-
A Study on Hyperparameters Configurations for an Efficient Human Activity Recognition System
Authors:
Paulo J. S. Ferreira,
João Mendes Moreira,
João M. P. Cardoso
Abstract:
Human Activity Recognition (HAR) has been a popular research field due to the widespread of devices with sensors and computational power (e.g., smartphones and smartwatches). Applications for HAR systems have been extensively researched in recent literature, mainly due to the benefits of improving quality of life in areas like health and fitness monitoring. However, since persons have different mo…
▽ More
Human Activity Recognition (HAR) has been a popular research field due to the widespread of devices with sensors and computational power (e.g., smartphones and smartwatches). Applications for HAR systems have been extensively researched in recent literature, mainly due to the benefits of improving quality of life in areas like health and fitness monitoring. However, since persons have different motion patterns when performing physical activities, a HAR system must adapt to user characteristics to maintain or improve accuracy. Mobile devices, such as smartphones, used to implement HAR systems, have limited resources (e.g., battery life). They also have difficulty adapting to the device's constraints to work efficiently for long periods. In this work, we present a kNN-based HAR system and an extensive study of the influence of hyperparameters (window size, overlap, distance function, and the value of k) and parameters (sampling frequency) on the system accuracy, energy consumption, and inference time. We also study how hyperparameter configurations affect the model's user and activity performance. Experimental results show that adapting the hyperparameters makes it possible to adjust the system's behavior to the user, the device, and the target service. These results motivate the development of a HAR system capable of automatically adapting the hyperparameters for the user, the device, and the service.
△ Less
Submitted 25 August, 2023;
originally announced August 2023.
-
Van Der Waals Heteroepitaxy of GaSe and InSe, Quantum Wells and Superlattices
Authors:
Marcel S. Claro,
Juan P. Martínez-Pastor,
Alejandro Molina-Sánchez,
Khalil El Hajraoui,
Justyna Grzonka,
Hamid Pashaei Adl,
David Fuertes Marrón,
Paulo J. Ferreira,
Alex Bondarchuk,
Sascha Sadewasser
Abstract:
Bandgap engineering and quantum confinement in semiconductor heterostructures provide the means to fine-tune material response to electromagnetic fields and light in a wide range of the spectrum. Nonetheless, forming semiconductor heterostructures on lattice-mismatched substrates has been a challenge for several decades, leading to restrictions for device integration and the lack of efficient devi…
▽ More
Bandgap engineering and quantum confinement in semiconductor heterostructures provide the means to fine-tune material response to electromagnetic fields and light in a wide range of the spectrum. Nonetheless, forming semiconductor heterostructures on lattice-mismatched substrates has been a challenge for several decades, leading to restrictions for device integration and the lack of efficient devices in important wavelength bands. Here, we show that the van der Waals epitaxy of two-dimensional (2D) GaSe and InSe heterostructures occur on substrates with substantially different lattice parameters, namely silicon and sapphire. The GaSe/InSe heterostructures were applied in the growth of quantum wells and superlattices presenting photoluminescence and absorption related to interband transitions. Moreover, we demonstrate a self-powered photodetector based on this heterostructure on Si that works in the visible-NIR wavelength range. Fabricated at wafer-scale, these results pave the way for an easy integration of optoelectronics based on these layered 2D materials in current Si technology.
△ Less
Submitted 11 October, 2022; v1 submitted 13 June, 2022;
originally announced June 2022.
-
Wafer-scale fabrication of two-dimensional beta-In2Se3 photodetectors
Authors:
Marcel S. Claro,
Justyna Grzonka,
Nicoleta Nicoara,
Paulo J. Ferreira,
Sascha Sadewasser
Abstract:
The epitaxial growth of two-dimensional (2D) $β-In_2Se_3$ material was obtained over 2-inches c-sapphire wafers using molecular beam epitaxy (MBE). Excellent quality of thick (90 nm) and very thin films, down to two quintuple layers (2 nm), was confirmed by x-ray diffraction (XRD), Raman spectroscopy, and aberration-corrected scanning transmission electron microscopy (ac-STEM). Wafer-scale fabrica…
▽ More
The epitaxial growth of two-dimensional (2D) $β-In_2Se_3$ material was obtained over 2-inches c-sapphire wafers using molecular beam epitaxy (MBE). Excellent quality of thick (90 nm) and very thin films, down to two quintuple layers (2 nm), was confirmed by x-ray diffraction (XRD), Raman spectroscopy, and aberration-corrected scanning transmission electron microscopy (ac-STEM). Wafer-scale fabrication of photodetectors based on five quintuple layers was produced using photolithography and other standard semiconductor processing methods. The photodetectors exhibit responsivity of 3 mA/W, peak specific detectivity (D*) of $10^9$ Jones, external quantum efficiency (EQE) of 0.67 % at 550 nm, and response-time of ~7 ms, which is faster than any result previously reported for $β-In_2Se_3$ photodetectors. From the photocurrent measurements, an optical bandgap of 1.38 eV was observed. These results on wafer-scale deposition of 2D $In_2Se_3$, as well as its fabrication into optoelectronic devices provide the missing link that will enable the commercialization of 2D materials.
△ Less
Submitted 30 November, 2020; v1 submitted 21 May, 2020;
originally announced May 2020.
-
Information profiles for DNA pattern discovery
Authors:
Armando J. Pinho,
Diogo Pratas,
Paulo J. S. G. Ferreira
Abstract:
Finite-context modeling is a powerful tool for compressing and hence for representing DNA sequences. We describe an algorithm to detect genomic regularities, within a blind discovery strategy. The algorithm uses information profiles built using suitable combinations of finite-context models. We used the genome of the fission yeast Schizosaccharomyces pombe strain 972 h- for illustration, unveillin…
▽ More
Finite-context modeling is a powerful tool for compressing and hence for representing DNA sequences. We describe an algorithm to detect genomic regularities, within a blind discovery strategy. The algorithm uses information profiles built using suitable combinations of finite-context models. We used the genome of the fission yeast Schizosaccharomyces pombe strain 972 h- for illustration, unveilling locations of low information content, which are usually associated with DNA regions of potential biological interest.
△ Less
Submitted 19 January, 2014;
originally announced January 2014.
-
Unusual Properties of Superoscillating Particles
Authors:
Achim Kempf,
Paulo J. S. G. Ferreira
Abstract:
It has been found that functions can oscillate locally much faster than their Fourier transform would suggest is possible - a phenomenon called superoscillation. Here, we consider the case of superoscillating wave functions in quantum mechanics. We find that they possess rather unusual properties which raise measurement theoretic, thermodynamic and information theoretic issues. We explicitly det…
▽ More
It has been found that functions can oscillate locally much faster than their Fourier transform would suggest is possible - a phenomenon called superoscillation. Here, we consider the case of superoscillating wave functions in quantum mechanics. We find that they possess rather unusual properties which raise measurement theoretic, thermodynamic and information theoretic issues. We explicitly determine the wave functions with the most pronounced superoscillations, together with their scaling behavior. We also address the question how superoscillating wave functions could be produced.
△ Less
Submitted 24 May, 2003;
originally announced May 2003.
