Bayes-Sufficient Compression Is Not Enough: How Does Communication Help Multi-Agent Systems? 35DNCV14M4

URI	`m2://res/35DNCV14M4`
Kind	preprint
Creators	Yi Xie, Zhanke Zhou, Yi Fan, Yong Ge, Bo Han, Bo Liu
Date	2026-09-29
Subjects	arxiv cs.LG, arxiv cs.IT, arxiv cs.MA, arxiv math.IT
Scope	border

Expressions

Date	2026-09-29
Ids	arXiv 2610.03769

The mentions of the inventory that name this resource. Each name links to the card of its inventory row.

Item	Kind	Where	Source
Variational Lower Bound on Idealized Benefit	statement	proposition 11	arxiv
Sufficient Compression Wins When Compatibility Beats Tax	statement	corollary B.5	arxiv
Bayes-Sufficient Compression Need Not Be Receiver-Optimal	statement	theorem 3	arxiv
Receiver-Class Compression Certificate	statement	proposition 8	arxiv
Compression Wins by Saving More Tax Than It Loses	statement	theorem 2	arxiv
Post-Message Loss Decomposes into Three Stages	statement	theorem 4	arxiv
Fixed-Message Replay Isolates Receiver-Side Loss	statement	proposition 5	arxiv
Compatibility Condition for Compression Dominance	statement	corollary B.2	arxiv
Communication-Tax Identity	statement	proposition 1	arxiv
CNRYQW75G5	statement	lemma	arxiv
Proper Bayes-Sufficient Compression	definition	definition 2.2	arxiv
Sender-Upgrade Burden Threshold under Single Crossing	statement	theorem 6	arxiv
Factorized Target under Additive Loss	statement	proposition 10	arxiv
Action-Sufficient Coordination Target	definition	definition 2.1	arxiv
Receiver-Compatible Sufficient Compression	definition	definition B.4	arxiv
Burden-Indexed Receiver Family	definition	definition 2.3	arxiv
Slack is Nonnegative	statement	lemma B.1	arxiv
Alternating Labels Are Hard for Thresholds	statement	lemma B.3	arxiv
Conditional Propagation	statement	proposition 7	arxiv
Plug-In Selector Regret	statement	proposition 9	arxiv
Termwise Calibration Error	statement	corollary B.7	arxiv