Adaptive Resonance Theory: How a brain learns to consciously attend, learn, and recognize a changing world

FedART: A neural model integrating federated learning and adaptive resonance theory

Federated Learning (FL) has emerged as a promising paradigm for collaborative model training across distributed clients while preserving data privacy. However, prevailing FL approaches aggregate the clients' local models into a global model through multi-round iterative parameter averaging. This leads to the undesirable bias of the aggregated model towards certain clients in the presence of heterogeneous data distributions among the clients. Moreover, such approaches are restricted to supervised classification tasks and do not support unsupervised clustering. To address these limitations, we propose a novel one-shot FL approach called Federated Adaptive Resonance Theory (FedART) which leverages self-organizing Adaptive Resonance Theory (ART) models to learn category codes, where each code represents a cluster of similar data samples. In FedART, the clients learn to associate their private data with various local category codes. Under heterogeneity, the local codes across different clients represent heterogeneous data. In turn, a global model takes these local codes as inputs and aggregates them into global category codes, wherein heterogeneous client data is indirectly represented by distinctly encoded global codes, in contrast to the averaging out of parameters in the existing approaches. This enables the learned global model to handle heterogeneous data. In addition, FedART employs a universal learning mechanism to support both federated classification and clustering tasks. Our experiments conducted on various federated classification and clustering tasks show that FedART consistently outperforms state-of-the-art FL methods on data with heterogeneous distribution across clients.

Data Generation for Machine Learning Interatomic Potentials and Beyond

The field of data-driven chemistry is undergoing an evolution, driven by innovations in machine learning models for predicting molecular properties and behavior. Recent strides in ML-based interatomic potentials have paved the way for accurate modeling of diverse chemical and structural properties at the atomic level. The key determinant defining MLIP reliability remains the quality of the training data. A paramount challenge lies in constructing training sets that capture specific domains in the vast chemical and structural space. This Review navigates the intricate landscape of essential components and integrity of training data that ensure the extensibility and transferability of the resulting models. We delve into the details of active learning, discussing its various facets and implementations. We outline different types of uncertainty quantification applied to atomistic data acquisition and the correlations between estimated uncertainty and true error. The role of atomistic data samplers in generating diverse and informative structures is highlighted. Furthermore, we discuss data acquisition via modified and surrogate potential energy surfaces as an innovative approach to diversify training data. The Review also provides a list of publicly available data sets that cover essential domains of chemical space.

Causal inference predicts the transition from integration to segmentation in motion perception

Motion provides a powerful sensory cue for segmenting a visual scene into objects and inferring the causal relationships between objects. Fundamental mechanisms involved in this process are the integration and segmentation of local motion signals. However, the computations that govern whether local motion signals are perceptually integrated or segmented remain unclear. Hierarchical Bayesian causal inference has recently been proposed as a model for these computations, yet a hallmark prediction of the model - its dependency on sensory uncertainty - has remained untested. We used a recently developed hierarchical stimulus configuration to measure how human subjects integrate or segment local motion signals while manipulating motion coherence to control sensory uncertainty. We found that (a) the perceptual transition from motion integration to segmentation shifts with sensory uncertainty, and (b) that perceptual variability is maximal around this transition point. Both findings were predicted by the model and challenge conventional interpretations of motion repulsion effects.

A dendritic mechanism for balancing synaptic flexibility and stability

Biological and artificial neural networks learn by modifying synaptic weights, but it is unclear how these systems retain previous knowledge and also acquire new information. Here, we show that cortical pyramidal neurons can solve this plasticity-versus-stability dilemma by differentially regulating synaptic plasticity at distinct dendritic compartments. Oblique dendrites of adult mouse layer 5 cortical pyramidal neurons selectively receive monosynaptic thalamic input, integrate linearly, and lack burst-timing synaptic potentiation. In contrast, basal dendrites, which do not receive thalamic input, exhibit conventional NMDA receptor (NMDAR)-mediated supralinear integration and synaptic potentiation. Congruently, spiny synapses on oblique branches show decreased structural plasticity in vivo. The selective decline in NMDAR activity and expression at synapses on oblique dendrites is controlled by a critical period of visual experience. Our results demonstrate a biological mechanism for how single neurons can safeguard a set of inputs from ongoing plasticity by altering synaptic properties at distinct dendritic domains.

A landscape of consciousness: Toward a taxonomy of explanations and implications

Diverse explanations or theories of consciousness are arrayed on a roughly physicalist-to-nonphysicalist landscape of essences and mechanisms. Categories: Materialism Theories (philosophical, neurobiological, electromagnetic field, computational and informational, homeostatic and affective, embodied and enactive, relational, representational, language, phylogenetic evolution); Non-Reductive Physicalism; Quantum Theories; Integrated Information Theory; Panpsychisms; Monisms; Dualisms; Idealisms; Anomalous and Altered States Theories; Challenge Theories. There are many subcategories, especially for Materialism Theories. Each explanation is self-described by its adherents, critique is minimal and only for clarification, and there is no attempt to adjudicate among theories. The implications of consciousness explanations or theories are assessed with respect to four questions: meaning/purpose/value (if any); AI consciousness; virtual immortality; and survival beyond death. A Landscape of Consciousness, I suggest, offers perspective.

Bioinspired Control Architecture for Adaptive and Resilient Navigation of Unmanned Underwater Vehicle in Monitoring Missions of Submerged Aquatic Vegetation Meadows

Submerged aquatic vegetation plays a fundamental role as a habitat for the biodiversity of marine species. To carry out the research and monitoring of submerged aquatic vegetation more efficiently and accurately, it is important to use advanced technologies such as underwater robots. However, when conducting underwater missions to capture photographs and videos near submerged aquatic vegetation meadows, algae can become entangled in the propellers and cause vehicle failure. In this context, a neurobiologically inspired control architecture is proposed for the control of unmanned underwater vehicles with redundant thrusters. The proposed control architecture learns to control the underwater robot in a non-stationary environment and combines the associative learning method and vector associative map learning to generate transformations between the spatial and velocity coordinates in the robot actuator. The experimental results obtained show that the proposed control architecture exhibits notable resilience capabilities while maintaining its operation in the face of thruster failures. In the discussion of the results obtained, the importance of the proposed control architecture is highlighted in the context of the monitoring and conservation of underwater vegetation meadows. Its resilience, robustness, and adaptability capabilities make it an effective tool to face challenges and meet mission objectives in such critical environments.

An operating principle of the cerebral cortex, and a cellular mechanism for attentional trial-and-error pattern learning and useful classification extraction

A feature of the brains of intelligent animals is the ability to learn to respond to an ensemble of active neuronal inputs with a behaviorally appropriate ensemble of active neuronal outputs. Previously, a hypothesis was proposed on how this mechanism is implemented at the cellular level within the neocortical pyramidal neuron: the apical tuft or perisomatic inputs initiate "guess" neuron firings, while the basal dendrites identify input patterns based on excited synaptic clusters, with the cluster excitation strength adjusted based on reward feedback. This simple mechanism allows neurons to learn to classify their inputs in a surprisingly intelligent manner. Here, we revise and extend this hypothesis. We modify synaptic plasticity rules to align with behavioral time scale synaptic plasticity (BTSP) observed in hippocampal area CA1, making the framework more biophysically and behaviorally plausible. The neurons for the guess firings are selected in a voluntary manner via feedback connections to apical tufts in the neocortical layer 1, leading to dendritic Ca2+ spikes with burst firing, which are postulated to be neural correlates of attentional, aware processing. Once learned, the neuronal input classification is executed without voluntary or conscious control, enabling hierarchical incremental learning of classifications that is effective in our inherently classifiable world. In addition to voluntary, we propose that pyramidal neuron burst firing can be involuntary, also initiated via apical tuft inputs, drawing attention toward important cues such as novelty and noxious stimuli. We classify the excitations of neocortical pyramidal neurons into four categories based on their excitation pathway: attentional versus automatic and voluntary/acquired versus involuntary. Additionally, we hypothesize that dendrites within pyramidal neuron minicolumn bundles are coupled via depolarization cross-induction, enabling minicolumn functions such as the creation of powerful hierarchical "hyperneurons" and the internal representation of the external world. We suggest building blocks to extend the microcircuit theory to network-level processing, which, interestingly, yields variants resembling the artificial neural networks currently in use. On a more speculative note, we conjecture that principles of intelligence in universes governed by certain types of physical laws might resemble ours.

Human Crowds as Social Networks: Collective Dynamics of Consensus and Polarization

A ubiquitous type of collective behavior and decision-making is the coordinated motion of bird flocks, fish schools, and human crowds. Collective decisions to move in the same direction, turn right or left, or split into subgroups arise in a self-organized fashion from local interactions between individuals without central plans or designated leaders. Strikingly similar phenomena of consensus (collective motion), clustering (subgroup formation), and bipolarization (splitting into extreme groups) are also observed in opinion formation. As we developed models of crowd dynamics and analyzed crowd networks, we found ourselves going down the same path as models of opinion dynamics in social networks. In this article, we draw out the parallels between human crowds and social networks. We show that models of crowd dynamics and opinion dynamics have a similar mathematical form and generate analogous phenomena in multiagent simulations. We suggest that they can be unified by a common collective dynamics, which may be extended to other psychological collectives. Models of collective dynamics thus offer a means to account for collective behavior and collective decisions without appealing to a priori mental structures.

Analyzing Biomedical Datasets with Symbolic Tree Adaptive Resonance Theory

Biomedical datasets distill many mechanisms of human diseases, linking diseases to genes and phenotypes (signs and symptoms of disease), genetic mutations to altered protein structures, and altered proteins to changes in molecular functions and biological processes. It is desirable to gain new insights from these data, especially with regard to the uncovering of hierarchical structures relating disease variants. However, analysis to this end has proven difficult due to the complexity of the connections between multi-categorical symbolic data. This article proposes symbolic tree adaptive resonance theory (START), with additional supervised, dual-vigilance (DV-START), and distributed dual-vigilance (DDV-START) formulations, for the clustering of multi-categorical symbolic data from biomedical datasets by demonstrating its utility in clustering variants of Charcot–Marie–Tooth disease using genomic, phenotypic, and proteomic data.

CL3: Generalization of Contrastive Loss for Lifelong Learning

Lifelong learning portrays learning gradually in nonstationary environments and emulates the process of human learning, which is efficient, robust, and able to learn new concepts incrementally from sequential experience. To equip neural networks with such a capability, one needs to overcome the problem of catastrophic forgetting, the phenomenon of forgetting past knowledge while learning new concepts. In this work, we propose a novel knowledge distillation algorithm that makes use of contrastive learning to help a neural network to preserve its past knowledge while learning from a series of tasks. Our proposed generalized form of contrastive distillation strategy tackles catastrophic forgetting of old knowledge, and minimizes semantic drift by maintaining a similar embedding space, as well as ensures compactness in feature distribution to accommodate novel tasks in a current model. Our comprehensive study shows that our method achieves improved performances in the challenging class-incremental, task-incremental, and domain-incremental learning for supervised scenarios.

Is predictive coding falsifiable?

Predictive-coding has justifiably become a highly influential theory in Neuroscience. However, the possibility of its unfalsifiability has been raised. We argue that if predictive-coding were unfalsifiable, it would be a problem, but there are patterns of behavioural and neuroimaging data that would stand against predictive-coding. Contra (vanilla) predictive patterns are those in which the more expected stimulus generates the largest evoked-response. However, basic formulations of predictive-coding mandate that an expected stimulus should generate little, if any, prediction error and thus little, if any, evoked-response. It has, though, been argued that contra (vanilla) predictive patterns can be obtained if precision is higher for expected stimuli. Certainly, using precision, one can increase the amplitude of an evoked-response, turning a predictive into a contra (vanilla) predictive pattern. We demonstrate that, while this is true, it does not present an absolute barrier to falsification. This is because increasing precision also reduces latency and increases the frequency of the response. These properties can be used to determine whether precision-weighting in predictive-coding justifiably explains a contra (vanilla) predictive pattern, ensuring that predictive-coding is falsifiable.

Information-theoretic analyses of neural data to minimize the effect of researchers' assumptions in predictive coding studies

Studies investigating neural information processing often implicitly ask both, which processing strategy out of several alternatives is used and how this strategy is implemented in neural dynamics. A prime example are studies on predictive coding. These often ask whether confirmed predictions about inputs or prediction errors between internal predictions and inputs are passed on in a hierarchical neural system-while at the same time looking for the neural correlates of coding for errors and predictions. If we do not know exactly what a neural system predicts at any given moment, this results in a circular analysis-as has been criticized correctly. To circumvent such circular analysis, we propose to express information processing strategies (such as predictive coding) by local information-theoretic quantities, such that they can be estimated directly from neural data. We demonstrate our approach by investigating two opposing accounts of predictive coding-like processing strategies, where we quantify the building blocks of predictive coding, namely predictability of inputs and transfer of information, by local active information storage and local transfer entropy. We define testable hypotheses on the relationship of both quantities, allowing us to identify which of the assumed strategies was used. We demonstrate our approach on spiking data collected from the retinogeniculate synapse of the cat (N = 16). Applying our local information dynamics framework, we are able to show that the synapse codes for predictable rather than surprising input. To support our findings, we estimate quantities applied in the partial information decomposition framework, which allow to differentiate whether the transferred information is primarily bottom-up sensory input or information transferred conditionally on the current state of the synapse. Supporting our local information-theoretic results, we find that the synapse preferentially transfers bottom-up information.

Deep learning in preclinical antibody drug discovery and development

Antibody drugs have become a key part of biotherapeutics. Patients suffering from various diseases have benefited from antibody therapies. However, its development process is rather long, expensive and risky. To speed up the process, reduce cost and improve success rate, artificial intelligence, especially deep learning methods, have been widely used in all aspects of preclinical antibody drug development, from library generation to hit identification, developability screening, lead selection and optimization. In this review, we systematically summarize antibody encodings, deep learning architectures and models used in preclinical antibody drug discovery and development. We also critically discuss challenges and opportunities, problems and possible solutions, current applications and future directions of deep learning in antibody drug development.

VLAD: Task-agnostic VAE-based lifelong anomaly detection

Lifelong learning represents an emerging machine learning paradigm that aims at designing new methods providing accurate analyses in complex and dynamic real-world environments. Although a significant amount of research has been conducted in image classification and reinforcement learning, very limited work has been done to solve lifelong anomaly detection problems. In this context, a successful method has to detect anomalies while adapting to changing environments and preserving knowledge to avoid catastrophic forgetting. While state-of-the-art online anomaly detection methods are able to detect anomalies and adapt to a changing environment, they are not designed to preserve past knowledge. On the other hand, while lifelong learning methods are focused on adapting to changing environments and preserving knowledge, they are not tailored for detecting anomalies, and often require task labels or task boundaries which are not available in task-agnostic lifelong anomaly detection scenarios. This paper proposes VLAD, a novel VAE-based Lifelong Anomaly Detection method addressing all these challenges simultaneously in complex task-agnostic scenarios. VLAD leverages the combination of lifelong change point detection and an effective model update strategy supported by experience replay with a hierarchical memory maintained by means of consolidation and summarization. An extensive quantitative evaluation showcases the merit of the proposed method in a variety of applied settings. VLAD outperforms state-of-the-art methods for anomaly detection, presenting increased robustness and performance in complex lifelong settings.

Hybrid predictive coding: Inferring, fast and slow

Predictive coding is an influential model of cortical neural activity. It proposes that perceptual beliefs are furnished by sequentially minimising "prediction errors"-the differences between predicted and observed data. Implicit in this proposal is the idea that successful perception requires multiple cycles of neural activity. This is at odds with evidence that several aspects of visual perception-including complex forms of object recognition-arise from an initial "feedforward sweep" that occurs on fast timescales which preclude substantial recurrent activity. Here, we propose that the feedforward sweep can be understood as performing amortized inference (applying a learned function that maps directly from data to beliefs) and recurrent processing can be understood as performing iterative inference (sequentially updating neural activity in order to improve the accuracy of beliefs). We propose a hybrid predictive coding network that combines both iterative and amortized inference in a principled manner by describing both in terms of a dual optimization of a single objective function. We show that the resulting scheme can be implemented in a biologically plausible neural architecture that approximates Bayesian inference utilising local Hebbian update rules. We demonstrate that our hybrid predictive coding model combines the benefits of both amortized and iterative inference-obtaining rapid and computationally cheap perceptual inference for familiar data while maintaining the context-sensitivity, precision, and sample efficiency of iterative inference schemes. Moreover, we show how our model is inherently sensitive to its uncertainty and adaptively balances iterative and amortized inference to obtain accurate beliefs using minimum computational expense. Hybrid predictive coding offers a new perspective on the functional relevance of the feedforward and recurrent activity observed during visual perception and offers novel insights into distinct aspects of visual phenomenology.

Lifelong Machine Learning Potentials

Machine learning potentials (MLPs) trained on accurate quantum chemical data can retain the high accuracy, while inflicting little computational demands. On the downside, they need to be trained for each individual system. In recent years, a vast number of MLPs have been trained from scratch because learning additional data typically requires retraining on all data to not forget previously acquired knowledge. Additionally, most common structural descriptors of MLPs cannot represent efficiently a large number of different chemical elements. In this work, we tackle these problems by introducing element-embracing atom-centered symmetry functions (eeACSFs), which combine structural properties and element information from the periodic table. These eeACSFs are key for our development of a lifelong machine learning potential (lMLP). Uncertainty quantification can be exploited to transgress a fixed, pretrained MLP to arrive at a continuously adapting lMLP, because a predefined level of accuracy can be ensured. To extend the applicability of an lMLP to new systems, we apply continual learning strategies to enable autonomous and on-the-fly training on a continuous stream of new data. For the training of deep neural networks, we propose the continual resilient (CoRe) optimizer and incremental learning strategies relying on rehearsal of data, regularization of parameters, and the architecture of the model.

Where is the mind within the brain? Transient selection of subnetworks by metabotropic receptors and G protein-gated ion channels

Perhaps the most important question posed by brain research is: How the brain gives rise to the mind. To answer this question, we have primarily relied on the connectionist paradigm: The brain's entire knowledge and thinking skills are thought to be stored in the connections; and the mental operations are executed by network computations. I propose here an alternative paradigm: Our knowledge and skills are stored in metabotropic receptors (MRs) and the G protein-gated ion channels (GPGICs). Here, mental operations are assumed to be executed by the functions of MRs and GPGICs. As GPGICs have the capacity to close or open branches of dendritic trees and axon terminals, their states transiently re-route neural activity throughout the nervous system. First, MRs detect ligands that signal the need to activate GPGICs. Next, GPGICs transiently select a subnetwork within the brain. The process of selecting this new subnetwork is what constitutes a mental operation - be it in a form of directed attention, perception or making a decision. Synaptic connections and network computations play only a secondary role, supporting MRs and GPGICs. According to this new paradigm, the mind emerges within the brain as the function of MRs and GPGICs whose primary function is to continually select the pathways over which neural activity will be allowed to pass. It is argued that MRs and GPGICs solve the scaling problem of intelligence from which the connectionism paradigm suffers.

Topological biclustering ARTMAP for identifying within bicluster relationships

Biclustering is a powerful tool for exploratory data analysis in domains such as social networking, data reduction, and differential gene expression studies. Topological learning identifies connected regions that are difficult to find using other traditional clustering methods and produces a graphical representation. Therefore, to improve the quality of biclustering and module extraction, this work combines the adaptive resonance theory (ART)-based methods of biclustering ARTMAP (BARTMAP) and topological ART (TopoART), to produce TopoBARTMAP. The latter inherits the ability to detect topological associations while performing data reduction. The capabilities of TopoBARTMAP were benchmarked using 35 real world cancer datasets and contrasted with other (bi)clustering methods, where it showed a statistically significant improvement over the other assessed methods on ordered and shuffled data experiments. In experiments with 12 synthetic datasets, the method was observed to perform better at identifying constant, scale, shift, and shift scale type biclusters. The produced graphical representation was refined to represent gene bicluster associations and was assessed on the NCBI GSE89116 dataset containing expression levels of 39,326 probes sampled over 38 observations.

Synapses, predictions, and prediction errors: A neocortical computational study of MDD using the temporal memory algorithm of HTM

Introduction

Synapses and spines play a significant role in major depressive disorder (MDD) pathophysiology, recently highlighted by the rapid antidepressant effect of ketamine and psilocybin. According to the Bayesian brain and interoception perspectives, MDD is formalized as being stuck in affective states constantly predicting negative energy balance. To understand how spines and synapses relate to the predictive function of the neocortex and thus to symptoms, we used the temporal memory (TM), an unsupervised machine-learning algorithm. TM models a single neocortical layer, learns in real-time, and extracts and predicts temporal sequences. TM exhibits neocortical biological features such as sparse firing and continuous online learning using local Hebbian-learning rules.

Methods

We trained a TM model on random sequences of upper-case alphabetical letters, representing sequences of affective states. To model depression, we progressively destroyed synapses in the TM model and examined how that affected the predictive capacity of the network. We found that the number of predictions decreased non-linearly.

Results

Destroying 50% of the synapses slightly reduced the number of predictions, followed by a marked drop with further destruction. However, reducing the synapses by 25% distinctly dropped the confidence in the predictions. Therefore, even though the network was making accurate predictions, the network was no longer confident about these predictions.

Discussion

These findings explain how interoceptive cortices could be stuck in limited affective states with high prediction error. Connecting ketamine and psilocybin's proposed mechanism of action to depression pathophysiology, the growth of new synapses would allow representing more futuristic predictions with higher confidence. To our knowledge, this is the first study to use the TM model to connect changes happening at synaptic levels to the Bayesian formulation of psychiatric symptomatology. Linking neurobiological abnormalities to symptoms will allow us to understand the mechanisms of treatments and possibly, develop new ones.

Temporal stability of human heading perception

Humans are capable of accurately judging their heading from optic flow during straight forward self-motion. Despite the global coherence in the optic flow field, however, visual clutter and other naturalistic conditions create constant flux on the eye. This presents a problem that must be overcome to accurately perceive heading from optic flow-the visual system must maintain sensitivity to optic flow variations that correspond with actual changes in self-motion and disregard those that do not. One solution could involve integrating optic flow over time to stabilize heading signals while suppressing transient fluctuations. Stability, however, may come at the cost of sluggishness. Here, we investigate the stability of human heading perception when subjects judge their heading after the simulated direction of self-motion changes. We found that the initial heading exerted an attractive influence on judgments of the final heading. Consistent with an evolving heading representation, bias toward the initial heading increased with the size of the heading change and as the viewing duration of the optic flow consistent with the final heading decreased. Introducing periods of sensory dropout (blackouts) later in the trial increased bias whereas an earlier one did not. Simulations of a neural model, the Competitive Dynamics Model, demonstrates that a mechanism that produces an evolving heading signal through recurrent competitive interactions largely captures the human data. Our findings characterize how the visual system balances stability in heading perception with sensitivity to change and support the hypothesis that heading perception evolves over time.

A review of recent advances and applications of machine learning in tribology

In tribology, a considerable number of computational and experimental approaches to understand the interfacial characteristics of material surfaces in motion and tribological behaviors of materials have been considered to date. Despite being useful in providing important insights on the tribological properties of a system, at different length scales, a vast amount of data generated from these state-of-the-art techniques remains underutilized due to lack of analysis methods or limitations of existing analysis techniques. In principle, this data can be used to address intractable tribological problems including structure-property relationships in tribological systems and efficient lubricant design in a cost and time effective manner with the aid of machine learning. Specifically, data-driven machine learning methods have shown potential in unraveling complicated processes through the development of structure-property/functionality relationships based on the collected data. For example, neural networks are incredibly effective in modeling non-linear correlations and identifying primary hidden patterns associated with these phenomena. Here we present several exemplary studies that have demonstrated the proficiency of machine learning in understanding these critical factors. A successful implementation of neural networks, supervised, and stochastic learning approaches in identifying structure-property relationships have shed light on how machine learning may be used in certain tribological applications. Moreover, ranging from the design of lubricants, composites, and experimental processes to studying fretting wear and frictional mechanism, machine learning has been embraced either independently or integrated with optimization algorithms by scientists to study tribology. Accordingly, this review aims at providing a perspective on the recent advances in the applications of machine learning in tribology. The review on referenced simulation approaches and subsequent applications of machine learning in experimental and computational tribology shall motivate researchers to introduce the revolutionary approach of machine learning in efficiently studying tribology.

The Grossberg Code: Universal Neural Network Signatures of Perceptual Experience

Two universal functional principles of Grossberg’s Adaptive Resonance Theory decipher the brain code of all biological learning and adaptive intelligence. Low-level representations of multisensory stimuli in their immediate environmental context are formed on the basis of bottom-up activation and under the control of top-down matching rules that integrate high-level, long-term traces of contextual configuration. These universal coding principles lead to the establishment of lasting brain signatures of perceptual experience in all living species, from aplysiae to primates. They are re-visited in this concept paper on the basis of examples drawn from the original code and from some of the most recent related empirical findings on contextual modulation in the brain, highlighting the potential of Grossberg’s pioneering insights and groundbreaking theoretical work for intelligent solutions in the domain of developmental and cognitive robotics.

AAPM task group report 273: Recommendations on best practices for AI and machine learning for computer-aided diagnosis in medical imaging

Rapid advances in artificial intelligence (AI) and machine learning, and specifically in deep learning (DL) techniques, have enabled broad application of these methods in health care. The promise of the DL approach has spurred further interest in computer-aided diagnosis (CAD) development and applications using both "traditional" machine learning methods and newer DL-based methods. We use the term CAD-AI to refer to this expanded clinical decision support environment that uses traditional and DL-based AI methods. Numerous studies have been published to date on the development of machine learning tools for computer-aided, or AI-assisted, clinical tasks. However, most of these machine learning models are not ready for clinical deployment. It is of paramount importance to ensure that a clinical decision support tool undergoes proper training and rigorous validation of its generalizability and robustness before adoption for patient care in the clinic. To address these important issues, the American Association of Physicists in Medicine (AAPM) Computer-Aided Image Analysis Subcommittee (CADSC) is charged, in part, to develop recommendations on practices and standards for the development and performance assessment of computer-aided decision support systems. The committee has previously published two opinion papers on the evaluation of CAD systems and issues associated with user training and quality assurance of these systems in the clinic. With machine learning techniques continuing to evolve and CAD applications expanding to new stages of the patient care process, the current task group report considers the broader issues common to the development of most, if not all, CAD-AI applications and their translation from the bench to the clinic. The goal is to bring attention to the proper training and validation of machine learning algorithms that may improve their generalizability and reliability and accelerate the adoption of CAD-AI systems for clinical decision support.

Neural Aspects of Prospective Control through Resonating Taus in an Interceptive Timing Task

High-density electroencephalography from visual and motor cortices in addition to kinematic hand and target movement recordings were used to investigate τ-coupling between brain activity patterns and physical movements in an interceptive timing task. Twelve adult participants were presented with a target car moving towards a destination at three constant accelerations, and an effector dot was available to intercept the car at the destination with a swift movement of the finger. A τ-coupling analysis was used to investigate involvement of perception and action variables at both the ecological scale of behavior and neural scale. By introducing the concept of resonance, the underlying dynamics of interceptive actions were investigated. A variety of one- and two-scale τ-coupling analyses showed significant differences in distinguishing between slow, medium, and fast target speed when car motion and finger movement, VEP and MRP brain activity, VEP and car motion, and MRP and finger movement were involved. These results suggested that the temporal structure present at the ecological scale is reflected at the neural scale. The results further showed a strong effect of target speed, indicating that τ-coupling constants k and k_res increased with higher speeds of the moving target. It was concluded that τ-coupling can be considered a valuable tool when combining different types of variables at both the ecological and neural levels of analysis.

A neural network account of memory replay and knowledge consolidation

Replay can consolidate memories through offline neural reactivation related to past experiences. Category knowledge is learned across multiple experiences, and its subsequent generalization is promoted by consolidation and replay during rest and sleep. However, aspects of replay are difficult to determine from neuroimaging studies. We provided insights into category knowledge replay by simulating these processes in a neural network which approximated the roles of the human ventral visual stream and hippocampus. Generative replay, akin to imagining new category instances, facilitated generalization to new experiences. Consolidation-related replay may therefore help to prepare us for the future as much as remember the past. Generative replay was more effective in later network layers functionally similar to the lateral occipital cortex than layers corresponding to early visual cortex, drawing a distinction between neural replay and its relevance to consolidation. Category replay was most beneficial for newly acquired knowledge, suggesting replay helps us adapt to changes in our environment. Finally, we present a novel mechanism for the observation that the brain selectively consolidates weaker information, namely a reinforcement learning process in which categories were replayed according to their contribution to network performance. This reinforces the idea of consolidation-related replay as an active rather than passive process.

Sleep-like unsupervised replay reduces catastrophic forgetting in artificial neural networks

Artificial neural networks are known to suffer from catastrophic forgetting: when learning multiple tasks sequentially, they perform well on the most recent task at the expense of previously learned tasks. In the brain, sleep is known to play an important role in incremental learning by replaying recent and old conflicting memory traces. Here we tested the hypothesis that implementing a sleep-like phase in artificial neural networks can protect old memories during new training and alleviate catastrophic forgetting. Sleep was implemented as off-line training with local unsupervised Hebbian plasticity rules and noisy input. In an incremental learning framework, sleep was able to recover old tasks that were otherwise forgotten. Previously learned memories were replayed spontaneously during sleep, forming unique representations for each class of inputs. Representational sparseness and neuronal activity corresponding to the old tasks increased while new task related activity decreased. The study suggests that spontaneous replay simulating sleep-like dynamics can alleviate catastrophic forgetting in artificial neural networks.

What if consciousness is not an emergent property of the brain? Observational and empirical challenges to materialistic models

The nature of consciousness is considered one of science's most perplexing and persistent mysteries. We all know the subjective experience of consciousness, but where does it arise? What is its purpose? What are its full capacities? The assumption within today's neuroscience is that all aspects of consciousness arise solely from interactions among neurons in the brain. However, the origin and mechanisms of qualia (i.e., subjective or phenomenological experience) are not understood. David Chalmers coined the term "the hard problem" to describe the difficulties in elucidating the origins of subjectivity from the point of view of reductive materialism. We propose that the hard problem arises because one or more assumptions within a materialistic worldview are either wrong or incomplete. If consciousness entails more than the activity of neurons, then we can contemplate new ways of thinking about the hard problem. This review examines phenomena that apparently contradict the notion that consciousness is exclusively dependent on brain activity, including phenomena where consciousness appears to extend beyond the physical brain and body in both space and time. The mechanisms underlying these "non-local" properties are vaguely suggestive of quantum entanglement in physics, but how such effects might manifest remains highly speculative. The existence of these non-local effects appears to support the proposal that post-materialistic models of consciousness may be required to break the conceptual impasse presented by the hard problem of consciousness.

Discriminating and Clustering Ordered Permutations Using Artificial Neural Networks: A Potential Application in ANN-Guided Genetic Algorithms

Traveling salesman, linear ordering, quadratic assignment, and flow shop scheduling are typical examples of permutation-based combinatorial optimization problems with real-life applications. These problems naturally represent solutions as an ordered permutation of objects. However, as the number of objects increases, finding optimal permutations is extremely difficult when using exact optimization methods. In those circumstances, approximate algorithms such as metaheuristics are a plausible way of finding acceptable solutions within a reasonable computational time. In this paper, we present a technique for clustering and discriminating ordered permutations with potential applications in developing neural network-guided metaheuristics to solve this class of problems. In this endeavor, we developed two different techniques to convert ordered permutations to binary-vectors and considered Adaptive Resonate Theory (ART) neural networks for clustering the resulting binary vectors. The proposed binary conversion techniques and two neural networks (ART-1 and Improved ART-1) are examined under various performance indicators. Numerical examples show that one of the binary conversion methods provides better results than the other, and Improved ART-1 is superior to ART-1. Additionally, we apply the proposed clustering and discriminating technique to develop a neural-network-guided Genetic Algorithm (GA) to solve a flow-shop scheduling problem. The investigation shows that the neural network-guided GA outperforms pure GA.

A Quantum Predictive Brain: Complementarity Between Top-Down Predictions and Bottom-Up Evidence

Predictive brain theory challenges the general assumption of a brain extracting knowledge from sensations and considers the brain as an organ of inference, actively constructing explanations about reality beyond its sensory evidence. Predictive brain has been formalized through Bayesian updating, where top-down predictions are compared with bottom-up evidence. In this article, we propose a different approach to predictive brain based on quantum probability-we call it Quantum Predictive Brain (QPB). QPB is consistent with the Bayesian framework, but considers it as a special case. The tenet of QPB is that top-down predictions and bottom-up evidence are complementary, as they cannot be co-jointly determined to pursue a univocal model of brain functioning. QPB can account for several high-order cognitive phenomena (which are problematic in current predictive brain theories) and offers new insights into the mechanisms of neural reuse.

BepFAMN: A Method for Linear B-Cell Epitope Predictions Based on Fuzzy-ARTMAP Artificial Neural Network

The public health system is extremely dependent on the use of vaccines to immunize the population from a series of infectious and dangerous diseases, preventing the system from collapsing and millions of people dying every year. However, to develop these vaccines and effectively monitor these diseases, it is necessary to use accurate diagnostic methods capable of identifying highly immunogenic regions within a given pathogenic protein. Existing experimental methods are expensive, time-consuming, and require arduous laboratory work, as they require the screening of a large number of potential candidate epitopes, making the methods extremely laborious, especially for application to larger microorganisms. In the last decades, researchers have developed in silico prediction methods, based on machine learning, to identify these markers, to drastically reduce the list of potential candidate epitopes for experimental tests, and, consequently, to reduce the laborious task associated with their mapping. Despite these efforts, the tools and methods still have low accuracy, slow diagnosis, and offline training. Thus, we develop a method to predict B-cell linear epitopes which are based on a Fuzzy-ARTMAP neural network architecture, called BepFAMN (B Epitope Prediction Fuzzy ARTMAP Artificial Neural Network). This was trained using a linear averaging scheme on 15 properties that include an amino acid ratio scale and a set of 14 physicochemical scales. The database used was obtained from the IEDB website, from which the amino acid sequences with the annotations of their positive and negative epitopes were taken. To train and validate the knowledge models, five-fold cross-validation and competition techniques were used. The BepiPred-2.0 database, an independent database, was used for the tests. In our experiment, the validation dataset reached sensitivity = 91.50%, specificity = 91.49%, accuracy = 91.49%, MCC = 0.83, and an area under the curve (AUC) ROC of approximately 0.9289. The result in the testing dataset achieves a significant improvement, with sensitivity = 81.87%, specificity = 74.75%, accuracy = 78.27%, MCC = 0.56, and AOC = 0.7831. These achieved values demonstrate that BepFAMN outperforms all other linear B-cell epitope prediction tools currently used. In addition, the architecture provides mechanisms for online training, which allow the user to find a new B-cell linear epitope, and to improve the model without need to re-train itself with the whole dataset. This fact contributes to a considerable reduction in the number of potential linear epitopes to be experimentally validated, reducing laboratory time and accelerating the development of diagnostic tests, vaccines, and immunotherapeutic approaches.

Design Principles for Neurorobotics

In their book "How the Body Shapes the Way We Think: A New View of Intelligence," Pfeifer and Bongard put forth an embodied approach to cognition. Because of this position, many of their robot examples demonstrated "intelligent" behavior despite limited neural processing. It is our belief that neurorobots should attempt to follow many of these principles. In this article, we discuss a number of principles to consider when designing neurorobots and experiments using robots to test brain theories. These principles are strongly inspired by Pfeifer and Bongard, but build on their design principles by grounding them in neuroscience and by adding principles based on neuroscience research. Our design principles fall into three categories. First, organisms must react quickly and appropriately to events. Second, organisms must have the ability to learn and remember over their lifetimes. Third, organisms must weigh options that are crucial for survival. We believe that by following these design principles a robot's behavior will be more naturalistic and more successful.

Resonance as a Design Strategy for AI and Social Robots

Resonance, a powerful and pervasive phenomenon, appears to play a major role in human interactions. This article investigates the relationship between the physical mechanism of resonance and the human experience of resonance, and considers possibilities for enhancing the experience of resonance within human-robot interactions. We first introduce resonance as a widespread cultural and scientific metaphor. Then, we review the nature of "sympathetic resonance" as a physical mechanism. Following this introduction, the remainder of the article is organized in two parts. In part one, we review the role of resonance (including synchronization and rhythmic entrainment) in human cognition and social interactions. Then, in part two, we review resonance-related phenomena in robotics and artificial intelligence (AI). These two reviews serve as ground for the introduction of a design strategy and combinatorial design space for shaping resonant interactions with robots and AI. We conclude by posing hypotheses and research questions for future empirical studies and discuss a range of ethical and aesthetic issues associated with resonance in human-robot interactions.

Metacognition as a Consequence of Competing Evolutionary Time Scales

Evolution is full of coevolving systems characterized by complex spatio-temporal interactions that lead to intertwined processes of adaptation. Yet, how adaptation across multiple levels of temporal scales and biological complexity is achieved remains unclear. Here, we formalize how evolutionary multi-scale processing underlying adaptation constitutes a form of metacognition flowing from definitions of metaprocessing in machine learning. We show (1) how the evolution of metacognitive systems can be expected when fitness landscapes vary on multiple time scales, and (2) how multiple time scales emerge during coevolutionary processes of sufficiently complex interactions. After defining a metaprocessor as a regulator with local memory, we prove that metacognition is more energetically efficient than purely object-level cognition when selection operates at multiple timescales in evolution. Furthermore, we show that existing modeling approaches to coadaptation and coevolution-here active inference networks, predator-prey interactions, coupled genetic algorithms, and generative adversarial networks-lead to multiple emergent timescales underlying forms of metacognition. Lastly, we show how coarse-grained structures emerge naturally in any resource-limited system, providing sufficient evidence for metacognitive systems to be a prevalent and vital component of (co-)evolution. Therefore, multi-scale processing is a necessary requirement for many evolutionary scenarios, leading to de facto metacognitive evolutionary outcomes.

Toward Understanding the Brain Dynamics of Music: Learning and Conscious Performance of Lyrics and Melodies With Variable Rhythms and Beats

A neural network architecture models how humans learn and consciously perform musical lyrics and melodies with variable rhythms and beats, using brain design principles and mechanisms that evolved earlier than human musical capabilities, and that have explained and predicted many kinds of psychological and neurobiological data. One principle is called factorization of order and rhythm: Working memories store sequential information in a rate-invariant and speaker-invariant way to avoid using excessive memory and to support learning of language, spatial, and motor skills. Stored invariant representations can be flexibly performed in a rate-dependent and speaker-dependent way under volitional control. A canonical working memory design stores linguistic, spatial, motoric, and musical sequences, including sequences with repeated words in lyrics, or repeated pitches in songs. Stored sequences of individual word chunks and pitch chunks are categorized through learning into lyrics chunks and pitches chunks. Pitches chunks respond selectively to stored sequences of individual pitch chunks that categorize harmonics of each pitch, thereby supporting tonal music. Bottom-up and top-down learning between working memory and chunking networks dynamically stabilizes the memory of learned music. Songs are learned by associatively linking sequences of lyrics and pitches chunks. Performance begins when list chunks read word chunk and pitch chunk sequences into working memory. Learning and performance of regular rhythms exploits cortical modulation of beats that are generated in the basal ganglia. Arbitrary performance rhythms are learned by adaptive timing circuits in the cerebellum interacting with prefrontal cortex and basal ganglia. The same network design that controls walking, running, and finger tapping also generates beats and the urge to move with a beat.

Exploring Orthographic Representation in Chinese Handwriting: A Mega-Study Based on a Pedagogical Corpus of CFL Learners

Writing and reading are closely related and are thus likely to have a common orthographic representation. A fundamental question in the literature on the production of written Chinese characters concerns the structure of orthographic representations. We report on a Chinese character handwriting pedagogical corpus involving a class of 22 persons, 232 composite character types, 1,913 tokens, and 13,057 stroke records, together with the inter-stroke interval (ISI), which reflects the parallel processing of multilevel orthographic representation during the writing execution, and 50 orthographic variables from the whole character, logographeme, and stroke. The results of regression analyses show that orthographic representation has a hierarchy and that different representational levels are active simultaneously. In the multilevel structure of orthographic representation, the representation of the logographeme is absolutely dominant. Writing and reading have both commonalities and individual differences in their orthographic representations. The online processing of the logographeme unit probably occurs at the ISI before the initial stroke of the current logographeme, which may also cascade to the first subsequent logographeme. In addition, we propose a new effective character structure unit for describing orthographic complexity.

SuPART: supervised projective adapted resonance theory for automatic quality assurance approval of radiotherapy treatment plans

Radiotherapy is a common treatment modality for the treatment of cancer, where treatments must be carefully designed to deliver appropriate dose to targets while avoiding healthy organs. The comprehensive multi-disciplinary quality assurance (QA) process in radiotherapy is designed to ensure safe and effective treatment plans are delivered to patients. However, the plan QA process is expensive, often time-intensive, and requires review of large quantities of complex data, potentially leading to human error in QA assessment. We therefore develop an automated machine learning algorithm to identify ‘acceptable’ plans (plans that are similar to historically approved plans) and ‘unacceptable’ plans (plans that are dissimilar to historically approved plans). This algorithm is a supervised extension of projective adaptive resonance theory, called SuPART, that learns a set of distinctive features, and considers deviations from them indications of unacceptable plans. We test SuPART on breast and prostate radiotherapy datasets from our institution, and find that SuPART outperforms common classification algorithms in several measures of accuracy. When no falsely approved plans are allowed, SuPART can correctly auto-approve 34% of the acceptable breast and 32% of the acceptable prostate plans, and can also correctly reject 53% of the unacceptable breast and 56% of the unacceptable prostate plans. Thus, usage of SuPART to aid in QA could potentially yield significant time savings.

LAG-1: A dynamic, integrative model of learning, attention, and gaze

It is clear that learning and attention interact, but it is an ongoing challenge to integrate their psychological and neurophysiological descriptions. Here we introduce LAG-1, a dynamic neural field model of learning, attention and gaze, that we fit to human learning and eye-movement data from two category learning experiments. LAG-1 comprises three control systems: one for visuospatial attention, one for saccadic timing and control, and one for category learning. The model is able to extract a kind of information gain from pairwise differences in simple associations between visual features and categories. Providing this gain as a reentrant signal with bottom-up visual information, and in top-down spatial priority, appropriately influences the initiation of saccades. LAG-1 provides a moment-by-moment simulation of the interactions of learning and gaze, and thus simultaneously produces phenomena on many timescales, from the duration of saccades and gaze fixations, to the response times for trials, to the slow optimization of attention toward task relevant information across a whole experiment. With only three free parameters (learning rate, trial impatience, and fixation impatience) LAG-1 produces qualitatively correct fits for learning, behavioural timing and eye movement measures, and also for previously unmodelled empirical phenomena (e.g., fixation orders showing stimulus-specific attention, and decreasing fixation counts during feedback). Because LAG-1 is built to capture attention and gaze generally, we demonstrate how it can be applied to other phenomena of visual cognition such as the free viewing of visual stimuli, visual search, and covert attention.

Deficient Recurrent Cortical Processing in Congenital Deafness

The influence of sensory experience on cortical feedforward and feedback interactions has rarely been studied in the auditory cortex. Previous work has documented a dystrophic effect of deafness in deep cortical layers, and a reduction of interareal couplings between primary and secondary auditory areas in congenital deafness which was particularly pronounced in the top-down direction (from the secondary to the primary area). In the present study, we directly quantified the functional interaction between superficial (supragranular, I to III) and deep (infragranular, V and VI) layers of feline’s primary auditory cortex A1, and also between superficial/deep layers of A1 and a secondary auditory cortex, namely the posterior auditory field (PAF). We compared adult hearing cats under acoustic stimulation and cochlear implant (CI) stimulation to adult congenitally deaf cats (CDC) under CI stimulation. Neuronal activity was recorded from auditory fields A1 and PAF simultaneously with two NeuroNexus electrode arrays. We quantified the spike field coherence (i.e., the statistical dependence of spike trains at one electrode with local field potentials on another electrode) using pairwise phase consistency (PPC). Both the magnitude as well as the preferred phase of synchronization was analyzed. The magnitude of PPC was significantly smaller in CDCs than in controls. Furthermore, controls showed no significant difference between the preferred phase of synchronization between supragranular and infragranular layers, both in acoustic and electric stimulation. In CDCs, however, there was a large difference in the preferred phase between supragranular and infragranular layers. These results demonstrate a loss of synchrony and for the first time directly document a functional decoupling of the interaction between supragranular and infragranular layers of the primary auditory cortex in congenital deafness. Since these are key for the influence of top-down to bottom-up computations, the results suggest a loss of recurrent cortical processing in congenital deafness and explain the outcomes of previous studies by deficits in intracolumnar microcircuitry.

Study of Transmission Line Boundary Protection Using a Multilayer Perceptron Neural Network with Back Propagation and Wavelet Transform

Protection schemes are usually implemented in the planning of transmission line operations. These schemes are expected to protect not only the network of transmission lines but also the entire power systems network during fault conditions. However, it is often a challenge for these schemes to differentiate accurately between various fault locations. This study analyses the deficiencies identified in existing protection schemes and investigates a different method that proposes to overcome these shortcomings. The proposed scheme operates by performing a wavelet transform on the fault-generated signal, which reduces the signal into frequency components. These components are then used as the input data for a multilayer perceptron neural network with backpropagation that can classify between different fault locations in the system. The study uses the transient signal generated during fault conditions to identify faults. The scientific research paradigm was adopted for the study. It also adopted the deduction research approach as it requires data collection via simulation using the Simscape electrical sub-program of Simulink within Matrix laboratory (MATLAB). The outcome of the study shows that the simulation correctly classifies 70.59% of the faults when tested. This implies that the majority of the faults can be detected and accurately isolated using boundary protection of transmission lines with the help of wavelet transforms and a neural network. The outcome also shows that more accurate fault identification and classification are achievable by using neural network than by the conventional system currently in use.

Cognitive Based Authentication Protocol for Distributed Data and Web Technologies

The objective of the verification process, besides guaranteeing security, is also to be effective and robust. This means that the login should take as little time as possible, and each time allow for a successful authentication of the authorised account. In recent years, however, online users have been experiencing more and more issues with recalling their own passwords on the spot. According to research done in 2017 by LastPass on its employees, the number of personal accounts assigned to one business user currently exceeds 191 profiles and keeps growing. Remembering these many passwords, especially to applications which are not used every week, seems to be impossible without storing them either on paper, in a password manager, or saved in a file somewhere on a PC. In this article a new verification model using a Google Street View image as well as the user’s personal experience and knowledge will be presented. The purpose of this scheme is to assure secure verification by creating longer passwords as well as delivering a ‘password reminder’ already embedded into the login scheme.

The fronto-parietal network is not a flexible hub during naturalistic cognition

The fronto‐parietal network (FPN) is crucial for cognitively demanding tasks as it selectively represents task‐relevant information and controls other brain regions. To implement these functions, it has been argued that it is a flexible hub that reconfigures its functional connectivity with other networks. This was supported by a study in which a set of demanding tasks were presented, that varied in their sensory features, comparison rules, and response mappings, and the FPN showed greater reconfiguration of functional connectivity between tasks than any other network. However, this task set was designed to engage the FPN, and therefore it remains an open question whether the FPN is in a flexible hub in general or only for such task sets. Using two freely available datasets (Experiment 1, N = 15, Experiment 2, N = 644), we examined dynamic functional connectivity during naturalistic cognition, while participants watched a movie. Many differences in the flexibility were found across networks but the FPN was not the most flexible hub in the brain, during either movie for any of two measures, using a regression model or a correlation model and across five timescales. We, therefore, conclude that the FPN does not have the trait of being a flexible hub, although it may adopt this state for particular task sets.