Sensitivity analysis for feedforward artificial neural networks with differentiable activation functions

“…Importantly, computational techniques for determining first-order partial derivatives of certain ANNs have been available for some time. One such technique, outlined by Hashem (1992), involves the application of a simple backward chaining partial differentiation rule. His general rule is adapted in Eq.…”

Section: Enabling the Ddmmf For Ann Models: Revealing Mechanistic Behmentioning

confidence: 99%

Legitimising data-driven models: exemplification of a new data-driven mechanistic modelling framework

Mount

Dawson

Abrahart

2013

Hydrol. Earth Syst. Sci.

View full text Add to dashboard Cite

Abstract.In this paper the difficult problem of how to legitimise data-driven hydrological models is addressed using an example of a simple artificial neural network modelling problem. Many data-driven models in hydrology have been criticised for their black-box characteristics, which prohibit adequate understanding of their mechanistic behaviour and restrict their wider heuristic value. In response, presented here is a new generic data-driven mechanistic modelling framework. The framework is significant because it incorporates an evaluation of the legitimacy of a data-driven model's internal modelling mechanism as a core element in the modelling process. The framework's value is demonstrated by two simple artificial neural network river forecasting scenarios. We develop a novel adaptation of first-order partial derivative, relative sensitivity analysis to enable each model's mechanistic legitimacy to be evaluated within the framework. The results demonstrate the limitations of standard, goodness-of-fit validation procedures by highlighting how the internal mechanisms of complex models that produce the best fit scores can have lower mechanistic legitimacy than simpler counterparts whose scores are only slightly inferior.

show abstract

“…The network input and output are x and the learned version of ϕ(x) defined in (27), respectively. There is also one extra output corresponding to the sensitivity of the cost function, i.e., ∂ϕ/∂x, [20].…”

Section: A Learning and Optimizationmentioning

confidence: 99%

“…The first order output derivative of the MMF networks, i.e., ∂a 1 (2)/∂a 1 (0), can be calculated by applying a backward chaining partial differentiation rule that is described in detail in [20]:…”

Section: A Learning and Optimizationmentioning

confidence: 99%

Analytical and learning‐based spectrum sensing time optimisation in cognitive radio systems

2013

View full text Add to dashboard Cite

PostprintThis is the accepted version of a paper published in IET Communications. This paper has been peerreviewed but does not include the final publisher proof-corrections or journal pagination.Citation for the original published paper (version of record):Analytical and learning-based spectrum sensing time optimisation in cognitive radio systems. AbstractIn this paper, the average throughput maximization of a secondary user by optimizing its spectrum sensing time is formulated, assuming that a prior knowledge of the presence and absence probabilities of the primary users is available. The energy consumed to find a transmission opportunity is evaluated, and a discussion on the impacts of the number of the primary users on the secondary user throughput and consumed energy are presented.In order to avoid the challenges associated with the analytical method, as a second solution, a systematic adaptive neural network-based sensing time optimization approach is also proposed. The proposed scheme is able to find the optimum value of the channel sensing time without any prior knowledge or assumption about the wireless environment. The structure, performance, and cooperation of the artificial neural networks used in the proposed method are explained in detail, and a set of illustrative simulation results is presented to validate the analytical results as well as the performance of the proposed learning-based optimization scheme. Index TermsCognitive radio, spectrum handover, artificial neural networks, maximum secondary user's throughput.

show abstract

Sensitivity analysis for feedforward artificial neural networks with differentiable activation functions

Cited by 76 publications

References 10 publications

Feature Selection With Neural Networks

Feature Selection With Neural Networks

Legitimising data-driven models: exemplification of a new data-driven mechanistic modelling framework

Analytical and learning‐based spectrum sensing time optimisation in cognitive radio systems

Contact Info

Product

Resources

About