arXiv subjects

Osvaldo Agamennoni

Publications and source records attributed to Osvaldo Agamennoni.

4 recordsLinked to original sources

arXiv2018

Eye-Movement behavior identification for AD diagnosis

In the present work, we develop a deep-learning approach for differentiating the eye-movement behavior of people with neurodegenerative diseases over healthy control subjects during reading well-defined sentences. We define an information compaction of the eye-tracking data of subjects without and with probable Alzheimer's disease when reading a set of well-defined, previously validated, sentences including high-, low-predictable sentences, and proverbs. Using this information we train a set of denoising sparse-autoencoders and build a deep neural network with these and a softmax classifier. Our results are very promising and show that these models may help to understand the dynamics of eye movement behavior and its relationship with underlying neuropsychological correlates.

cs.NE↗

arXiv2017

Pupil size behavior during on line processing of sentences

In the present work we analyzed the pupil size behavior of forty subjects while they read well defined sentences with different contextual predictability (i.e., regular sentences and proverbs). In general, pupil size increased when reading regular sentences, but when readers realized that they were reading proverbs their pupils strongly increase until finishing proverbs' reading. Our results suggest that an increased pupil size is not limited to cognitive load (i.e., relative difficulty in processing) because when participants accurately recognized words during reading proverbs, theirs pupil size increased too. Our results show that pupil size dynamics may be a reliable measure to investigate the cognitive processes involved in sentence processing and memory functioning.

q-bio.NC↗

arXiv2010

Attraction and Stability of Nonlinear Ode's using Continuous Piecewise Linear Approximations

In this paper, several results concerning attraction and asymptotic stability in the large of nonlinear ordinary differential equations are presented. The main result is very simple to apply yielding a sufficient condition under which the equilibrium point (assuming a unique equilibrium) is attractive and also provides a variety of options among them the classical linearization and other existing results are special cases of the this main theorem in this paper including and extension of the well known Markus-Yamabe conjecture. Several application examples are presented in order to analyze the advantages and drawbacks of the proposed result and to compare such results with successful existing techniques for analysis available in the literature nowadays.

math.OC↗

arXiv2009

Closed-Form Trajectories of the Unicycle

This paper presents a methodology to stabilize some kind of Nonlinear Control system known as Driftless, utilizing the concept of \textit{Pseudo-Kinetic Energy} introduced in this work. Once this controller is applied to the Unicycle-type robot, stability is guaranteed with the salient property that the structure of the controller allows to solve in closed-form the trajectories of the vehicle. While the proposed controller only ensures stability (not asymptotic stability) the obtained closed-form solutions will show a path to obtain in closed form the solutions for the general control problem of the unicycle. Some conclusions and future directions for research are also depicted.

math.OC↗