Feed
researchAI Infrastructure3d ago

PUMA-Former Framework Introduced for Agentic Computation Offloading in MEO-LEO Satellite Networks

Facts

Researchers introduced PUMA-Former, an agentic framework designed to optimize workload offloading in MEO-LEO satellite networks under conditions of incomplete information. The model utilizes uncertainty-aware mechanisms, including latent-delay attention and Transformer-based allocation, to improve service completion and reduce latency compared to existing baselines.

  • ▸PUMA-Former uses an uncertainty-aware agentic framework to coordinate mobile workload distribution.
  • ▸The framework addresses stale or missing CPU state data common in MEO-LEO satellite networks.
  • ▸Performance testing shows an 8.62% increase in admitted compute-work completion.
  • ▸Mean service delay was reduced by 34.78% and mean episode-P95 delay by 71.08% compared to ATLA-adapt.
  • ▸The system incorporates a structured Transformer allocator for bandwidth, CPU, and power management.
Occurred
2026-11-13 13:08 UTC
First reported
2026-11-13 13:08 UTC
Last updated
2026-11-16 07:15 UTC
Event type
research
Region
global

Entities

Categories

Source timeline · 1

  • arXivFirst-party2026-11-13 13:08
Community vote

Sign in to vote on importance and sentiment.

Platform Signals

Impact score
54
AI sentiment
Positive
Event significance30%0
Entity importance20%0
Source credibility20%1
Novelty15%1
Corroboration15%1
Confidence1
Novelty1
Corroboration1
Community
ImportanceNo votes
Strategic sentiment
Positive 0%Neutral 0%Negative 0%

Platform signals are algorithmic and community-derived indicators. They are not investment recommendations and do not predict valuation or performance.

Related events

No events yet.