QoS-Aware Token Scheduling and Private Data Valuation for Multi-Modal Agentic Networks

Yao Du; Jing Liu; Pengfei Xu; Zehua Wang; Victor C. M. Leung; Cyril Leung; Victoria Lemieux

cs.AIApr 2, 2026

QoS-Aware Token Scheduling and Private Data Valuation for Multi-Modal Agentic Networks

Authors: Yao Du, Jing Liu, Pengfei Xu, Zehua Wang, Victor C. M. Leung, Cyril Leung, Victoria Lemieux

Organizations: The University of British Columbia · LazAI Network · China University of Mining and Technology · SMBU ·

Abstract

In agentic systems, human-generated data records anchor the value of AI services. Yet cloud compute pipelines centralize processing on remote servers. Data centralization reduces personal data sovereignty and may potentially degrade the quality of service (QoS). Meanwhile, user contributions are diverse in quantity and quality: decentralized records can be biased, noisy, and heterogeneously distributed. To address the data challenge, we study fair token allocation and private data valuation for decentralized and resource-constrained agentic systems. Our approach embeds multi-modal representations in a shared semantic space and releases differentially private (DP) prototypes to preserve utility while reducing semantic leakage. With the DP guarantee, we design a fair token allocation scheme that rewards effective contributions and remains robust to data heterogeneity and AI resource scarcity. Extensive simulations demonstrate improved contribution-based fairness and QoS compared to standard benchmarks. The improved resistance to image reconstruction attacks indicates enhanced privacy for multi-modal personal data.

Figures & tables

Fig. 1: Cloud-edge-end collaborative agentic framework. An example with $M=2$ and $N=3$ .

Fig. 2: Token-bucket scheduling in a three-agent system. Utility quotas are represented as tokens capped by per-agent bucket sizes. Blockchain servers read the DAT information to distribute tokens.

Fig. 3: DP noise is added to high-dimensional private data before data valuation. Each semantic cluster forms an iDAO to govern data valuation and reward settlement.

Panel of Figure 4 — Fig. 4: Fairness metrics under different $\alpha$ . Our proposed token allocation improves fairness across diverse contribution distributions.

Panel of Figure 5 — Fig. 5: Fairness metrics with different $N$ . Our method supports real-world deployment under constrained AI quota and compute budgets.

Panel of Figure 6 — Fig. 6: Data valuation in novel discovery. Our proposed scheme incentivizes early contributions while preserving privacy by applying DP before prototypes are shared.

Panel of Figure 7 — Fig. 7: Attacks on shared DP prototypes. Our method improves privacy level on multi-modal data.

Panel of Figure 8 — Fig. S.1: Reconstruction attack comparison: (top) Image-DP benchmark [ 11 ] shows substantial semantic leakage; (middle) our Image-DP prototype is well-protected; (bottom) our Text-DP prototype is well-protected.

Fig. S.2: Multi-modal datasets are highly skewed in quantity by nature: An example with the COCO validation dataset.

Fig. S.3: Image samples from the shareable cluster: Detected cluster in the person class.

Fig. S.4: Image samples from the shareable cluster: Detected cluster in the person class.

Data Valuation

Data valuation estimates how examples, datasets, or contributors affect model performance to guide selection, auditing, and fair compensation.

See all 44 papers

Agentic AI Systems Should Be Designed as Marginal Token Allocators
May 2, 2026Siqi ZhuLLM Agent Orchestration LLM Agent Evaluation
University of Illinois Urbana-Champaign
Token Management in Multi-Tenant AI Inference Platforms
Feb 27, 2026William J. CunninghamLLM Inference Serving Fair Resource Allocation
DataRobot, Boston, Massachusetts, USA
Utility-Aware Data Pricing: Token-Level Quality and Empirical Training Gain for LLMs
Apr 24, 2026Minghui Xu, Qi Luo, Kun Li +1Data Valuation LLM Training
Shandong University, China

May 2, 2026J/K move · Enter open · S save

Agentic AI Systems Should Be Designed as Marginal Token Allocators

Siqi Zhu

University of Illinois Urbana-Champaign

LLM Agent Orchestration LLM Agent Evaluation LLM Agent Training Agentic AI

Open paper AskPDF arXiv

This position paper argues that agentic AI systems should be designed and evaluated as marginal token allocation economies rather than as text generators priced by the unit. We follow a single request -- a developer asking a coding agent to fix a failing test -- through four economic layers that today are designed in isolation: a router that decides which model answers, an agent that decides whether to plan, act, verify, or defer, a serving stack that decides how to produce each token, and a training pipeline that decides whether the trace is worth learning from. We show that all four layers are solving the same first-order condition -- marginal benefit equals marginal cost plus latency cost plus risk cost -- with different index sets and different prices. The framing is deliberately minimal: we do not propose a complete theory of AI economics. But adopting marginal token allocation as the shared accounting object explains why systems that locally minimize tokens globally misallocate them, predicts a small set of recurring failure modes (over-routing, over-delegation, under-verification, serving congestion, stale rollouts, cache misuse), and points to a concrete research agenda in token-aware evaluation, autonomy pricing, congestion-priced serving, and risk-adjusted RL budgeting.

Abstract

Data Valuation

Explore similar work

Agentic AI Systems Should Be Designed as Marginal Token Allocators

Token Management in Multi-Tenant AI Inference Platforms

Utility-Aware Data Pricing: Token-Level Quality and Empirical Training Gain for LLMs