BlockBeats report: On March 18, the LazAI team’s paper, “QoS-Aware Token Scheduling and Private Data Valuation for Multi-Modal Agentic Networks,” was officially accepted by IEEE ICME 2026.
IEEE ICME (International Conference on Multimedia and Expo) is one of the premier academic conferences in the fields of multimedia and intelligent computing. This year, the conference received 3,810 valid submissions, and ultimately accepted 1,101 papers, resulting in an acceptance rate of 28.89%.
This paper focuses on two core issues in AI Agent networks:
Token Scheduling Optimization: How to achieve efficient resource allocation and service quality assurance in multimodal agent collaboration scenarios;
Privacy Data Valuation: How to accurately assess the value of data contributions in AI training and inference while safeguarding data privacy;
These two research directions align closely with the decentralized AI infrastructure LazAI is building—DAT (Data Anchoring Token) addresses data assetization and revenue distribution, while the verification computation framework ensures trustworthy execution of Agent behaviors. This paper acceptance signifies that LazAI’s technical approach is gaining recognition from the mainstream academic community.
The paper will be officially published and presented in person during the ICME 2026 conference.