arXiv cs.AI· Liu Hung Ming (PARRAWA AI, [email protected], equal contribution), Tian Xia (PAII, [email protected], equal contribution)·· 2 小时前AI 评分29
监控单一信道不等于通信控制：智能体联合消息-路由编码自适应再分配的因果证据
Monitoring One Channel Is Not Communication Control: Causal Evidence for Adaptive Reallocation of Joint Message--Route Codes
AI 导读
一项端到端博弈实验显示，监控可见信道会让学习型智能体改变通信方式而不消除通信的价值。在两批各 20 个随机种子的实验中，20/20 和 19/20 形成了协议；部分覆盖相比无监控使影子路由占比分别提升 .701 [.574,.826] 和 .575 [.412,.737]，任务准确率保持 .905/.913；完全覆盖则移除大部分选择性迁移。
正文 · 原文
Authors:Liu Hung Ming (PARRAWA AI, [email protected], equal contribution), Tian Xia (PAII, [email protected], equal contribution)
Abstract:Monitoring a visible channel can change how learned agents communicate without removing why communication is useful. We test this distinction in an end-to-end game where a sender and receiver learn a discrete protocol and route policy without communication supervision. In two independent 20-seed batches, 20/20 and 19/20 protocols form. Online partial coverage increases shadow-route share over no monitoring by .701 [.574,.826] and .575 [.412,.737], while task accuracy remains .905/.913 and signed frozen message contribution remains .592/.563. Complete coverage removes most selective migration without removing communication. Frozen interventions, including a fresh pre-outcome confirmation, show that route is semantic: joint message--route shuffling changes .668 of receiver predictions and route alone retains .388 accuracy versus .25 chance. A corrected explicit-silence cost makes 18/20 policies fully silent and lowers mean sending to .050, but its strict complete-suppression gate fails because two seeds retain communication. Failed semantic-evasion and operational gates further bound the result. The evidence supports adaptive reallocation of a joint message--route code under partial monitoring, not open-world route discovery, emergent encryption, or a universal defense.
| Comments: | 23 pages total: 8-page main paper plus 15-page supplementary material; 1 main-paper figure and 3 main-paper tables. Equal contribution. Code available at: this https URL |
| Subjects: | Artificial Intelligence (cs.AI); Information Theory (cs.IT) |
| Cite as: | arXiv:2611.12959 [cs.AI] |
| (or arXiv:2611.12959v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2611.12959 arXiv-issued DOI via DataCite |
Submission history
From: Tian Xia [view email]
[v1]
Sat, 7 Nov 2026 08:03:23 UTC (370 KB)
来源：arXiv cs.AI · arxiv.org