Search results for "NEURAL NETWORK"
showing 10 items of 1385 documents
Foetal ECG recovery using dynamic neural networks
2002
Non-invasive electrocardiography has proven to be a very interesting method for obtaining information about the foetus state and thus to assure its well-being during pregnancy. One of the main applications in this field is foetal electrocardiogram (ECG) recovery by means of automatic methods. Evident problems found in the literature are the limited number of available registers, the lack of performance indicators, and the limited use of non-linear adaptive methods. In order to circumvent these problems, we first introduce the generation of synthetic registers and discuss the influence of different kinds of noise to the modelling. Second, a method which is based on numerical (correlation coe…
A Novel Multi-step Finite-State Automaton for Arbitrarily Deterministic Tsetlin Machine Learning
2020
Due to the high energy consumption and scalability challenges of deep learning, there is a critical need to shift research focus towards dealing with energy consumption constraints. Tsetlin Machines (TMs) are a recent approach to machine learning that has demonstrated significantly reduced energy usage compared to neural networks alike, while performing competitively accuracy-wise on several benchmarks. However, TMs rely heavily on energy-costly random number generation to stochastically guide a team of Tsetlin Automata (TA) to a Nash Equilibrium of the TM game. In this paper, we propose a novel finite-state learning automaton that can replace the TA in TM learning, for increased determinis…
Using the Hermite Regression Formula to Design a Neural Architecture with Automatic Learning of the “Hidden” Activation Functions
2000
The value of the output function gradient of a neural network, calculated in the training points, plays an essential role for its generalization capability. In this paper a feed forward neural architecture (αNet) that can learn the activation function of its hidden units during the training phase is presented. The automatic learning is obtained through the joint use of the Hermite regression formula and the CGD optimization algorithm with the Powell restart conditions. This technique leads to a smooth output function of αNet in the nearby of the training points, achieving an improvement of the generalization capability and the flexibility of the neural architecture. Experimental results, ob…
Aging and the fluctuation dissipation ratio in a Lennard-Jones fluid
1999
We discuss numerically the relaxation dynamics of a simple structural glass which has been quenched below its (computer) glass transition temperature. We demonstrate that time correlation functions show strong aging effects and compute the fluctuation dissipation ratio of this non-equilibrium system.
Robustness, Stability, and Fidelity of Explanations for a Deep Skin Cancer Classification Model
2022
Skin cancer is one of the most prevalent of all cancers. Because of its being widespread and externally observable, there is a potential that machine learning models integrated into artificial intelligence systems will allow self-screening and automatic analysis in the future. Especially, the recent success of various deep machine learning models shows promise that, in the future, patients could self-analyse their external signs of skin cancer by uploading pictures of these signs to an artificial intelligence system, which runs such a deep learning model and returns the classification results. However, both patients and dermatologists, who might use such a system to aid their work, need to …
Sea Surface Temperatures and Paleoenvironmental Variability in the Central Mediterranean During Historical Times Reconstructed Using Planktonic Foram…
2019
The ongoing anthropogenic‐induced warming assessment requires a robust background from regional sea surface temperature (SST) reconstructions. Planktonic foraminifera have yielded valuable insights into late Quaternary SST dynamics, but the techniques to estimate SST from fossil assemblages have only rarely been used in very recent sedimentary records (the last 2,000 years). Here we use two transfer function methods, modern analog technique and artificial neural networks, to reconstruct SST variability in two cores from the Central Mediterranean Sea that span the last five centuries. Both cores show similar and considerable changes in the planktonic foraminifera assemblages. However, the in…
Glass Transition and Glass Dynamics
2014
The transition from an undercooled liquid towards a glass (glass transition) is introduced and discussed in terms of mode-coupling theory. It is demonstrated that mode-coupling theory leads to a two-step relaxation scenario near the transition with time-critical exponents, which characterize the two relaxation steps (beta and alpha relaxation). The anomalous vibrational properties of a disordered solid (glass) is explained in terms of a model with spatially fluctuating harmonic force constants.
Neural network models for prediction of trichothecene content in wheat
2008
Fusarium graminearum is a mould that causes serious diseases in cereals worldwide and that synthesises mycotoxins such as deoxynivalenol (DON), which can seriously affect human and animal health. Predicting the level of mycotoxin accumulation in food is very difficult, because of the complexity of the influencing parameters. In this work, we have studied the possibility of using artificial neural networks (NN) to predict DON level attained in F. graminearum wheat cultures taking as inputs the fungal contamination level of the cereal, the water activity as a measure of the available water for fungal growth in the cereal, the temperature and time. DON analysis was performed by gas chromatogr…
A tale of two trade-offs: Effects of opening pathways from vocational to higher education
2021
Abstract This paper studies the effects of a vocational secondary school reform implemented in Finland between 1999 and 2001. The reform extended vocational two-year programs to three years and made all graduates eligible to apply for university. For identification, we exploit the gradual implementation of the reform, and use a differences-in-differences approach and administrative register data up to 13 years after the reform. We find no long-term effect on enrollment in further education or labor market outcomes. However, our results illustrate that the reform increased the dropout probability. Thus, the benefits of opening pathways from vocational to higher education may be outweighed by…
A genetic integrated fuzzy classifier
2005
This paper introduces a new classifier, that is based on fuzzy-integration schemes controlled by a genetic optimisation procedure. Two different types of integration are proposed here, and are validated by experiments on real data sets of biological cells. The performance of our classifier is tested against a feed-forward neural network and a Support Vector Machine. Results show the good performance and robustness of the integrated classifier strategies.