Fitted Q-iteration by Functional Networks for control problems

Gaeta, Matteo; Loia, Vincenzo; Miranda, Sergio; Tomasiello, Stefania

doi:10.1016/j.apm.2016.05.049

Cited by 5 publications

(1 citation statement)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In particular, this problem is well formulated in the typical setup of reinforcement learning. Advanced reinforcement learning methods have been applied to robot control problems, with remarkable results [15,16]. In addition, according to the recent advent of deep learning techniques, reinforcement learning methods that can use images as a state by using a convolutional neural network, have attracted increasing attention [17][18][19].…”

Section: Related Workmentioning

confidence: 99%

Neural Model Extraction for Model-Based Control of a Neural Network Forward Model

et al. 2021

View full text Add to dashboard Cite

Neural networks have been widely used to model nonlinear systems that are difficult to formulate. Thus far, because neural networks are a radically different approach to mathematical modeling, control theory has not been applied to them, even if they approximate the nonlinear state equation of a control object. In this research, we propose a new approach-i.e., neural model extraction, that enables model-based control for a feed-forward neural network trained for a nonlinear state equation. Specifically, we propose a method for extracting the linear state equations that are equivalent to the neural network corresponding to given input vectors. We conducted simple simulations of a two degrees-of-freedom planar manipulator to verify how the proposed method enables model-based control on neural network forward models. Through simulations, where different settings of the manipulator's state observation are assumed, we successfully confirm the validity of the proposed method.

show abstract

Section: Related Workmentioning

confidence: 99%

Neural Model Extraction for Model-Based Control of a Neural Network Forward Model

et al. 2021

View full text Add to dashboard Cite

show abstract

Fitted Q-iteration and functional networks for ubiquitous recommender systems

2016

View full text Add to dashboard Cite

Ubiquitous recommender systems facilitate users on-location by personalized recommendations of items in the proximity via mobile devices. Due to a high variability of situations and preferences, an efficient resource processing is needed in order to assist the user in a proper way. In this paper, we consider a recommender system, able to learn preferences/habits of users through contextual information, such as location and time, using a new offline model-free approximate Q-iteration. Following the basic idea of Fitted Q-Iteration, the paper focuses on a computational scheme, based on functional networks, and that, unlike the well-known neural ones, does not require a large number of training samples. A preliminary case study, which deals with a shopping mall, is useful to show the approximation capabilities of the proposed approach

show abstract