Search

While neural networks have been successfully used in a number of machine learning applications, logical languages have been the standard for the representation of argumentative reasoning. In this paper, we establish a relationship between neural networks and argumentation networks, combining reasoning and learning in the same argumentation framework. We do so by presenting a new neural argumentation algorithm, responsible for translating argumentation networks into standard neural networks. We then show a correspondence between the two networks. The algorithm works not only for acyclic argumentation networks, but also for circular networks, and it enables the accrual of arguments through learning as well as the parallel computation of arguments. [ABSTRACT FROM AUTHOR]

The effective integration of knowledge representation, reasoning, and learning in a robust computational model is one of the key challenges of computer science and artificial intelligence. In particular, temporal knowledge and models have been fundamental in describing the behavior of computational systems. However, knowledge acquisition of correct descriptions of a system's desired behavior is a complex task. In this paper, we present a novel neural-computation model capable of representing and learning temporal knowledge in recurrent networks. The model works in an integrated fashion. It enables the effective representation of temporal knowledge, the adaptation of temporal models given a set of desirable system properties, and effective learning from examples, which in turn can lead to temporal knowledge extraction from the corresponding trained networks. The model is sound from a theoretical standpoint, but it has also been tested on a case study in the area of model verification and adaptation. The results contained in this paper indicate that model verification and learning can be integrated within the neural computation paradigm, contributing to the development of predictive temporal knowledge-based systems and offering interpretable results that allow system researchers and engineers to improve their models and specifications. The model has been implemented and is available as part of a neural-symbolic computational toolkit.

Continuous reductions in the dimensions of semiconductor devices have led to an increasing number of noise sources, including random telegraph signals (RTS) due to the capture and emission of electrons by traps at random positions between oxide and semiconductor. The models traditionally used for microscopic devices become of limited validity in nano- and mesoscale systems since, in such systems, distributed quantities such as electron and trap densities, and concepts like electron mobility, become inadequate to model electrical behaviour. In addition, current experimental works have shown that RTS in semiconductor devices based on carbon nanotubes lead to giant current fluctuations. Therefore, the physics of this phenomenon and techniques to decrease the amplitudes of RTS need to be better understood. This problem can be described as a collective Poisson process under different, but time-independent, rates, τ(c) and τ(e), that control the capture and emission of electrons by traps distributed over the oxide. Thus, models that consider calculations performed under time-dependent periodic capture and emission rates should be of interest in order to model more efficient devices. We show a complete theoretical description of a model that is capable of showing a noise reduction of current fluctuations in the time domain, and a reduction of the power spectral density in the frequency domain, in semiconductor devices as predicted by previous experimental work. We do so through numerical integrations and a novel Monte Carlo Markov chain (MCMC) algorithm based on microscopic discrete values. The proposed model also handles the ballistic regime, relevant in nano- and mesoscale devices. Finally, we show that the ballistic regime leads to nonlinearity in the electrical behaviour.