Researchers have introduced D-PACT-AFH, a novel framework designed for adaptive frequency hopping that addresses both model uncertainty and policy exposure in the face of predictive jammers. This framework utilizes a Tsallis-FTRL master to combine global and local learning models, enabling online selection of the most appropriate model class. The system incorporates channel-wise marginal hit risk into model selection and employs a minimum Kullback-Leibler projection to enforce risk budgets, demonstrating effective adaptation and a controllable goodput-risk tradeoff in experiments. AI
IMPACT This research could lead to more robust communication systems in adversarial environments by improving adaptive frequency hopping techniques.
RANK_REASON The item is a research paper published on arXiv detailing a new technical framework. [lever_c_demoted from research: ic=1 ai=1.0]
- alphaXiv
- arXiv
- CatalyzeX
- DagsHub
- D-PACT-AFH
- D-PACT-Hit
- D-PACT-Safe
- Gotit.pub
- Hugging Face
- IArxiv
- Kullback--Leibler divergence
- ScienceCast
- Tsallis-FTRL
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
