Attention-driven auditory stream segregation using a SOM coupled with an excitatory-inhibitory ANN

The 2013 International Joint Conference on Neural Networks (IJCNN)

et al. 2013

Self Cite

Abstract-In this paper, a human-mimicking model for sound source recognition is presented. It consists of an artificial neural network with three neuron layers (input, middle and output) that are connected by feedback connections between the output and middle layer, on top of feedforward connections from the input to middle and middle to output layers. Learning is accomplished by the model following the Hebb principle, dictating that "cells that fire together, wire together", with some important alterations, compared to standard Hebbian learning, in order to prevent the model from forgetting previously learned patterns, when learning new ones. In addition, short-term memory is introduced into the model in order to facilitate and guide learning of neuronal synapses (long-term memory). As auditory attention is an essential part of human auditory scene analysis (ASA), it is also indispensable in any computational model mimicking it, and it is shown that different auditory attention mechanism naturally emerge from the neuronal behaviour as implemented in the model described in this paper. The learning behavior of the model is further investigated in the context of an urban sonic environment, and the importance of shortterm memory in this process is demonstrated. Finally, the effectiveness of the model is evaluated by comparing model output on presented sound recordings to a human expert listeners evaluation of the same fragments.

Section: A Network Architecture and Neural Activationmentioning

confidence: 96%

Section: A Network Architecture and Neural Activationmentioning

confidence: 99%

See 1 more Smart Citation

A biologically inspired recurrent neural network for sound source recognition incorporating auditory attention

Boes

Oldoni

The 2013 International Joint Conference on Neural Networks (IJCNN)

et al. 2013

Self Cite

“…The concept of inhibition-of-return is usually introduced in human attention models in order to prevent attention from permanently staying fixed on one single item [12]. Whereas these attention mechanisms could be included in existing human perception models by artificially adding extra parameters and submodels [6], in the current model, they automatically arise from the way in which biological neural behavior has been implemented.…”

Section: Auditory Attentionmentioning

confidence: 99%

“…In previous work by the same authors, a biologically inspired neural network model for auditory scene analysis, incorporating both auditory attention and learning flexibility, was developed [6] [7], upon which the model presented in this paper builds. The model consists of 3 neural layers, connected to each other by feedforward excitatory connections as well as feedback excitatory connections between the last two layers.…”

Section: Introductionmentioning

confidence: 99%

Long-term learning behavior in a recurrent neural network for sound recognition

Boes

Oldoni

2014 International Joint Conference on Neural Networks (IJCNN)

et al. 2014

Self Cite

Abstract-In this paper, the long-term learning properties of an artificial neural network model, designed for sound recognition and computational auditory scene analysis in general, are investigated. The model is designed to run for long periods of time (weeks to months) on low-cost hardware, used in a noise monitoring network, and builds upon previous work by the same authors. It consists of three neural layers, connected to each other by feedforward and feedback excitatory connections. It is shown that the different mechanisms that drive auditory attention emerge naturally from the way in which neural activation and intra-layer inhibitory connections are implemented in the model. Training of the artifical neural network is done following the Hebb principle, dictating that "Cells that fire together, wire together", with some important modifications, compared to standard Hebbian learning. As the model is designed to be on-line for extended periods of time, also learning mechanisms need to be adapted to this. The learning needs to be strongly attention-and saliency-driven, in order not to waste available memory space for sounds that are of no interest to the human listener. The model also implements plasticity, in order to deal with new or changing input over time, without catastrophically forgetting what it already learned. On top of that, it is shown that also the implementation of shortterm memory plays an important role in the long-term learning properties of the model. The above properties are investigated and demonstrated by training on real urban sound recordings.

Auditory sensory saliency as a better predictor of change than sound amplitude in pleasantness assessment of reproduced urban soundscapes

Filipan

Building and Environment

Aumond

et al. 2019

Self Cite

The sonic environment of the urban public space is often experienced while walking through it. Nevertheless, city dwellers are usually not actively listening to the environment when traversing the city. Therefore, sound events that are salient, i.e. stand out of the sonic environment, are the ones that trigger attention and contribute highly to the perception of the soundscape. In a previously reported audiovisual perception experiment, the pleasantness of a recorded urban sound walk was continuously evaluated by a group of participants. To detect salient events in the soundscape, a biologically-inspired computational model for auditory sensory saliency based on spectrotemporal modulations is proposed. Using the data from a sound walk, the present study validates the hypothesis that salient events detected by the model contribute to changes in soundscape rating and are therefore important when evaluating the urban soundscape. Finally, when using the data from an additional experiment without a strong visual component, the importance of auditory sensory saliency as a predictor for change in pleasantness assessment is found to be even more pronounced.