New Stochastic Reset Pathfinding framework introduced for graph-based learning

Researchers have introduced Stochastic Reset Pathfinding (SRP), a new episodic learning problem designed for scenarios involving unknown edge success probabilities on directed graphs. This framework is applicable to diverse fields such as quantum repeater networks, payment routing on the Lightning Network, and delivery in mesh networks. The proposed solution, Log-Dijkstra meta-algorithm with UCB (PathUCB) and Thompson Sampling (PathTS) instantiations, offers a path-level regret bound that provides more detailed insights for structured graphs compared to existing edge-level bounds. Experiments indicate PathTS generally performs best empirically, though adversarial instances exist where its convergence is not guaranteed. AI

IMPACT Introduces a new theoretical framework and algorithms for graph-based learning problems with potential applications in various domains.

RANK_REASON The cluster contains a research paper detailing a new theoretical framework and algorithms for a machine learning problem. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• Dijkstra
• Lightning Network
• PathTS
• PathUCB
• Stochastic Reset Pathfinding
• Thompson sampling
• University of California, Berkeley

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →