The authors prove supply-demand converses and a receiver-internal compute min-cut bound for reliable remote inference under unreliable communication and computation, showing that certain losses are closure-dependent rather than universal.
Title resolution pending
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.IT 2years
2026 2verdicts
UNVERDICTED 2representative citing papers
A polynomial-time additive approximation scheme for rate-constrained dyadic coding of discrete distributions such as LLM token probabilities, with application to steganography.
citing papers explorer
-
Reliable Remote Inference from Unreliable Components: Joint Communication and Computation Limits
The authors prove supply-demand converses and a receiver-internal compute min-cut bound for reliable remote inference under unreliable communication and computation, showing that certain losses are closure-dependent rather than universal.
-
An Additive Approximation Scheme for Generating Dyadic Codings for the Outputs of an LLM
A polynomial-time additive approximation scheme for rate-constrained dyadic coding of discrete distributions such as LLM token probabilities, with application to steganography.