LYRA: Label-Free Structural Synchronization and Resource Allocation for UAV Edge Networks

Alireza Furutanpey; Feng He; Jiangchuan Liu; Jiannong Cao; Paolo Bellavista; Schahram Dustdar; Yu Qiu

REVIEW 4 major objections 5 minor 43 references

LYRA: Label-Free Structural Synchronization and Resource Allocation for UAV Edge Networks

T0 review · 4 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read LYRA, a label-free scheduling framework for UAV edge models, claims up to 33.3% lower average risk backlog than baselines while respecting long-term energy budgets.

desk verdict A solid framework paper with a valid OSDR bound and real recovered-accuracy evidence, but the headline backlog numbers depend on uncalibrated parameters in the queue model. read the letter →

arxiv 2608.07392 v1 pith:KJ3GYN7B submitted 2026-08-07 cs.NI

Feng He , Alireza Furutanpey , Paolo Bellavista , Yu Qiu , Jiangchuan Liu , Jiannong Cao , Schahram Dustdar This is my paper · ORCID

classification cs.NI

keywords UAVedgenetworksmodelupdateschedulinglabel-freesemanticproxystructuralsynchronizationLyapunovoptimizationdeepreinforcementlearninghierarchicalvisionmodelsriskbacklog

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

LYRA is a joint model-update scheduling and resource-allocation framework for UAV edge systems that run hierarchical vision models under low-level environmental corruption such as fog or sudden illumination changes. Its central claim is that the Online Semantic Disagreement Rate (OSDR), defined as the disagreement between the deployed edge model and a quasi-static server-hosted oracle on unlabelled probe batches, can replace real-time ground-truth labels for deciding when a model must be updated. The paper further claims that partial front-to-back structural synchronization, regulated by a sensitivity threshold, can be scheduled and resourced optimally, and that a Lyapunov-guided deep reinforcement learning algorithm can make the resulting hybrid continuous-discrete decision tractable. On trace-driven simulations, LYRA is reported to reduce average semantic risk backlog by up to 33.3% over the best baseline while satisfying long-term energy budgets, and to cut communication overhead by 80.5% compared with back-to-front updates. If correct, this means drones can maintain reliable inference under environmental stress without ground-truth labels and within tight bandwidth and energy constraints.

What carries the argument

Three coupled mechanisms carry the argument. (1) OSDR, the empirical disagreement rate between the deployed edge model and a quasi-static server oracle over $m=32$ unlabelled compressed frames, acts as a label-free proxy for true error; Theorem 3.1 bounds $|e_t-\hat{d}_t|$ by $\epsilon_t+\sqrt{\log(2/\delta)/2m}$. (2) SASS models a front-to-back structural update: a sensitivity threshold $\tau_t\in[0,1]$ determines cumulative patch size $L(\tau_t)=L_{\max}e^{-K\tau_t}$ and recovery efficiency $\Phi(\tau_t)=1-\tau_t^\gamma$; the semantic risk queue drains as $\mu_t=\alpha_t\Phi(\tau_t)Q_{\mathrm{sem}}(t)$, a multiplicative proportional catch-up model. (3) The Lyapunov-guided DRL collapses $(\alpha_t,\tau_t,b_t,p_t)$ into one discrete action $a_t\in\{0,\dots,K\}$ through a structural mapping $M(a_t)$, then allocates bandwidth by bisection on the strictly convex physical penalty $J_{\mathrm{phys}}(b_t\mid a_t)$, with virtual queues $Q_{\mathrm{sem}}(t)$ and $Z(t)$ converting long-term constraints into per-slot reward penalties.

What would settle it

Run the actual SASS protocol on a drone-in-the-loop testbed: record OSDR and true accuracy on a corrupted stream, apply each structural update depth, and compare measured accuracy recovery with $\Phi(\tau_t)Q_{\mathrm{sem}}(t)$; a mismatch would show that the multiplicative drain model in Eq. (3) is wrong, making the 33.3% backlog reduction a simulation artifact.

Watch

Extended reading notes

Core claim

The paper claims that semantic fidelity of a UAV edge model under environmental corruption can be maintained by continuously tracking the disagreement between the deployed edge model and a quasi-static server oracle on unlabelled probe batches, and by using that disagreement signal to trigger partial updates that synchronize shallow layers first. The update decision is cast as minimizing a long-term cost subject to queue stability and an energy budget; the paper derives a per-slot Lyapunov drift-plus-penalty problem and solves it with a hierarchical PPO agent whose discrete action selects a structural synchronization depth, while a closed-form execution layer allocates bandwidth and transmit power. The central discovery is that this combination yields higher semantic recovery per unit of communication and more precise update triggering than periodic, myopic, or full-model baselines, with the reported 33.3% reduction in average risk backlog.

Load-bearing premise

The whole queue model assumes that a partial update of a given depth removes a fixed fraction of the accumulated semantic damage, and that this fraction follows a power-law curve; the paper never measures this recovery curve directly, and if real model recovery behaves differently, the reported backlog reductions would not survive contact with real hardware.

Editorial extensions

If this is right

UAVs can maintain model fidelity mid-flight without expert labels, as long as a quasi-static oracle is available at the ground server.
Front-to-back partial updates are more communication-efficient than full-model or back-to-front updates under low-level corruption; the paper reports an 80.5% overhead reduction.
Lyapunov-guided reward shaping keeps long-term energy debt bounded in simulation, unlike heuristic rewards, so the learned policy can respect hardware budgets over long horizons.
The discrete structural action space avoids the convergence failure of hybrid-action DRL; the Mixed-Action DDPG baseline in the original continuous-discrete space does not converge.
OSDR-based triggering generalizes across unseen drift temporal patterns (step and sinusoidal) without retuning, whereas fixed thresholds fail.

Reading between the lines

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

The paper does not demonstrate the OSDR proxy outside ResNet-18 on CIFAR-10-C; in principle the disagreement-triggered scheduling logic applies to any hierarchical model with block structure, but that transfer is an extension, not a proven claim.
The load-bearing simulation assumption is the multiplicative recovery model: the paper asserts that a partial update drains the semantic risk queue in proportion to its own size and to a power law of synchronization depth, without measuring that recovery curve or varying the shape parameter.
Because the oracle is only quasi-static and the probe consumes uplink bandwidth, there is an implicit cost and staleness trade-off; a testable extension is to make the oracle's refresh frequency a decision variable rather than a fixed schedule.
The Lyapunov optimality gap is stated as $O(1/V)$ but not numerically verified; a sensitivity study on $V$, $\gamma$, and the sparsity coefficient $K$ would show whether the reported gains persist across operating points.

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.

Desk Editor's Note

read the letter

LYRA is a real attempt at an unsolved problem: when to trigger partial model updates on a bandwidth-limited UAV and how deep to sync, without real-time labels. The paper does several things well. The OSDR proxy is well-motivated and Theorem 3.1 is a valid finite-sample bound relating edge-oracle disagreement to true error; that part I checked and it holds. The hierarchical action-space reduction (map trigger and depth to a discrete structural action, then solve bandwidth/power in closed form) is a sound engineering idea, and the ablation against MA-DDPG and HR-PPO shows the design choices matter. The recovered-accuracy curves in Fig. 4(a) provide non-circular evidence that the policy actually maintains real model accuracy under drift.

The soft spots are real but narrower than the stress-test note claims. The biggest problem is Eq. (3): the service rate mu_t = alpha_t (1 - tau^gamma) Q_sem(t) is the only mechanism reducing the risk backlog, and gamma is never assigned a value or subjected to sensitivity analysis. K in Eq. (4) is likewise unspecified. Because the headline 33.3% backlog reduction and the SRE/UTP metrics are computed from this assumed queue dynamics, those numbers are partly circular. The stress-test is right that if gamma is off, the learned policy could be chasing the wrong objective. But the stress-test overreaches when it says the results may be artifacts: the simulator does apply real partial updates to ResNet-18 on CIFAR-10-C, and Fig. 4(a) shows LYRA sustains real accuracy better than baselines. So the central claim is not purely a queue-model artifact, but the quantification of how much better is contaminated by uncalibrated parameters.

Other issues: Proposition 3.1 is asserted without proof, Theorem 5.1's i.i.d. assumption sits awkwardly with the non-stationary evaluation and its proof is omitted 'for brevity,' and there is no code or data release. The paper would be substantially stronger with a sensitivity analysis over gamma and K and a direct comparison between the assumed Phi(tau) and measured accuracy recovery per depth.

Who is this for? Researchers working on UAV-MEC model maintenance and semantic-aware scheduling. It deserves a serious referee, but the referee should require calibration of the free parameters and ideally code before acceptance.

Referee Report

4 major / 5 minor

Summary. This paper proposes LYRA, a joint model-update scheduling and resource-allocation framework for UAV edge systems running hierarchical vision models under low-level environmental corruption. LYRA introduces SASS, which selects front-to-back partial synchronization depth via a discrete structural action space, and OSDR, a label-free disagreement-based proxy for semantic degradation. A Lyapunov drift-plus-penalty formulation converts the long-term energy budget into a virtual queue, and a PPO agent with a closed-form bandwidth/power execution layer optimizes the per-slot penalty. Experiments on CIFAR-10-C with CRAWDAD mobility traces compare LYRA against periodic, myopic, threshold, and hybrid-action baselines, reporting up to 33.3% lower average risk backlog and 80.5% lower communication overhead than back-to-front updates.

Significance. If the empirical claims hold, LYRA would be a useful step toward label-free, bandwidth-aware model maintenance for UAV edge inference. The paper's strengths include a formal finite-sample OSDR bound (Theorem 3.1), a clean Lyapunov decomposition of the long-term energy constraint, and a hierarchical discrete-action design that avoids hybrid-action DRL instability. The structural ablation against B2F/HO/FM/RB is a good experimental idea. However, the numerical results currently rest on an uncalibrated multiplicative recovery model in Eq. (3), so the quantitative claims are not yet established for real model fidelity; with calibration or measured recovery curves, the contribution would be significant.

major comments (4)

[Section IV-B, Eqs. (3)-(4)] The semantic drain model mu_t = alpha_t * (1 - tau_t^gamma) * Q_sem(t) and the patch-size model L_t = L_max * e^{-K*tau_t} are asserted rather than calibrated: gamma and K never receive numerical values, and no sensitivity analysis is reported. Because Q_sem(t+1) in Eq. (2), the reward in Eq. (14), and the reported backlog and SRE numbers all propagate through these two equations, the headline 33.3% backlog reduction is currently a property of an unvalidated model, not a measured property of ResNet-18 under CIFAR-10-C corruption. Please either fit Phi(tau) to measured layer-wise accuracy recovery, report gamma and K with confidence intervals and sweep them, or explicitly downgrade the quantitative claims.
[Section VII-B, SRE definition] The Semantic Recovery Efficiency is defined as SRE = (1/|T+|) * sum_{t in T+} Phi(a_t)/L(a_t), using exactly the assumed recovery function Phi from Eq. (3). The SRE comparisons in Table I and Fig. 4(b) are therefore not independent evidence for the recovery model; they restate the assumption. Please compute SRE from measured accuracy deltas after actual partial updates, or explicitly label the current SRE as a model-based proxy and place the validation burden on the recovered-accuracy results.
[Section VII-A/C, Fig. 4(a)] The paper does not specify how an update action a_t is mapped to the reported 'Recovered Accuracy.' If the simulator literally evolves Eq. (3) or uses Phi to reconstruct accuracy, then Fig. 4(a) cannot serve as external validation of LYRA's real-world fidelity, and the Q_sem-based circularity is only partially mitigated. Please describe the accuracy simulator: either actual ResNet-18 forward passes after applying partial parameter restorations under CIFAR-10-C, or a separate analytical curve. If it is the latter, state this clearly and adjust the claims accordingly.
[Section III-C, Proposition 3.1] The regret bound lim_{T->infinity} R(T) <= W_1 * (epsilon_max + sqrt(log(2/delta)/2m)) is asserted without proof, and neither W_1 nor the cost functional Cost(.) is defined precisely enough to check the Lipschitz condition. Since this proposition is invoked to justify the practical gap to a label-aware policy, please supply a proof or state the precise assumptions under which it holds; otherwise present it as a conjecture.

minor comments (5)

[Theorem 5.1] The proof of Theorem 5.1 is omitted as 'standard'; given the paper's reliance on Lyapunov arguments, a concise proof or a direct reference to the exact theorem in [5] would improve verifiability.
[Algorithm 1, line 14] The description 'Mask the action by setting E_total^* = infinity' is not an action mask as normally understood; please clarify how the infinite penalty enters the reward and whether such actions are also excluded from the PPO update.
[Abstract and Section VII-A] The abstract says 'real traffic traces,' but reference [43] is a GPS mobility trace archive; the communication channel is simulated from this mobility. Please rephrase to avoid implying that measured wireless traffic was used.
[Lemma 6.1] The constant c in Lemma 6.1 is said to aggregate bandwidth-energy coefficients induced by optimal physical execution, but no derivation from Eq. (13) is given; as written, the Lipschitz bound is a sketch rather than a complete proof.
[Section VII-C, UTP metric] UTP is undefined for the NU policy, which never triggers updates, yet Figs. 5(b) and 6(c) appear to include NU UTP values; please specify the convention used for policies with no triggered slots.

Circularity Check

2 steps flagged · score 4.0 of 10

Partial circularity: the headline risk backlog is the algorithm's own Lyapunov state and the SRE metric is defined from the assumed recovery function Phi; independent recovered-accuracy results keep the central claim from being fully circular.

self definitional [Section IV.B Eqs. (2)-(3), Section VI.C Eq. (14), Section VII.B]
"Qsem(t+1) = max[0, Qsem(t) - mu_t(alpha_t, tau_t) + lambda_t] ... mu_t(alpha_t, tau_t) = alpha_t * Phi(tau_t) * Qsem(t) ... r_t(s_t,a_t) = Q_sem(t)*mu^a_t - (V*omega_e + Z(t))*E*_total(a_t) - V*omega_b*b*_t(a_t)"

The risk backlog Q_sem is defined by the recurrence in Eqs. (2)-(3), in which the only drain is mu = alpha*Phi*Q_sem. The reward in Eq. (14) includes +Q_sem*mu = alpha*Phi*Q_sem^2, i.e., it rewards draining this same Q_sem variable, and the headline metric 'average risk backlog Q_sem' (Sec. VII.B) is that same state variable. Reporting a 33.3% reduction in this backlog is therefore measuring the algorithm against the surrogate that defines both its dynamics and its reward; the real external check is only the recovered-accuracy curve in Fig. 4(a).
fitted input called prediction [Section IV.B Eq. (3), Section IV.C Eq. (4), Section VII.B metric definition]
"SRE is defined as SRE = 1/|T_+| * sum_{t in T_+} Phi(a_t)/L(a_t) ... Phi(tau_t) = 1 - tau_t^gamma ... L_t(tau_t) = L_max * e^{-K*tau_t}"

The Semantic Recovery Efficiency metric is computed directly from the assumed recovery function Phi (Eq. 3) and the fitted patch-size model L (Eq. 4), not from measured accuracy recovery. Thus the claimed 'superior semantic recovery efficiency' is an algebraic consequence of the definitions: any algorithm choosing small tau (large Phi, large L) scores high by construction. Since gamma and K are never calibrated and no sensitivity analysis is reported, the SRE comparison cannot validate the assumed Phi; only the separately reported recovered accuracy (Fig. 4a) provides external grounding.

full rationale

The derivation chain OSDR-to-SASS-to-Lyapunov-to-PPO is not itself circular: OSDR is a standard disagreement-rate bound, P2 is a Lyapunov drift-plus-penalty reduction, and the discrete structural mapping has a Lipschitz optimality gap. No load-bearing self-citation appears; the only cited authority for the omitted Lyapunov proof is an external paper [5]. The circularity burden is in the evaluation layer: the primary risk-backlog metric is the same variable the Lyapunov reward is built from, and SRE is defined through the paper's own assumed Phi rather than measured fidelity. This is mitigated by the independent recovered-accuracy result in Fig. 4(a), by the fact that all baselines are evaluated under the same simulated dynamics, and by the OSDR generalization bound. The unset gamma and absent sensitivity analysis are model-validity risks, not circularity. Score 4 reflects partial self-referential metrics while acknowledging independent content.

Assumptions & free parameters 4 free parameters · 7 assumptions · 0 invented entities

The central claim rests on two unvalidated functional forms (multiplicative risk drain, exponential patch size) with unreported coefficients gamma and K, and on an oracle with bounded error. No code or data is provided, and the theory sections omit some proofs.

free parameters (4)

gamma (recovery shape parameter)
Eq. (3) defines Phi(tau)=1-tau^gamma; no value is reported or swept in experiments, leaving the recovery model underdetermined.
K (patch-size sparsity coefficient) = not reported (exponential fit to Fig. 1(b))
Eq. (4) models cumulative patch size as Lmax*exp(-K*tau); K is fitted to empirical data in Section III.B but its value and sensitivity are not given.
structural action levels {tau_1,...,tau_K} = not reported
The discrete sensitivity levels for the action mapping M(a_t) in Eq. (11) are predefined but their values are not stated, so the action space is not fully specified.
Lipschitz constant W1 (Prop. 3.1) = not estimated
Proposition 3.1 asserts a regret bound using an unspecified Lipschitz constant W1; the bound is qualitative without a value.

assumptions (7)

standard math Data Processing Inequality: I(Y; Y_hat) <= I(Y; H_k) for hierarchical features
Used in Section III.B to argue shallow features limit downstream semantic information, motivating front-to-back updates.
ad hoc to paper Multiplicative recovery model (Eq. 3)
The service rate mu_t = alpha*Phi*Q_sem is asserted without empirical validation; it is central to queue dynamics and reward.
ad hoc to paper Exponential cumulative patch-size model (Eq. 4)
The exponential L(tau) is an approximation motivated by Fig. 1(b); its fitted coefficient K is not reported.
domain assumption Bounded oracle error Pr(g_t != Y) <= epsilon_t
Theorem 3.1 and Proposition 3.1 require the server oracle to have bounded error; the paper assumes this but does not specify how the oracle is maintained.
domain assumption i.i.d. drift and channel with bounded support (Theorem 5.1)
The Lyapunov bound assumes i.i.d. lambda_t and h_t, while the simulation uses non-stationary Markov/step/sinusoidal drift; the applicability to the reported settings is not formally established.
domain assumption Truncated semantic backlog Q_sem <= Q_max_sem
Theorem 5.1 relies on a bounded state space; the paper does not define Q_max_sem or justify the truncation physically.
domain assumption Lipschitz continuity of patch size and recovery functions
Lemma 6.1 assumes finite constants C_L and C_Phi; this is plausible but not proven from the physical model.

how reviews work

0 comments

Cite this review

Pith. "Pith review of LYRA: Label-Free Structural Synchronization and Resource Allocation for UAV Edge Networks." pith.science (2026). https://pith.science/paper/KJ3GYN7B

@misc{pith2026260807392,
  author       = {Pith},
  title        = {Pith review of: LYRA: Label-Free Structural Synchronization and Resource Allocation for UAV Edge Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KJ3GYN7B}},
  note         = {Machine review of arXiv:2608.07392}
}

read the original abstract

While deploying hierarchical vision models to process mission-critical tasks, UAV edge systems must adaptively update the models to sustain inference reliability under low-level environmental corruption. However, existing work has overlooked the optimal timing for model updates, the impracticality of relying on real-time expert labels, and the significant bandwidth and energy constraints of UAVs. This paper proposes a joint model update scheduling and resource allocation framework, aiming to maximize long-term semantic fidelity and resource efficiency of UAV edge intelligence systems. To address the challenge of label-free semantic evaluation, we formulate the Online Semantic Disagreement Rate (OSDR) as a proxy for timely update triggering, thereby enabling fine-grained Sensitivity-Aware Structural Synchronization (SASS). Furthermore, to overcome the curse of dimensionality in hybrid action spaces and effectively bound long-term energy budgets, we propose a Lyapunov-guided discrete reinforcement learning algorithm that performs action space dimensionality reduction and transforms constraints into virtual queue stability problems. The reported experimental results, based on real traffic traces, demonstrate that the proposed framework consistently outperforms representative baselines in semantic recovery efficiency and update triggering precision, by satisfying long-term energy budget and by reducing average risk backlog by up to 33.3\% in the dynamic environmental corruption scenario.

Figures

Figures reproduced from arXiv: 2608.07392 by the authors.

**Figure 2.** The overall architecture of the proposed LYRA frame [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 3.** Impact of (a) hierarchical action decoupling against [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗

Figures from the paper (2 more)

**Figure 4.** Figure 4: System performance comparison under dynamic environmental corruptions. [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]

**Figure 6.** Figure 6: Performance stability comparison under varying drift [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 41 canonical work pages

[1]

Pyramid: Enabling hierarchical neural networks with edge computing,

Q. He, Z. Dong, F. Chenet al., “Pyramid: Enabling hierarchical neural networks with edge computing,” inACM WWW’22, New York, NY , USA, 2022, p. 1860–1870

work page 2022
[2]

A distributed hierarchical deep computation model for federated learning in edge computing,

H. Zheng, M. Gao, Z. Chenet al., “A distributed hierarchical deep computation model for federated learning in edge computing,”IEEE Transactions on Industrial Informatics, vol. 17, no. 12, pp. 7946–7956, 2021

work page 2021
[3]

Vision-language models for edge networks: A comprehensive survey,

A. Sharshar, L. U. Khan, W. Ullahet al., “Vision-language models for edge networks: A comprehensive survey,”IEEE Internet Things Journal, vol. 12, no. 16, pp. 32 701–32 724, 2025

work page 2025
[4]

Tri-ring: Asynchronous service provisioning with online learning in edge cloud networks,

X. Ren, Q. Li, H. Duet al., “Tri-ring: Asynchronous service provisioning with online learning in edge cloud networks,” inIEEE INFOCOM, 2025, pp. 1–10

work page 2025
[5]

Freshness-aware inference services in edge computing via offloading or local processing,

X. Ai and W. Liang, “Freshness-aware inference services in edge computing via offloading or local processing,” inIEEE LCN. IEEE, 2025, pp. 1–10

work page 2025
[6]

Efficient coordination of federated learning and inference offloading at the edge: A proactive optimization paradigm,

K. Luo, K. Zhao, T. Ouyanget al., “Efficient coordination of federated learning and inference offloading at the edge: A proactive optimization paradigm,”IEEE Transactions Mobile Computing, vol. 24, no. 1, pp. 407–421, 2025

work page 2025
[7]

Efficient online dnn inference with continuous learning in edge computing,

Y . Zeng, R. Zhou, L. Jiaoet al., “Efficient online dnn inference with continuous learning in edge computing,” inIEEE/ACM IWQoS, 2024, pp. 1–10

work page 2024
[8]

Joint optimization of model retraining and inference services in dt-assisted edge computing,

X. Ai, W. Liang, and C. Liu, “Joint optimization of model retraining and inference services in dt-assisted edge computing,”IEEE/ACM Transac- tions Networking, vol. 34, pp. 1804–1819, 2025

work page 2025

Show all 43 references

[9]

Digital twin-enabled service provisioning in edge computing via continual learning,

J. Li, S. Guo, W. Lianget al., “Digital twin-enabled service provisioning in edge computing via continual learning,”IEEE Transactions Mobile Computing, vol. 23, no. 6, pp. 7335–7350, 2024

2024
[10]

Adaptive on-device model update for responsive video analytics in adverse environments,

Y . Kong, P. Yang, and Y . Cheng, “Adaptive on-device model update for responsive video analytics in adverse environments,”IEEE Transactions on Circuits and Systems for Video Technology, vol. 35, no. 1, pp. 857– 873, 2025

2025
[11]

Joint optimization of model deployment for freshness-sensitive task assignment in edge intelligence,

H. Liu, S. Liu, S. Longet al., “Joint optimization of model deployment for freshness-sensitive task assignment in edge intelligence,” inIEEE INFOCOM, 2024, pp. 1751–1760

2024
[12]

Federated learning while providing model as a service: Joint training and inference optimization,

P. Han, S. Wang, Y . Jiaoet al., “Federated learning while providing model as a service: Joint training and inference optimization,” 2023. [Online]. Available: https://arxiv.org/abs/2312.12863

2023 arXiv
[13]

Inference service fidelity maximization in dt-assisted edge computing,

J. Li, J. Wang, W. Lianget al., “Inference service fidelity maximization in dt-assisted edge computing,”IEEE Transactions Mobile Computing, 2025

2025
[14]

Aoi-aware inference services in edge computing via digital twin network slicing,

Y . Zhang, W. Liang, Z. Xuet al., “Aoi-aware inference services in edge computing via digital twin network slicing,”IEEE Transactions Services Computing, vol. 17, no. 6, pp. 3154–3170, 2024

2024
[15]

Fidelity-aware inference services in dt-assisted edge computing via service model retraining,

X. Ai, W. Liang, Y . Zhanget al., “Fidelity-aware inference services in dt-assisted edge computing via service model retraining,”IEEE Transactions Services Computing, 2025

2025
[16]

Online resource allocation for edge intelligence with colocated model retraining and inference,

H. Cai, Z. Zhou, and Q. Huang, “Online resource allocation for edge intelligence with colocated model retraining and inference,” 2024. [Online]. Available: https://arxiv.org/abs/2405.16029

2024 arXiv
[17]

Digital twin-based task-driven resource management in intelligent uav swarms,

T. Li, S. Leng, X. Liaoet al., “Digital twin-based task-driven resource management in intelligent uav swarms,”IEEE Transactions on Intelli- gent Transportation Systems, vol. 26, no. 4, pp. 5467–5480, 2025

2025
[18]

Energy-efficient and accuracy-aware dnn inference with iot device-edge collaboration,

W. Jiang, H. Han, D. Fenget al., “Energy-efficient and accuracy-aware dnn inference with iot device-edge collaboration,”IEEE Transactions on Services Computing, 2025

2025
[19]

Cost-effective retraining of ma- chine learning models,

A. Mahadevan and M. Mathioudakis, “Cost-effective retraining of ma- chine learning models,”arXiv preprint arXiv:2310.04216, 2023

2023 arXiv
[20]

Age of information under periodic updating: Time-dependent statistical characteristics,

T. Zhang, Z. Chen, A. Liuet al., “Age of information under periodic updating: Time-dependent statistical characteristics,”IEEE Transactions on Communications, 2026

2026
[21]

Aoi minimization in status update control with energy harvesting sensors,

M. Hatami, M. Leinonen, and M. Codreanu, “Aoi minimization in status update control with energy harvesting sensors,”IEEE Transactions on Communications, vol. 69, no. 12, pp. 8335–8351, 2021

2021
[22]

Aoi minimization for wsn data col- lection with periodic updating scheme,

G. Zhang, C. Shen, Q. Shiet al., “Aoi minimization for wsn data col- lection with periodic updating scheme,”IEEE Transactions on Wireless Communications, vol. 22, no. 1, pp. 32–46, 2022

2022
[23]

Efficient information updates in compute- first networking via reinforcement learning with joint aoi and voi,

J. Qi, C. Liu, C. Xuet al., “Efficient information updates in compute- first networking via reinforcement learning with joint aoi and voi,”IEEE Internet of Things Journal, 2026

2026
[24]

How important is periodic model update in recommender system?

H. Lee, S. Yoo, D. Leeet al., “How important is periodic model update in recommender system?” inProceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval, 2023, pp. 2661–2668

2023
[25]

From offline to periodic adaptation for pose-based shoplifting detection in real-world retail security,

S. Yao, N. Rashvand, A. D. Pazhoet al., “From offline to periodic adaptation for pose-based shoplifting detection in real-world retail security,”IEEE Internet of Things Journal, 2026

2026
[26]

Edgeta: Neuron-grained scaling of foundation models in edge-side retraining,

Q. Zhang, R. Han, C. H. Liuet al., “Edgeta: Neuron-grained scaling of foundation models in edge-side retraining,”IEEE Transactions on Mobile Computing, vol. 24, no. 4, pp. 2690–2707, 2024

2024
[27]

Fedquad: Adaptive layer-wise lora de- ployment and activation quantization for federated fine-tuning,

J. Liu, R. Li, H. Xuet al., “Fedquad: Adaptive layer-wise lora de- ployment and activation quantization for federated fine-tuning,”IEEE Transactions on Mobile Computing, 2025

2025
[28]

Semantic utility loss of information for energy efficient semantic status update communications,

L. Xu, J. Jiao, T. Yanget al., “Semantic utility loss of information for energy efficient semantic status update communications,”IEEE Transactions on Cognitive Communications and Networking, vol. 11, no. 1, pp. 59–74, 2024

2024
[29]

Semantic data sourcing for 6g edge intelligence,

K. Huang, Q. Lan, Z. Liuet al., “Semantic data sourcing for 6g edge intelligence,”IEEE Communications Magazine, vol. 61, no. 12, pp. 70– 76, 2023

2023
[30]

Robust semantic communications with masked vq-vae enabled codebook,

Q. Hu, G. Zhang, Z. Qinet al., “Robust semantic communications with masked vq-vae enabled codebook,”IEEE Transactions on Wireless Communications, vol. 22, no. 12, pp. 8707–8722, 2023

2023
[31]

Wireless resource management in in- telligent semantic communication networks,

L. Xia, Y . Sun, X. Liet al., “Wireless resource management in in- telligent semantic communication networks,” inIEEE INFOCOM 2022- IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS). IEEE, 2022, pp. 1–6

2022
[32]

Performance optimization of energy efficient semantic communications over wireless networks,

Z. Yang, M. Chen, Z. Zhanget al., “Performance optimization of energy efficient semantic communications over wireless networks,” in2022 IEEE 96th V ehicular Technology Conference (VTC2022-Fall). IEEE, 2022, pp. 1–5

2022
[33]

Task-oriented multi-user semantic communications,

H. Xie, Z. Qin, X. Taoet al., “Task-oriented multi-user semantic communications,”IEEE Journal on Selected Areas in Communications, vol. 40, no. 9, pp. 2584–2597, 2022

2022
[34]

Drl-based resource allocation and trajectory planning for noma-enabled multi-uav collaborative caching 6g network,

P. Qin, Y . Fu, J. Zhanget al., “Drl-based resource allocation and trajectory planning for noma-enabled multi-uav collaborative caching 6g network,”IEEE Transactions on V ehicular Technology, vol. 73, no. 6, pp. 8750–8764, 2024

2024
[35]

Joint trajectory optimization and resource allocation in uav-mec systems: A lyapunov-assisted drl ap- proach,

Y . Chen, Y . Yang, Y . Wuet al., “Joint trajectory optimization and resource allocation in uav-mec systems: A lyapunov-assisted drl ap- proach,”IEEE Transactions on Services Computing, 2025

2025
[36]

Joint optimization of trajectory control, resource allocation, and user association based on drl for multi-fixed- wing uav networks,

B. Yin, X. Fang, and X. Wang, “Joint optimization of trajectory control, resource allocation, and user association based on drl for multi-fixed- wing uav networks,”IEEE Transactions on Wireless Communications, vol. 23, no. 10, pp. 13 330–13 343, 2024

2024
[37]

Drl-based trajectory optimization and computation-aware resource allocation for uav-assisted edge computing networks,

X. Wu, L. Liang, W. Wenet al., “Drl-based trajectory optimization and computation-aware resource allocation for uav-assisted edge computing networks,”IEEE Internet of Things Journal, 2025

2025
[38]

Trajectory design and resource alloca- tion for multi-uav networks: Deep reinforcement learning approaches,

Z. Chang, H. Deng, L. Youet al., “Trajectory design and resource alloca- tion for multi-uav networks: Deep reinforcement learning approaches,” IEEE Transactions on Network Science and Engineering, vol. 10, no. 5, pp. 2940–2951, 2022

2022
[39]

Deep reinforcement learning based resource allocation and trajectory planning in integrated sensing and communications uav network,

Y . Qin, Z. Zhang, X. Liet al., “Deep reinforcement learning based resource allocation and trajectory planning in integrated sensing and communications uav network,”IEEE Transactions on Wireless Commu- nications, vol. 22, no. 11, pp. 8158–8169, 2023

2023
[40]

Deep reinforcement learning based trajec- tory design and resource allocation for uav-assisted communications,

C. Zhang, Z. Li, C. Heet al., “Deep reinforcement learning based trajec- tory design and resource allocation for uav-assisted communications,” IEEE Communications Letters, vol. 27, no. 9, pp. 2398–2402, 2023

2023
[41]

Joint uav placement optimization, resource allocation, and computation offloading for thz band: A drl approach,

H. Wang, H. Zhang, X. Liuet al., “Joint uav placement optimization, resource allocation, and computation offloading for thz band: A drl approach,”IEEE Transactions on Wireless Communications, vol. 22, no. 7, pp. 4890–4900, 2022

2022
[42]

A novel lyapunov based dynamic resource allocation for uavs-assisted edge computing,

J. Lin, L. Huang, H. Zhang, X. Yang, and P. Zhao, “A novel lyapunov based dynamic resource allocation for uavs-assisted edge computing,” Computer Networks, vol. 205, p. 108710, 2022

2022
[43]

Crawdad data set epfl/mobility (v. 2009-02-24),

M. Piorkowski, N. Sarafijanovic-Djukic, and M. Grossglauser, “Crawdad data set epfl/mobility (v. 2009-02-24),” Feb. 2009. [Online]. Available: https://crawdad.org/epfl/mobility/20090224

2009

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.

[1] [1]

Pyramid: Enabling hierarchical neural networks with edge computing,

Q. He, Z. Dong, F. Chenet al., “Pyramid: Enabling hierarchical neural networks with edge computing,” inACM WWW’22, New York, NY , USA, 2022, p. 1860–1870

work page 2022

[2] [2]

A distributed hierarchical deep computation model for federated learning in edge computing,

H. Zheng, M. Gao, Z. Chenet al., “A distributed hierarchical deep computation model for federated learning in edge computing,”IEEE Transactions on Industrial Informatics, vol. 17, no. 12, pp. 7946–7956, 2021

work page 2021

[3] [3]

Vision-language models for edge networks: A comprehensive survey,

A. Sharshar, L. U. Khan, W. Ullahet al., “Vision-language models for edge networks: A comprehensive survey,”IEEE Internet Things Journal, vol. 12, no. 16, pp. 32 701–32 724, 2025

work page 2025

[4] [4]

Tri-ring: Asynchronous service provisioning with online learning in edge cloud networks,

X. Ren, Q. Li, H. Duet al., “Tri-ring: Asynchronous service provisioning with online learning in edge cloud networks,” inIEEE INFOCOM, 2025, pp. 1–10

work page 2025

[5] [5]

Freshness-aware inference services in edge computing via offloading or local processing,

X. Ai and W. Liang, “Freshness-aware inference services in edge computing via offloading or local processing,” inIEEE LCN. IEEE, 2025, pp. 1–10

work page 2025

[6] [6]

Efficient coordination of federated learning and inference offloading at the edge: A proactive optimization paradigm,

K. Luo, K. Zhao, T. Ouyanget al., “Efficient coordination of federated learning and inference offloading at the edge: A proactive optimization paradigm,”IEEE Transactions Mobile Computing, vol. 24, no. 1, pp. 407–421, 2025

work page 2025

[7] [7]

Efficient online dnn inference with continuous learning in edge computing,

Y . Zeng, R. Zhou, L. Jiaoet al., “Efficient online dnn inference with continuous learning in edge computing,” inIEEE/ACM IWQoS, 2024, pp. 1–10

work page 2024

[8] [8]

Joint optimization of model retraining and inference services in dt-assisted edge computing,

X. Ai, W. Liang, and C. Liu, “Joint optimization of model retraining and inference services in dt-assisted edge computing,”IEEE/ACM Transac- tions Networking, vol. 34, pp. 1804–1819, 2025

work page 2025