Recognition of partly occluded patterns: a neural network model

 Human beings are often able to read a letter or word partly occluded by contaminating ink stains. However, if the stains are completely erased and the occluded areas of the letter are changed to white, we usually have difficulty in reading the letter. In this article I propose a hypothesis explaining why a pattern is easier to recognize when it is occluded by visible objects than by invisible opaque objects. A neural network model is constructed based on this hypothesis. The visual system extracts various visual features from the input pattern and then attempts to recognize it. If the occluding objects are not visible, the visual system will have difficulty in distinguishing which features are relevant to the original pattern and which are newly generated by the occlusion. If the occluding objects are visible, however, the visual system can easily discriminate between relevant and irrelevant features and recognize the occluded pattern correctly. The proposed model is an extended version of the neocognitron model. The activity of the feature-extracting cells whose receptive fields cover the occluding objects is suppressed in an early stage of the hierarchical network. Since the irrelevant features generated by the occlusion are thus eliminated, the model can recognize occluded patterns correctly, provided the occlusion is not so large as to prevent recognition even by human beings. Received: 21 February 2000 / Accepted in revised form: 11 September 2000     

Behavioural and neurophysiological evidence for face identity and face emotion processing in animals

Visual cues from faces provide important social information relating to individual identity, sexual attraction and emotional state. Behavioural and neurophysiological studies on both monkeys and sheep have shown that specialized skills and neural systems for processing these complex cues to guide behaviour have evolved in a number of mammals and are not present exclusively in humans. Indeed, there are remarkable similarities in the ways that faces are processed by the brain in humans and other mammalian species. While human studies with brain imaging and gross neurophysiological recording approaches have revealed global aspects of the face-processing network, they cannot investigate how information is encoded by specific neural networks. Single neuron electrophysiological recording approaches in both monkeys and sheep have, however, provided some insights into the neural encoding principles involved and, particularly, the presence of a remarkable degree of high-level encoding even at the level of a specific face. Recent developments that allow simultaneous recordings to be made from many hundreds of individual neurons are also beginning to reveal evidence for global aspects of a population-based code. This review will summarize what we have learned so far from these animal-based studies about the way the mammalian brain processes the faces and the emotions they can communicate, as well as associated capacities such as how identity and emotion cues are dissociated and how face imagery might be generated. It will also try to highlight what questions and advances in knowledge still challenge us in order to provide a complete understanding of just how brain networks perform this complex and important social recognition task.     

Estrogens and cognition: Friends or foes?: An evaluation of the opposing effects of estrogens on learning and memory

This article is part of a Special Issue “Estradiol and cognition”.Estrogens are becoming well known for their robust enhancement on cognition particularly for learning and memory that relies upon functioning of the hippocampus and related neural systems. What is also emerging is that estrogen modulation of cognition is not uniform, at times enhancing yet at other times impairing learning. This review explores the bidirectional effects of estrogens on learning from a multiple memory systems view, focusing on the hippocampus and striatum, whereby modulation by estrogens sorts according to task attributes and neural systems engaged during cognition. We highlight our findings showing that the ability to solve hippocampus-sensitive tasks typically improves under relatively high estrogen status while the ability to solve striatum-sensitive tasks degrades with estrogen exposures. Though constrained by dose and timing of exposure, these opposing enhancements and impairments of cognition can be observed following treatments with different estrogenic compounds including the hormone estradiol, the isoflavone genistein found in soybeans, and agonists that are selective for specific estrogen receptors, suggesting that activation of a single receptor type is sufficient to produce the observed shifts in learning strategies. Using this multi-dimensional framework will allow us to extend our thinking of the relationship between estrogens and cognition to other brain regions and cognitive functions.     

On the Relationships between Generative Encodings,Regularity, and Learning Abilities when Evolving Plastic Artificial Neural Networks

A major goal of bio-inspired artificial intelligence is to design artificial neural networks with abilities that resemble those of animal nervous systems. It is commonly believed that two keys for evolving nature-like artificial neural networks are (1) the developmental process that links genes to nervous systems, which enables the evolution of large, regular neural networks, and (2) synaptic plasticity, which allows neural networks to change during their lifetime. So far, these two topics have been mainly studied separately. The present paper shows that they are actually deeply connected. Using a simple operant conditioning task and a classic evolutionary algorithm, we compare three ways to encode plastic neural networks: a direct encoding, a developmental encoding inspired by computational neuroscience models, and a developmental encoding inspired by morphogen gradients (similar to HyperNEAT). Our results suggest that using a developmental encoding could improve the learning abilities of evolved, plastic neural networks. Complementary experiments reveal that this result is likely the consequence of the bias of developmental encodings towards regular structures: (1) in our experimental setup, encodings that tend to produce more regular networks yield networks with better general learning abilities; (2) whatever the encoding is, networks that are the more regular are statistically those that have the best learning abilities.     

Effect of Noise in Intelligent Cellular Decision Making

Similar to intelligent multicellular neural networks controlling human brains, even single cells, surprisingly, are able to make intelligent decisions to classify several external stimuli or to associate them. This happens because of the fact that gene regulatory networks can perform as perceptrons, simple intelligent schemes known from studies on Artificial Intelligence. We study the role of genetic noise in intelligent decision making at the genetic level and show that noise can play a constructive role helping cells to make a proper decision. We show this using the example of a simple genetic classifier able to classify two external stimuli.     

From Birdsong to Human Speech Recognition: Bayesian Inference on a Hierarchy of Nonlinear Dynamical Systems

Our knowledge about the computational mechanisms underlying human learning and recognition of sound sequences, especially speech, is still very limited. One difficulty in deciphering the exact means by which humans recognize speech is that there are scarce experimental findings at a neuronal, microscopic level. Here, we show that our neuronal-computational understanding of speech learning and recognition may be vastly improved by looking at an animal model, i.e., the songbird, which faces the same challenge as humans: to learn and decode complex auditory input, in an online fashion. Motivated by striking similarities between the human and songbird neural recognition systems at the macroscopic level, we assumed that the human brain uses the same computational principles at a microscopic level and translated a birdsong model into a novel human sound learning and recognition model with an emphasis on speech. We show that the resulting Bayesian model with a hierarchy of nonlinear dynamical systems can learn speech samples such as words rapidly and recognize them robustly, even in adverse conditions. In addition, we show that recognition can be performed even when words are spoken by different speakers and with different accents—an everyday situation in which current state-of-the-art speech recognition models often fail. The model can also be used to qualitatively explain behavioral data on human speech learning and derive predictions for future experiments.     

Knowledge extraction from evolving spiking neural networks with rank order population coding

This paper demonstrates how knowledge can be extracted from evolving spiking neural networks with rank order population coding. Knowledge discovery is a very important feature of intelligent systems. Yet, a disproportionally small amount of research is centered on the issue of knowledge extraction from spiking neural networks which are considered to be the third generation of artificial neural networks. The lack of knowledge representation compatibility is becoming a major detriment to end users of these networks. We show that a high-level knowledge can be obtained from evolving spiking neural networks. More specifically, we propose a method for fuzzy rule extraction from an evolving spiking network with rank order population coding. The proposed method was used for knowledge discovery on two benchmark taste recognition problems where the knowledge learnt by an evolving spiking neural network was extracted in the form of zero-order Takagi-Sugeno fuzzy IF-THEN rules.     

Discrete Neural Correlates for the Recognition of Negative Emotions: Insights from Frontotemporal Dementia

Patients with frontotemporal dementia have pervasive changes in emotion recognition and social cognition, yet the neural changes underlying these emotion processing deficits remain unclear. The multimodal system model of emotion proposes that basic emotions are dependent on distinct brain regions, which undergo significant pathological changes in frontotemporal dementia. As such, this syndrome may provide important insight into the impact of neural network degeneration upon the innate ability to recognise emotions. This study used voxel-based morphometry to identify discrete neural correlates involved in the recognition of basic emotions (anger, disgust, fear, sadness, surprise and happiness) in frontotemporal dementia. Forty frontotemporal dementia patients (18 behavioural-variant, 11 semantic dementia, 11 progressive nonfluent aphasia) and 27 healthy controls were tested on two facial emotion recognition tasks: The Ekman 60 and Ekman Caricatures. Although each frontotemporal dementia group showed impaired recognition of negative emotions, distinct associations between emotion-specific task performance and changes in grey matter intensity emerged. Fear recognition was associated with the right amygdala; disgust recognition with the left insula; anger recognition with the left middle and superior temporal gyrus; and sadness recognition with the left subcallosal cingulate, indicating that discrete neural substrates are necessary for emotion recognition in frontotemporal dementia. The erosion of emotion-specific neural networks in neurodegenerative disorders may produce distinct profiles of performance that are relevant to understanding the neurobiological basis of emotion processing.     

Animal Detection Precedes Access to Scene Category

The processes underlying object recognition are fundamental for the understanding of visual perception. Humans can recognize many objects rapidly even in complex scenes, a task that still presents major challenges for computer vision systems. A common experimental demonstration of this ability is the rapid animal detection protocol, where human participants earliest responses to report the presence/absence of animals in natural scenes are observed at 250–270 ms latencies. One of the hypotheses to account for such speed is that people would not actually recognize an animal per se, but rather base their decision on global scene statistics. These global statistics (also referred to as spatial envelope or gist) have been shown to be computationally easy to process and could thus be used as a proxy for coarse object recognition. Here, using a saccadic choice task, which allows us to investigate a previously inaccessible temporal window of visual processing, we showed that animal – but not vehicle – detection clearly precedes scene categorization. This asynchrony is in addition validated by a late contextual modulation of animal detection, starting simultaneously with the availability of scene category. Interestingly, the advantage for animal over scene categorization is in opposition to the results of simulations using standard computational models. Taken together, these results challenge the idea that rapid animal detection might be based on early access of global scene statistics, and rather suggests a process based on the extraction of specific local complex features that might be hardwired in the visual system.     

Human activity recognition based on Evolving Fuzzy Systems

Environments equipped with intelligent sensors can be of much help if they can recognize the actions or activities of their users. If this activity recognition is done automatically, it can be very useful for different tasks such as future action prediction, remote health monitoring, or interventions. Although there are several approaches for recognizing activities, most of them do not consider the changes in how a human performs a specific activity. We present an automated approach to recognize daily activities from the sensor readings of an intelligent home environment. However, as the way to perform an activity is usually not fixed but it changes and evolves, we propose an activity recognition method based on Evolving Fuzzy Systems.     

Neural signatures for sustaining object representations attributed to others in preverbal human infants

A major feat of social beings is to encode what their conspecifics see, know or believe. While various non-human animals show precursors of these abilities, humans perform uniquely sophisticated inferences about other people''s mental states. However, it is still unclear how these possibly human-specific capacities develop and whether preverbal infants, similarly to adults, form representations of other agents'' mental states, specifically metarepresentations. We explored the neurocognitive bases of eight-month-olds'' ability to encode the world from another person''s perspective, using gamma-band electroencephalographic activity over the temporal lobes, an established neural signature for sustained object representation after occlusion. We observed such gamma-band activity when an object was occluded from the infants'' perspective, as well as when it was occluded only from the other person (study 1), and also when subsequently the object disappeared, but the person falsely believed the object to be present (study 2). These findings suggest that the cognitive systems involved in representing the world from infants'' own perspective are also recruited for encoding others'' beliefs. Such results point to an early-developing, powerful apparatus suitable to deal with multiple concurrent representations, and suggest that infants can have a metarepresentational understanding of other minds even before the onset of language.     

Brain Systems for Probabilistic and Dynamic Prediction: Computational Specificity and Integration

A computational approach to functional specialization suggests that brain systems can be characterized in terms of the types of computations they perform, rather than their sensory or behavioral domains. We contrasted the neural systems associated with two computationally distinct forms of predictive model: a reinforcement-learning model of the environment obtained through experience with discrete events, and continuous dynamic forward modeling. By manipulating the precision with which each type of prediction could be used, we caused participants to shift computational strategies within a single spatial prediction task. Hence (using fMRI) we showed that activity in two brain systems (typically associated with reward learning and motor control) could be dissociated in terms of the forms of computations that were performed there, even when both systems were used to make parallel predictions of the same event. A region in parietal cortex, which was sensitive to the divergence between the predictions of the models and anatomically connected to both computational networks, is proposed to mediate integration of the two predictive modes to produce a single behavioral output.     

The slow process: A hypothetical cognitive adaptation for distributed cognitive networks

Human cognitive evolution is characterized by two special features that are truly novel in the primate line. The first is the emergence of "mindsharing" cultures that perform cooperative cognitive work, and serve as distributed cognitive networks. The second is the emergence of a brain that is specifically adapted for functioning within those distributed networks, and cannot realize its design potential without them. This paper proposes a hypothetical neural process at the core of this brain adaptation, called the "slow process". It enables the human brain to comprehend social events of much longer duration and complexity than those that characterize primate social life. It runs in the background of human cognitive life, with the faster moving sensorimotor interface running in the foreground. Most mammals can integrate events in the shorter time zone that corresponds to working memory. However, very few can comprehend complex events that extend over several hours (for example, a game or conversation) in what may be called the "intermediate" time zone. Adult humans typically live, plan, and imagine their lives in this time range, which seems to exceed the capabilities of our closest relatives, bonobos and chimpanzees. In summary, human cognition has both an individual and a collective dimension. Individual brains and minds function within cognitive-cultural networks, or CCNs, that store and transmit knowledge. The human brain relies on cultural input even to develop the basic cognitive capacities needed to gain access to that knowledge in the first place. The postulated slow process is a top-down executive capacity that evolved specifically to manage the cultural connection, and handle the cognitive demands imposed by increasingly complex distributed systems.     

Neural mechanisms of auditory species recognition in birds

Auditory communication in humans and other animals frequently takes place in noisy environments with many co‐occurring signallers. Receivers are thus challenged to rapidly recognize salient auditory signals and filter out irrelevant sounds. Most bird species produce a variety of complex vocalizations that function to communicate with other members of their own species and behavioural evidence broadly supports preferences for conspecific over heterospecific sounds (auditory species recognition). However, it remains unclear whether such auditory signals are categorically recognized by the sensory and central nervous system. Here, we review 53 published studies that compare avian neural responses between conspecific versus heterospecific vocalizations. Irrespective of the techniques used to characterize neural activity, distinct nuclei of the auditory forebrain are consistently shown to be repeatedly conspecific selective across taxa, even in response to unfamiliar individuals with distinct acoustic properties. Yet, species‐specific neural discrimination is not a stereotyped auditory response, but is modulated according to its salience depending, for example, on ontogenetic exposure to conspecific versus heterospecific stimuli. Neuromodulators, in particular norepinephrine, may mediate species recognition by regulating the accuracy of neuronal coding for salient conspecific stimuli. Our review lends strong support for neural structures that categorically recognize conspecific signals despite the highly variable physical properties of the stimulus. The available data are in support of a ‘perceptual filter’‐based mechanism to determine the saliency of the signal, in that species identity and social experience combine to influence the neural processing of species‐specific auditory stimuli. Finally, we present hypotheses and their testable predictions, to propose next steps in species‐recognition research into the emerging model of the neural conceptual construct in avian auditory recognition.     

Deep Neural Networks Rival the Representation of Primate IT Cortex for Core Visual Object Recognition

The primate visual system achieves remarkable visual object recognition performance even in brief presentations, and under changes to object exemplar, geometric transformations, and background variation (a.k.a. core visual object recognition). This remarkable performance is mediated by the representation formed in inferior temporal (IT) cortex. In parallel, recent advances in machine learning have led to ever higher performing models of object recognition using artificial deep neural networks (DNNs). It remains unclear, however, whether the representational performance of DNNs rivals that of the brain. To accurately produce such a comparison, a major difficulty has been a unifying metric that accounts for experimental limitations, such as the amount of noise, the number of neural recording sites, and the number of trials, and computational limitations, such as the complexity of the decoding classifier and the number of classifier training examples. In this work, we perform a direct comparison that corrects for these experimental limitations and computational considerations. As part of our methodology, we propose an extension of “kernel analysis” that measures the generalization accuracy as a function of representational complexity. Our evaluations show that, unlike previous bio-inspired models, the latest DNNs rival the representational performance of IT cortex on this visual object recognition task. Furthermore, we show that models that perform well on measures of representational performance also perform well on measures of representational similarity to IT, and on measures of predicting individual IT multi-unit responses. Whether these DNNs rely on computational mechanisms similar to the primate visual system is yet to be determined, but, unlike all previous bio-inspired models, that possibility cannot be ruled out merely on representational performance grounds.     

A comparative study of electronic and neural networks involved in pattern recognition

The comparative study of electronic and neural networks involved in pattern recognition starts with the analogies of structure and function which exist between the electronic “basic integrative unit” and the neuron. Both elements represent the basic components in each system of networks and may be considered as functionally equivalent.According to the kind of response given to a standard input signal, four types of integrative units, either electronic or neural, may be distinguished: the fixed, the accommodative, the signal prolongating and the adaptive type.The integrative units perform many different functions. Those involved in pattern recognition, however, can all be grouped into three categories according to one of the following functions they perform: contrast detection, pattern detection and pattern discrimination. A “contrast detecting unit” gives responses in two senses, positive or negative, according to the position of the stimulus over its receptive field. A “pattern detecting unit” gives responses in one sense only, with a maximum for a pattern having the spatial distribution corresponding to the positive acting receptors of its receptive field. For performing the function of discrimination, which leads to reliable identification of any pattern, a network arrangement called a “maximum amplitude filter” is necessary. Examples of such units and arrangements existing in the nervous system are provided.It is concluded that a “logical analysis of neural networks” based on engineering principles is possible and that this could provide a new tool to the neurophysiologist in the study of the nervous system.     

Cell‐cell recognition and social networking in bacteria

The ability to recognize self and to recognize partnering cells allows microorganisms to build social networks that perform functions beyond the capabilities of the individual. In bacteria, recognition typically involves genetic determinants that provide cell surface receptors or diffusible signalling chemicals to identify proximal cells at the molecular level that can participate in cooperative processes. Social networks also rely on discriminating mechanisms to exclude competing cells from joining and exploiting their groups. In addition to their appropriate genotypes, cell‐cell recognition also requires compatible phenotypes, which vary according to environmental cues or exposures as well as stochastic processes that lead to heterogeneity and potential disharmony in the population. Understanding how bacteria identify their social partners and how they synchronize their behaviours to conduct multicellular functions is an expanding field of research. Here, we review recent progress in the field and contrast the various strategies used in recognition and behavioural networking.     

Recognition of distantly related protein sequences using conserved motifs and neural networks.

A sensitive technique for protein sequence motif recognition based on neural networks has been developed. It involves three major steps. (1) At each appropriate alignment position of a set of N matched sequences, a set of N aligned oligopeptides is specified with preselected window length. N neural nets are subsequently and successively trained on N-1 amino acid spans after eliminating each ith oligopeptide. A test for recognition of each of the ith spans is performed. The average neural net recognition over N such trials is used as a measure of conservation for the particular windowed region of the multiple alignment. This process is repeated for all possible spans of given length in the multiple alignment. (2) The M most conserved regions are regarded as motifs and the oligopeptides within each are used to train intensively M individual neural networks. (3) The M networks are then applied in a search for related primary structures in a databank of known protein sequences. The oligopeptide spans in the database sequence with strongest neural net output for each of the M networks are saved and then scored according to the output signals and the proper combination that follows the expected N- to C-terminal sequence order. The motifs from the database with highest similarity scores can then be used to retrain the M neural nets, which can be subsequently utilized for further searches in the databank, thus providing even greater sensitivity to recognize distant familial proteins. This technique was successfully applied to the integrase, DNA-polymerase and immunoglobulin families.