Cloud inference can match or exceed on-device performance for latency-sensitive control in distributed CPS when high-throughput resources amortize network and queueing delays.
Ood-maml: Meta-learning for few-shot out-of- distribution detection and classification,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Cloud Is Closer Than It Appears: Revisiting the Tradeoffs of Distributed Real-Time Inference
Cloud inference can match or exceed on-device performance for latency-sensitive control in distributed CPS when high-throughput resources amortize network and queueing delays.