Mixture-of-Experts under Finite-Rate Gating: Communication–Generalization Trade-offs
Ali Khalesi, Mohammad Reza Deylam Salehi · 2026
Read Mixture-of-Experts under Finite-Rate Gating: Communication–Generalization Trade-offs by Ali Khalesi, Mohammad Reza Deylam Salehi (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-15091.txt.
Description
Khalesi and Salehi study mixture-of-experts models through the lens of communication theory. Such models split prediction across specialized expert sub-networks chosen by a gating step. This letter treats the gate as a stochastic channel working under a finite information rate, asking what limited gating bandwidth costs in learning.
Within an information-theoretic learning framework, the authors specialize a mutual-information generalization bound and develop a rate-distortion account of finite-rate gating. The analysis yields capacity-aware limits for communication-constrained expert systems. Numerical tests on synthetic multi-expert models confirm the predicted trade-offs among gating rate, expressivity, and generalization, tying how much the gate can say to how well the system learns.
Catalog record
| CREATOR | Ali Khalesi, Mohammad Reza Deylam Salehi |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 30 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-15091