collaborative Vehic- ular Networks,” in IEEE ICC, 2021

    2021

  7. [15]

    Joint Com- putation Offloading and Resource Allocation in UA V Swarms with Multi-access Edge Computing,

    W. Liu, Y . Xu, N. Qi, K. Yao, Y . Zhang, and W. He, “Joint Com- putation Offloading and Resource Allocation in UA V Swarms with Multi-access Edge Computing,” in International Confer- ence on Wireless Communications and Signal Processing, 2020

    2020

  8. [16]

    Cooperative Computation Offloading for UA Vs: A Joint Radio and Com- puting Resource Allocation Approach,

    S. Zhu, L. Gui, J. Chen, Q. Zhang, and N. Zhang, “Cooperative Computation Offloading for UA Vs: A Joint Radio and Com- puting Resource Allocation Approach,” in IEEE International Conference on Edge Computing (EDGE) , 2018

    2018

  9. [17]

    UA V-Assisted Cellular System: Offloading Strategy and Bandwidth Alloca- tion,

    L. Cao, W. Ni, H. Tian, M. Hua, and G. Hao, “UA V-Assisted Cellular System: Offloading Strategy and Bandwidth Alloca- tion,” in International Conference on Space-Air-Ground Com- puting (SAGC), 2020

    2020

  10. [18]

    Energy-Aware Task Offloading for Ultra-Dense Edge Computing,

    J. Zhang, H. Guo, and J. Liu, “Energy-Aware Task Offloading for Ultra-Dense Edge Computing,” in IEEE iThings and IEEE GreenCom and IEEE CPSCom and IEEE SmartData , 2018

    2018

  11. [19]

    Joint task offloading and resource allocation for energy-constrained mobile edge computing,

    H. Jiang, X. Dai, Z. Xiao, and A. Iyengar, “Joint task offloading and resource allocation for energy-constrained mobile edge computing,” IEEE Transactions on Mobile Computing , vol. 22, 13 no. 7, pp. 4000–4015, 2023

    2023

  12. [20]

    Multi-Armed Bandits Learning for Task Offloading in Maritime Edge Intelligence Networks,

    T. Yang, S. Gao, J. Li, M. Qin, X. Sun, R. Zhang, M. Wang, and X. Li, “Multi-Armed Bandits Learning for Task Offloading in Maritime Edge Intelligence Networks,” IEEE Transactions on Vehicular Technology, 2022

    2022

  13. [21]

    Sus- tainable Task Offloading in UA V Networks via Multi-Agent Reinforcement Learning,

    A. Sacco, F. Esposito, G. Marchetto, and P. Montuschi, “Sus- tainable Task Offloading in UA V Networks via Multi-Agent Reinforcement Learning,” IEEE Trans. on Vehi. Tech., 2021

    2021

  14. [22]

    A Self-Learning Strategy for Task Offloading in UA V Networks,

    ——, “A Self-Learning Strategy for Task Offloading in UA V Networks,” IEEE Transactions on Vehicular Technology, 2022

    2022

  15. [23]

    6DOF Virtual Reality Dataset and Performance Evaluation of Mil- limeter Wave vs. Free-Space-Optical Indoor Communications Systems for Lifelike Mobile VR Streaming,

    J. Chakareski, M. Khan, T. Ropitault, and S. Blandino, “6DOF Virtual Reality Dataset and Performance Evaluation of Mil- limeter Wave vs. Free-Space-Optical Indoor Communications Systems for Lifelike Mobile VR Streaming,” in Proc. 54 th Asilomar Conf. on Sig., Sys., and Compute...

    2020

  16. [24]

    MEC-Assisted FoV-Aware and QoE-Driven Adap- tive 360° Video Streaming for Virtual Reality,

    C.-H. Hsu, “MEC-Assisted FoV-Aware and QoE-Driven Adap- tive 360° Video Streaming for Virtual Reality,” in 16th Intern. Conf. on Mobility, Sensing and Networking (MSN) , 2020

    2020

  17. [25]

    A View Synthesis-Based 360° VR Caching System Over MEC-Enabled C-RAN,

    J. Dai, Z. Zhang, S. Mao, and D. Liu, “A View Synthesis-Based 360° VR Caching System Over MEC-Enabled C-RAN,” IEEE Transactions on Circuits and Systems for Video Technology , 2020

    2020

  18. [26]

    Tile-Based Edge Caching for 360° Live Video Streaming,

    P. Maniotis and N. Thomos, “Tile-Based Edge Caching for 360° Live Video Streaming,”IEEE Transactions on Circuits and Systems for Video Technology, 2021

    2021

  19. [27]

    Joint Radio Communication, Caching, and Computing Design for Mobile Virtual Reality Delivery in Fog Radio Access Networks,

    T. Dang and M. Peng, “Joint Radio Communication, Caching, and Computing Design for Mobile Virtual Reality Delivery in Fog Radio Access Networks,” IEEE Journal on Selected Areas in Communications, 2019

    2019

  20. [28]

    UA V-Assisted Edge Computing and Streaming for Wireless Virtual Reality: Analysis, Algorithm Design, and Performance Guarantees,

    L. Zhang and J. Chakareski, “UA V-Assisted Edge Computing and Streaming for Wireless Virtual Reality: Analysis, Algorithm Design, and Performance Guarantees,” IEEE Trans. on Vehicu- lar Technology, 2022

    2022

  21. [29]

    On the Placement of Edge Servers in Mobile Edge Computing,

    H. Liu, S. Wang, H. Huang, and Q. Ye, “On the Placement of Edge Servers in Mobile Edge Computing,” in2023 International Conference on Computing, Networking and Communications (ICNC), 2023

    2023

  22. [30]

    Design of Virtual Reality Edu- cation Platform based on 5G MEC,

    Z. Guo, P. Zhang, and J. Xia, “Design of Virtual Reality Edu- cation Platform based on 5G MEC,” in 2021 20th International Conference on Ubiquitous Computing and Communications , 2021

    2021

  23. [31]

    Edge-Enabled Metaverse: The Convergence of Metaverse and Mobile Edge Computing,

    N. Aung, S. Dhelim, L. Chen, H. Ning, L. Atzori, and T. Kechadi, “Edge-Enabled Metaverse: The Convergence of Metaverse and Mobile Edge Computing,” Tsinghua Science and Technology, 2024

    2024

  24. [32]

    A Real-Time Architecture of 360-Degree Panoramic Video Streaming System,

    M. Han, S.-H. Lee, and S. Ok, “A Real-Time Architecture of 360-Degree Panoramic Video Streaming System,” in 2019 IEEE 2nd International Conference on Knowledge Innovation and Invention (ICKII) , 2019

    2019

  25. [33]

    RealVR: Efficient, Economical, and Quality-of- Experience-Driven VR Video System Based on MPEG OMAF,

    Q. Zhang, J. Wei, S. Wang, S. Ma, and W. Gao, “RealVR: Efficient, Economical, and Quality-of- Experience-Driven VR Video System Based on MPEG OMAF,” IEEE Transactions on Multimedia, 2023

    2023

  26. [34]

    Communication-Constrained Mobile Edge Computing Systems for Wireless Virtual Reality: Scheduling and Tradeoff,

    X. Yang, Z. Chen, K. Li, Y . Sun, N. Liu, W. Xie, and Y . Zhao, “Communication-Constrained Mobile Edge Computing Systems for Wireless Virtual Reality: Scheduling and Tradeoff,” IEEE Access, 2018

    2018

  27. [35]

    Resource Management for Pervasive-Edge-Computing- Assisted Wireless VR Streaming in Industrial Internet of Things,

    P. Lin, Q. Song, D. Wang, F. R. Yu, L. Guo, and V . C. M. Leung, “Resource Management for Pervasive-Edge-Computing- Assisted Wireless VR Streaming in Industrial Internet of Things,” IEEE Transactions on Industrial Informatics , 2021

    2021

  28. [36]

    An Adaptive Wireless Virtual Reality Framework in Future Wireless Networks: A Distributed Learning Approach,

    F. Guo, R. Yu, H. Zhang, H. Ji, V . C. M. Leung, and X. Li, “An Adaptive Wireless Virtual Reality Framework in Future Wireless Networks: A Distributed Learning Approach,” IEEE Transactions on Vehicular Technology, 2020

    2020

  29. [37]

    Data Correlation- Aware Resource Management in Wireless Virtual Reality (VR): An Echo State Transfer Learning Approach,

    M. Chen, W. Saad, C. Yin, and M. Debbah, “Data Correlation- Aware Resource Management in Wireless Virtual Reality (VR): An Echo State Transfer Learning Approach,” IEEE Trans. on Communications, 2019

    2019

  30. [38]

    MAC Scheduling for Multiuser Wireless Virtual Reality in 5G MIMO-OFDM Systems,

    M. Huang and X. Zhang, “MAC Scheduling for Multiuser Wireless Virtual Reality in 5G MIMO-OFDM Systems,” in2018 IEEE International Conference on Communications Workshops (ICC Workshops), 2018

    2018

  31. [39]

    Live 360 ◦ Video Streaming to Heterogeneous Clients in 5G Networks,

    J. Chakareski and M. Khan, “Live 360 ◦ Video Streaming to Heterogeneous Clients in 5G Networks,” IEEE Transactions on Multimedia, 2024

    2024

  32. [40]

    Multipath Cooperative Communications Networks for Augmented and Virtual Reality Transmission,

    X. Ge, L. Pan, Q. Li, G. Mao, and S. Tu, “Multipath Cooperative Communications Networks for Augmented and Virtual Reality Transmission,” IEEE Transactions on Multimedia , 2017

    2017

  33. [41]

    MEC-Assisted Panoramic VR Video Streaming Over Millimeter Wave Mobile Networks,

    Y . Liu, J. Liu, A. Argyriou, and S. Ci, “MEC-Assisted Panoramic VR Video Streaming Over Millimeter Wave Mobile Networks,” IEEE Transactions on Multimedia , 2019

    2019

  34. [42]

    mmWave Network- ing and Edge Computing for Scalable 360 ◦ Video Multi-User Virtual Reality,

    S. Gupta, J. Chakareski, and P. Popovski, “mmWave Network- ing and Edge Computing for Scalable 360 ◦ Video Multi-User Virtual Reality,”IEEE Transactions on Image Processing, 2023

    2023

  35. [43]

    An Edge-Computing Based Arch. for Mobile Aug. Reality,

    J. Ren, Y . He, G. Huang, G. Yu, Y . Cai, and Z. Zhang, “An Edge-Computing Based Arch. for Mobile Aug. Reality,” IEEE Network, 2019

    2019

  36. [44]

    Latency-aware hybrid edge cloud framework for mobile augmented reality applications,

    A. Younis, B. Qiu, and D. Pompili, “Latency-aware hybrid edge cloud framework for mobile augmented reality applications,” in 2020 17th Annual IEEE International Conference on Sensing, Communication, and Networking (SECON) , 2020, pp. 1–9

    2020

  37. [45]

    Computing Offloading Based on Deep Reinforcement Learning For Virtual Reality Scene,

    Y . Song and Y . Shen, “Computing Offloading Based on Deep Reinforcement Learning For Virtual Reality Scene,” in 2023 IEEE International Symposium on Broadband Multimedia Sys- tems and Broadcasting (BMSB) , 2023

    2023

  38. [46]

    Design and Analysis of MEC- and Proactive Caching- Based 360◦ Mobile VR Video Streaming,

    Q. Cheng, H. Shan, W. Zhuang, L. Yu, Z. Zhang, and T. Q. S. Quek, “Design and Analysis of MEC- and Proactive Caching- Based 360◦ Mobile VR Video Streaming,” IEEE Transactions on Multimedia, 2022

    2022

  39. [47]

    ear: An edge-assisted and energy- efficient mobile augmented reality framework,

    N. Didar and M. Brocanelli, “ear: An edge-assisted and energy- efficient mobile augmented reality framework,” IEEE Transac- tions on Mobile Computing, vol. 22, no. 7, pp. 3898–3909, 2023

    2023

  40. [48]

    Leaf + aio: Edge-assisted energy-aware object detection for mobile augmented reality,

    H. Wang, B. Kim, J. Xie, and Z. Han, “Leaf + aio: Edge-assisted energy-aware object detection for mobile augmented reality,” IEEE Transactions on Mobile Computing , vol. 22, no. 10, pp. 5933–5948, 2023

    2023

  41. [49]

    Edge-Computing- Assisted Virtual Reality Computation Offloading: An Empirical Study,

    B. W. Nyamtiga, A. A. Hermawan, Y . F. Luckyarno, T.-W. Kim, D.-Y . Jung, J. S. Kwak, and J.-H. Yun, “Edge-Computing- Assisted Virtual Reality Computation Offloading: An Empirical Study,” IEEE Access, 2022

    2022

  42. [50]

    Energy-Optimal Mobile Cloud Computing under Stochastic Wireless Channel,

    W. Zhang, Y . Wen, K. Guan, D. Kilper, H. Luo, and D. O. Wu, “Energy-Optimal Mobile Cloud Computing under Stochastic Wireless Channel,” IEEE Trans. on Wireless Comm. , 2013

    2013

  43. [51]

    T. D. Burd and R. W. Brodersen, Processor Design for Portable Systems, 1996

    1996

  44. [52]

    CoPoMo: a context-aware power consumption model for LTE user equip- ment,

    B. Dusza, C. Ide, L. Cheng, and C. Wietfeld, “CoPoMo: a context-aware power consumption model for LTE user equip- ment,” Trans. on Emerging Telecomm. Tech., 2013

    2013

  45. [53]

    Zappone and E

    A. Zappone and E. Jorswieck, Energy Efficiency in Wireless Networks via Fractional Programming Theory , 2015

    2015

  46. [54]

    A Variegated Look at 5G in the Wild: Performance, Power, and QoE Implications,

    A. Narayanan, X. Zhang, R. Zhu, A. Hassan, S. Jin, X. Zhu, X. Zhang, D. Rybkin, Z. Yang, Z. M. Mao, F. Qian, and Z.-L. Zhang, “A Variegated Look at 5G in the Wild: Performance, Power, and QoE Implications,” in Proceedings of ACM SIG- COMM, 2021

    2021

  47. [55]

    Mastering Complex Control in MOBA Games with Deep Reinforcement Learning,

    D. Ye, Z. Liu, M. Sun, B. Shi, P. Zhao, H. Wu, H. Yu, S. Yang, X. Wu, Q. Guo, Q. Chen, Y . Yin, H. Zhang, T. Shi, L. Wang, Q. Fu, W. Yang, and L. Huang, “Mastering Complex Control in MOBA Games with Deep Reinforcement Learning,” in The 34 AAAI Conf. on AI , 2020

    2020

  48. [56]

    Buffer Awareness Neural Adaptive Video Streaming for Avoid- ing Extra Buffer Consumption,

    C. W. L. S. Tianchi Huang ; Chao Zhou ; Rui-Xiao Zhang, “Buffer Awareness Neural Adaptive Video Streaming for Avoid- ing Extra Buffer Consumption,” in IEEE INFOCOM Conf. , 2023

    2023

  49. [57]

    Proximal Policy Optimization Algorithms,

    J. Schulman, F. Wolski, P. Dhariwal, A. Radford, and O. Klimov, “Proximal Policy Optimization Algorithms,” arXiv:1707.06347, 2017

    2017 arXiv

  50. [58]

    The surprising effectiveness of ppo in cooperative 14 multi-agent games,

    C. Yu, A. Velu, E. Vinitsky, J. Gao, Y . Wang, A. Bayen, and Y . Wu, “The surprising effectiveness of ppo in cooperative 14 multi-agent games,” in Proceedings of the 36th International Conference on Neural Information Processing Systems , ser. NIPS ’22. Red Hook, NY , USA: Cur...

    2022

  51. [59]

    Parameter sharing for heterogeneous agents in multi-agent reinforcement learning,

    J. K. Terry, N. Grammel, S. Son, and B. Black, “Parameter sharing for heterogeneous agents in multi-agent reinforcement learning,” CoRR, vol. abs/2005.13625, 2020. [Online]. Available: https://arxiv.org/abs/2005.13625

    2005 arXiv

  52. [60]

    Neural Contextual Bandits with UCB-Based Exploration,

    D. Zhou, L. Li, and Q. Gu, “Neural Contextual Bandits with UCB-Based Exploration,” in Proceedings of 37th ICML , 2020

    2020

  53. [61]

    Full UHD 360-Degree Video Dataset and Modeling of Rate- Distortion Characteristics and Head Movement Navigation,

    J. Chakareski, R. Aksu, V . Swaminathan, and M. Zink, “Full UHD 360-Degree Video Dataset and Modeling of Rate- Distortion Characteristics and Head Movement Navigation,” in ACM Multimedia Sys. Conf. , 2021

    2021

  54. [62]

    Image quality assessment: from error visibility to structural similarity,

    Z. Wang, A. Bovik, H. Sheikh, and E. Simoncelli, “Image quality assessment: from error visibility to structural similarity,” IEEE Transactions on Image Processing, vol. 13, no. 4, pp. 600– 612, 2004

    2004

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.