Research Papers

A control system and a method for controlling search agents to perform search with noisy observations and probabilistic guarantees is provided. The control system collects confidence bounds of a probabilistic classification of at least one region within at least one path of a set of paths. The control system compares aggregations of the confidence bounds of the probabilistic classifications of eac...

We address the challenging problem of non-stationary multi-armed bandits where the reward distributions change over time in a periodic manner. Traditional approaches struggle with time-varying environments due to the exploration-exploitation trade-off becoming more complex when the optimal actions shift cyclically. We propose a novel framework that harnesses Ramanujan Periodicity Transforms to det...

We address autonomous search using teams of multiple agents, requiring tractable coordination strategies that can lower the time to identify interesting areas in the search environment, lower the costs/energy usage by the search agents during movement and sensing, and be resilient to the noise present in the sensed data due to the use of low-cost and low-weight sensors. We propose a data-driven, m...

We consider the pure exploration problem in stochastic multi-armed bandits where the similarities between the arms are captured by a graph and the rewards may be represented as a smooth signal on this graph. We specifically examine the problem of finding the arm with the maximum reward (maximizing problem) or one with a sufficiently high reward (satisficing problem) under this model. We propose no...

Hyperspectral unmixing is an important remote sensing task with applications including material identification and analysis. Characteristic spectral features make many pure materials identifiable from their visible-to-infrared spectra, but quantifying their presence within a mixture is a challenging task due to nonlinearities and factors of variation. We consider spectral variation from a physics-...

We consider the problem of recovering a complex vector from quadratic measurements, known as quadratic feasibility, which encompasses the well known phase retrieval problem and has applications in power system state estimation and x-ray crystallography. While the quadratic feasibility problem is generally NP-hard and may be unidentifiable, we establish conditions under which this problem becomes i...

We consider a strategic problem where multiple players compete to access a shared server platform that operates intermittently, switching between ON and OFF periods. Each player incurs costs to sample the server state and receives payoffs inversely proportional to the number of simultaneously connected players. We propose a distributed randomized learning algorithm that enables players to minimize...