-
Performance Analysis of a Foreground Segmentation Neural Network Model
Authors:
Joel Tomás Morais,
António Ramires Fernandes,
André Leite Ferreira,
Bruno Faria
Abstract:
In recent years the interest in segmentation has been growing, being used in a wide range of applications such as fraud detection, anomaly detection in public health and intrusion detection. We present an ablation study of FgSegNet_v2, analysing its three stages: (i) Encoder, (ii) Feature Pooling Module and (iii) Decoder. The result of this study is a proposal of a variation of the aforementioned…
▽ More
In recent years the interest in segmentation has been growing, being used in a wide range of applications such as fraud detection, anomaly detection in public health and intrusion detection. We present an ablation study of FgSegNet_v2, analysing its three stages: (i) Encoder, (ii) Feature Pooling Module and (iii) Decoder. The result of this study is a proposal of a variation of the aforementioned method that surpasses state of the art results. Three datasets are used for testing: CDNet2014, SBI2015 and CityScapes. In CDNet2014 we got an overall improvement compared to the state of the art, mainly in the LowFrameRate subset. The presented approach is promising as it produces comparable results with the state of the art (SBI2015 and Cityscapes datasets) in very different conditions, such as different lighting conditions.
△ Less
Submitted 25 May, 2021;
originally announced May 2021.
-
Cluster structure of optimal solutions in bipartitioning of small worlds
Authors:
Adam Lipowski,
Antonio L. Ferreira,
Dorota Lipowska
Abstract:
Using a simulated annealing, we examine a bipartitioning of small worlds obtained by adding a fraction of randomly chosen links to a one-dimensional chain or a square lattice. Models defined on small worlds typically exhibit a mean-field behaviour, regardless of the underlying lattice. Our work demonstrates that the bipartitioning of small worlds does depend on the underlying lattice. Simulations…
▽ More
Using a simulated annealing, we examine a bipartitioning of small worlds obtained by adding a fraction of randomly chosen links to a one-dimensional chain or a square lattice. Models defined on small worlds typically exhibit a mean-field behaviour, regardless of the underlying lattice. Our work demonstrates that the bipartitioning of small worlds does depend on the underlying lattice. Simulations show that for one-dimensional small worlds, optimal partitions are finite size clusters for any fraction of additional links. In the two-dimensional case, we observe two regimes: when the fraction of additional links is sufficiently small, the optimal partitions have a stripe-like shape, which is lost for larger number of additional links as optimal partitions become disordered. Some arguments, which interpret additional links as thermal excitations and refer to the thermodynamics of Ising models, suggest a qualitatitve explanation of such a behaviour. The histogram of overlaps suggests that a replica symmetry is broken in a one-dimensional small world. In the two-dimensional case, the replica symmetry seems to hold but with some additional degeneracy of stripe-like partitions.
△ Less
Submitted 19 November, 2020;
originally announced November 2020.
-
Deep Dense and Convolutional Autoencoders for Unsupervised Anomaly Detection in Machine Condition Sounds
Authors:
Alexandrine Ribeiro,
Luis Miguel Matos,
Pedro Jose Pereira,
Eduardo C. Nunes,
Andre L. Ferreira,
Paulo Cortez,
Andre Pilastri
Abstract:
This technical report describes two methods that were developed for Task 2 of the DCASE 2020 challenge. The challenge involves an unsupervised learning to detect anomalous sounds, thus only normal machine working condition samples are available during the training process. The two methods involve deep autoencoders, based on dense and convolutional architectures that use melspectogram processed sou…
▽ More
This technical report describes two methods that were developed for Task 2 of the DCASE 2020 challenge. The challenge involves an unsupervised learning to detect anomalous sounds, thus only normal machine working condition samples are available during the training process. The two methods involve deep autoencoders, based on dense and convolutional architectures that use melspectogram processed sound features. Experiments were held, using the six machine type datasets of the challenge. Overall, competitive results were achieved by the proposed dense and convolutional AE, outperforming the baseline challenge method.
△ Less
Submitted 19 June, 2020; v1 submitted 18 June, 2020;
originally announced June 2020.
-
Bipartitioning of directed and mixed random graphs
Authors:
Adam Lipowski,
Antonio Luis Ferreira,
Dorota Lipowska,
Manuel A. Barroso
Abstract:
We show that an intricate relation of cluster properties and optimal bipartitions, which takes place in undirected random graphs, extends to directed and mixed random graphs. In particular, the satisfability threshold coincides with the relative size of the giant OUT component reaching~{1/2}. Moreover, when counting undirected links as two directed ones, the partition cost, and cluster properties,…
▽ More
We show that an intricate relation of cluster properties and optimal bipartitions, which takes place in undirected random graphs, extends to directed and mixed random graphs. In particular, the satisfability threshold coincides with the relative size of the giant OUT component reaching~{1/2}. Moreover, when counting undirected links as two directed ones, the partition cost, and cluster properties, as well as location of the replica symmetry breaking transition for these random graphs depend primarily on the total number of directed links and not on their specific distribution.
△ Less
Submitted 18 June, 2019; v1 submitted 26 January, 2019;
originally announced January 2019.
-
Emergence of Social Structures via Preferential Selection
Authors:
Adam Lipowski,
Dorota Lipowska,
Antonio Luis Ferreira
Abstract:
We examine a weighted-network multi-agent model with preferential selection such that agents choose partners with the probability $p(w)$, where $w$ is the number of their past selections. When $p(w)$ increases sublinearly with the number of past selections ($p(w)\sim w^α, \ α<1$), agents develop a uniform preference for all other agents. At $α=1$, this state looses stability and more complex struc…
▽ More
We examine a weighted-network multi-agent model with preferential selection such that agents choose partners with the probability $p(w)$, where $w$ is the number of their past selections. When $p(w)$ increases sublinearly with the number of past selections ($p(w)\sim w^α, \ α<1$), agents develop a uniform preference for all other agents. At $α=1$, this state looses stability and more complex structures form. For a superlinear increase ($α>1$), strong heterogeneities emerge and agents make selections mainly within small and sometimes asymmetric clusters. Even in a few-agent case, formation of such clusters resembles phase transitions with spontaneous symmetry breaking.
△ Less
Submitted 17 September, 2014; v1 submitted 7 June, 2014;
originally announced June 2014.
-
Critical behavior and correlations on scale-free small-world networks. Application to network design
Authors:
M. Ostilli,
A. L. Ferreira,
J. F. F. Mendes
Abstract:
We analyze critical phenomena on networks generated as the union of hidden variables models (networks with any desired degree sequence) with arbitrary graphs. The resulting networks are general small-worlds similar to those a` la Watts and Strogatz but with a heterogeneous degree distribution. We prove that the critical behavior (thermal or percolative) remains completely unchanged by the presence…
▽ More
We analyze critical phenomena on networks generated as the union of hidden variables models (networks with any desired degree sequence) with arbitrary graphs. The resulting networks are general small-worlds similar to those a` la Watts and Strogatz but with a heterogeneous degree distribution. We prove that the critical behavior (thermal or percolative) remains completely unchanged by the presence of finite loops (or finite clustering). Then, we show that, in large but finite networks, correlations of two given spins may be strong, i.e., approximately power law like, at any temperature. Quite interestingly, if $γ$ is the exponent for the power law distribution of the vertex degree, for $γ\leq 3$ and with or without short-range couplings, such strong correlations persist even in the thermodynamic limit, contradicting the common opinion that in mean-field models correlations always disappear in this limit. Finally, we provide the optimal choice of rewiring under which percolation phenomena in the rewired network are best performed; a natural criterion to reach best communication features, at least in non congested regimes.
△ Less
Submitted 28 June, 2011; v1 submitted 1 January, 2011;
originally announced January 2011.
-
A model of student's dilemma
Authors:
Adam Lipowski,
Antonio L. Ferreira
Abstract:
Each year perhaps millions of young people face the following dilemma: should I continue my education or rather start working with already acquired skills. Right decision must take into account somebody's own abilities, accessibility to education institutions, competition, and potential benefits. A multi-agent, evolutionary model of this dilemma predicts a transition between stratified and homog…
▽ More
Each year perhaps millions of young people face the following dilemma: should I continue my education or rather start working with already acquired skills. Right decision must take into account somebody's own abilities, accessibility to education institutions, competition, and potential benefits. A multi-agent, evolutionary model of this dilemma predicts a transition between stratified and homogeneous phases, evolution that diminishes fitness, fewer applicants per seat for decreased capacity of the university, and presence of poor students at élite universities.
△ Less
Submitted 8 April, 2005; v1 submitted 22 October, 2004;
originally announced October 2004.
