Neural Network

Know thyself: Metacognitive networks and measures of consciousness

by Bert Timmermans

Pasquali A and* Timmermans B and* Cleeremans A (2010). Know thyself: Metacognitive networks and measures of consciousness. Cognition, 117(2), 182-190. *equal contributions

Subjective measures of awareness rest on the assumption that conscious knowledge is knowledge that participants know... more Subjective measures of awareness rest on the assumption that conscious knowledge is knowledge that participants know they possess. Post-Decision Wagering (PDW), recently proposed as a new measure of awareness, requires participants to place a high or a low wager on their decisions. Whereas advantageous wagering indicates awareness of the knowledge on which the decisions are based, cases in which participants fail to optimize their wagers suggest performance without awareness. Here, we hypothesize that wagering and other subjective measures of awareness reflect metacognitive capacities subtended by self-developed metarepresentations that inform an agent about its own internal states. To support this idea, we present three simulations in which neural networks learn to wager on their own responses. The simulations illustrate essential properties that are required for such metarepresentations to influence PDW as a measure of awareness. Results demon strate a good fit to human data. We discuss the implications of this modeling work for our understanding of consciousness and its measures.

The confusion effect in predatory neural networks

by Andrew Jackson

A simple artificial neural network model of image reconstruction in sensory maps is presented to explain the... more A simple artificial neural network model of image reconstruction in sensory maps is presented to explain the difficulty predators experience in targeting prey in large groups (the confusion effect). Networks are trained to reconstruct multiple randomly conformed “retinal” images of prey groups in an internal spatial map of their immediate environment. They are then used to simulate prey targeting by predators on groups of specific conformation. Networks trained with the biologically plausible associative reward‐penalty method produce a more realistic model of the confusion effect than those trained with the popular but biologically implausible backpropagation method. The associative reward‐penalty model makes the novel prediction that the accuracy–group size relationship is U shaped, and this prediction is confirmed by empirical data gathered from interactive computer simulation experiments with humans as “predators.” The model further predicts all factors known from previous empirical work (and most factors suspected) to alleviate the confusion effect: increased relative intensity of the target object, heterogeneity of group composition, and isolation of the target. Interestingly, group compaction per se is not predicted to worsen predator confusion. This study indicates that the relatively simple, nonattentional mechanism of information degradation in the sensory mapping process is potentially important in generating the confusion effect.

Neural Network Time Series Models for Financial Risk Management

by Francesco Virili

Nonstationarity and data preprocessing for neural network predictions of an economic time series

by Francesco Virili

Consciousness and Metarepresentation: A Computational Sketch

by Bert Timmermans

Cleeremans A, Timmermans B, & Pasquali A (2007). Consciousness and metarepresentation: A computational sketch. Neural Networks, 20, 1032-9.

When one is conscious of something, one is also conscious that one is conscious. Higher-Order Thought Theory... more When one is conscious of something, one is also conscious that one is conscious. Higher-Order Thought Theory [Rosenthal, D. (1997). A theory of consciousness. In N. Block, O. Flanagan, & G. Güzeldere (Eds.), The nature of consciousness: Philosophical debates. Cambridge, MA: MIT Press] takes it that it is in virtue of the fact that one is conscious of being conscious, that one is conscious. Here, we ask what the computational mechanisms may be that implement this intuition. Our starting point is Clark and Karmiloff-Smith’s [Clark, A., & Karmiloff-Smith, A. (1993). The cognizer’s innards: A psychological and philosophical perspective on the development of thought. Mind and Language, 8, 487–519] point that knowledge acquired by a connectionist network always remains “knowledge in the network rather than knowledge for the network”. That is, while connectionist networks may become exquisitely sensitive to regularities contained in their input–output environment, they never exhibit the ability to access and manipulate this knowledge as knowledge: The knowledge can only be expressed through performing the task upon which the network was trained; it remains forever embedded in the causal pathways that developed as a result of training. To address this issue, we present simulations in which two networks interact. The states of a first-order network trained to perform a simple categorization task become input to a second-order network trained either as an encoder or on another categorization task. Thus, the second-order network “observes” the states of the first-order network and has, in the first case, to reproduce these states on its output units, and in the second case, to use the states as cues in order to solve the secondary task. This implements a limited form of metarepresentation, to the extent that the second-order network’s internal representations become re-representations of the first-order network’s internal states. We conclude that this mechanism provides the beginnings of a computational mechanism to account for mental attitudes, that is, an understanding by a cognitive system of the manner in which its first-order knowledge is held (belief, hope, fear, etc.). Consciousness, in this light, thus involves knowledge of the geography of one own’s internal representations — a geography that is itself learned over time as a result of an agent’s attributing value to the various experiences it enjoys through interaction with itself, the world, and others.

A Fully CMOS Low-Cost Chaotic Neural Network

by Josep L. Rossello

Higher-Order Thoughts in Action: Consciousness as an unconscious redescription process

by Bert Timmermans

Timmermans B, Schilbach L, Pasquali A, & Cleeremans, A (2012) Higher-Order Thoughts in Action: Consciousness as an unconscious redescription process. Philosophical Transactions of the Royal Society B: Biological Sciences, 367(1594), 1412-23. (here only the final draft, as I'm not allowed to post the actual paper)

Metacognition is usually construed as a conscious, intentional process whereby people reflect upon their own mental... more Metacognition is usually construed as a conscious, intentional process whereby people reflect upon their own mental activity. Here, we instead suggest that metacognition is but an instance of a larger class of representational redescription processes that we assume occur unconsciously and automatically. From this perspective, the brain continuously and unconsciously learns to anticipate the consequences of action or activity on itself, on the world, and on other people through three predictive loops: An inner loop, a perception-action loop, and a self-other (social cognition) loop, which together form a tangled hierarchy. We ask what kinds of mechanisms may subtend this form of enactive metacognition. We extend previous neural network simulations and compare the model with Signal Detection Theory, highlighting that while the latter approach assumes that both Type I (objective) and Type II (subjective, metacognition-based) decisions tap into the same signal at different hierarchical levels, our approach is closer to dual-route models in that it assumes that the redescriptions made possible by the emergence of metarepresentations occur independently and outside of the first-order causal chain. We close by reviewing relevant neurological evidence for the idea that awareness, self-awareness and social cognition involve the same mechanisms.

Neural network technique applied to estimate hourly diffuse solar radiation in the city of São Paulo

by Jacyra Soares

An application of neural network technique to correct the dome temperature effects on pyrgeometer measurements

by Jacyra Soares

AN APPLICATION OF NEURAL NETWORK TECHNIQUE TO IMPROVE QUALITY OF METEOROLOGICAL MEASUREMENTS

by Jacyra Soares

ARTIFICIAL NEURAL NETWORKS AS A SUPPORTING TOOL FOR SOLVING METEOROLOGICAL AND AIR QUALITY MEASUREMENT/MODELING …

by Jacyra Soares

A computational feed-forward model predicts categorization of masked emotional body language for longer, but not for shorter latencies

by Bernard Stienen

Speech enhancement using microphone array neural switched Griffiths-Jim beamformer

by Vaitheki Yoganathan

Published in: 2010 International Conference on Wireless Communications and Signal Processing (WCSP)

There is a great need for speech enhancement in today's world due to the increasing demand for speech based... more There is a great need for speech enhancement in today's world due to the increasing demand for speech based applications. These applications vary from hearing-aids, hands-free telephony to speech controlled devices. The main goal is to minimize the interference from an acquired speech signal. The interference we considered here could be from any noise source such as competing speaker, radio, TV and so on. This paper proposes a solution to improve the current design of the switched Griffiths-Jim beamformer structure. It introduces an adaptive nonlinear neural network algorithm for the noise reduction section. The network topology used here is a partially connected three-layer feedforward neural network structure. The error backpropagation algorithm is used here as the learning algorithm. Comparison analysis of the traditional four channel linear beamformer and the proposed four-channel neural switched Griffiths-Jim beamformer structure is discussed here. They are both tested with different types of interference signal from the Noise-X database. All the experiments are conducted in real-world surrounding. The nonlinear approach introduced here shows remarkable improvement over the previous linear adaptive beamformer approach.

Multi-microphone adaptive neural switched Griffiths-Jim beamformer for noise reduction

by Vaitheki Yoganathan

Published in: 2010 IEEE 10th International Conference on Signal Processing (ICSP),

This paper proposes a novel multi-microphone nonlinear neural network based switched Griffiths-Jim beamformer... more This paper proposes a novel multi-microphone nonlinear neural network based switched Griffiths-Jim beamformer structure for speech enhancement. The main objective of this algorithm is to reduce real-world interference signals such as radio, television or computer fan noise from an acquired speech signal. The proposed algorithm improves the current design of the switched Griffiths-Jim beamformer structure by introducing an adaptive nonlinear neural network filter for the noise reduction section. The network topology used here is a partially connected three-layer feedforward neural network structure. The error backpropagation algorithm is used here as the learning algorithm. A comparison analysis of the traditional three-microphone linear beamformer and the proposed three-microphone neural switched Griffiths-Jim beamformer structure is discussed here. They are both tested with different types of interference signal from the Noise-X database. All the experiments are conducted in real-world surroundings. The nonlinear approach introduced here shows remarkable improvement over the previous linear adaptive beamformer approach.

Quantum logical neural networks

by Wilson de Oliveira

10th Brazilian Symposium on Neural Networks

Quantum analogues of the (classical) Logical Neural Networks (LNN) models are proposed in [6] (q-LNN for short). We... more Quantum analogues of the (classical) Logical Neural Networks (LNN) models are proposed in [6] (q-LNN for short). We shall here further develop and investigate the q- LNN composed of the quantum analogue of the Probabilis- tic Logic Node (PLN) and the multiple-valued PLN (MPLN) variations, dubbed q-PLN and q-MPLN respectively. Be- sides a clearer mathematical description, we present a com- putationally efficient and simply described quantum learn- ing algorithm in contrast to what has been proposed to the quantum weighted version.

Echo State Incremental Gaussian Mixture Network for Spatio-Temporal Pattern Processing

by Rafael Pinto

RECURSIVE INCREMENTAL GAUSSIAN MIXTURE NETWORK FOR SPATIO-TEMPORAL PATTERN PROCESSING

by Rafael Pinto

A Study for Prediction of Minerals in Rock Images using Back Propagation Neural Networks

by Imran Sarwar Bajwa

IEEE International Conference on Advances in Space Technologies (ICAST 2006) pp:185-189

This paper presents a novel approach for the segmentation of ground based images of rocks using back propagation... more This paper presents a novel approach for the segmentation of ground based images of rocks using back propagation neural network architecture. The designed system actually identifies the possible minerals by analyzing the surface color of the rocks. The rocks in Balochistan are very hard and defined. Such rocks are typically full of minerals. The rocks in the province of Balochistan are peculiar in their shape and surface colour. Usually, these colours are developed due to the reaction of the particles of the minerals with air. The upper layer of dust upon these rocks can be really useful in identifying the possible minerals concealing inside the rocks. The designed mechanism uses conventional artificial neural networks to identify various coloured parts of the rocks which are further classified into different minerals using histograms. The BPNN helps to learn to solve the task through a dynamic adaptation of its classification context. The designed system is trained by providing it the basic information related to the physical features of various mineral and types of rocks. The designed system highlights the various parts of the images by using various colours for various minerals.

Intelligent Paging Strategy for Multi-Carrier CDMA System

by Sheikh Shanawaz Mostafa

Sheikh Shanawaz Mostafa, Khondker Jahid Reza, Md. Ziaul Amin and Mohiuddin Ahmad, published in International Journal of Computer Science Issues, Vol. 8, Issue 6, No 1, pp. 254-260, November 2011.

Subscriber satisfaction and maximum radio resource utilization are the pivotal criteria in communication system... more Subscriber satisfaction and maximum radio resource utilization are the pivotal criteria in communication system design. In multi-Carrier CDMA system, different paging algorithms are used for locating user within the shortest possible time and best possible utilization of radio resources. Different paging algorithms underscored different techniques based on the different purposes. However, low servicing time of sequential search and better utilization of radio resources of concurrent search can be utilized simultaneously by swapping of the algorithms. In this paper, intelligent mechanism has been developed for dynamic algorithm assignment basing on time-varying traffic demand, which is predicted by radial basis neural network; and its performance has been analyzed are based on prediction efficiency of different types of data. High prediction efficiency is observed with a good correlation coefficient (0.99) and subsequently better performance is achieved by dynamic paging algorithm assignment. This claim is substantiated by the result of proposed intelligent paging strategy.

Supervised Hybrid SOM-NG Algorithm

by Mario J. Crespo-Ramos

The hybrid SOM-NG algorithm was formulated to improve the quantization precision in Self Organizing Maps by the means... more The hybrid SOM-NG algorithm was formulated to improve the quantization precision in Self Organizing Maps by the means of combine both SOM and Neural Gas properties using a parameter gamma to tune the topology preservation. A supervised learning algorithm is proposed to take advantage of the balanced hybrid algorithm. The proposed algorithm makes a linear approximation of the goal function for every Voronoi region. The algorithm gives good estimations and well balanced prototype positions combining the benefits of the original algorithms.