The notion of intelligence is a complex theoretical construct. Artificial General Intelligence (AGI) is confronted with different aspects of intelligence. What kind of intelligence does AI represent? Artificial neural networks may internalize much more information than humans. In what way do functional processes in the brain resemble those in AI algorithms? Internal models of the world are already emerging in large AI systems. Progress in hardware and software inevitably leads to the development of autonomous AI systems capable of setting their own goals, self-reflection, self-improvement, and even survival instinct. They will deserve the status of Digital Intelligent Beings (DIBs). How will it influence our society?1

One important application of machine learning is the analysis of electroencephalographic recordings (EEG). Such oscillatory signals are noisy, non-stationary, full of artifacts, contain transients, and chaotic transitions between meta-stable states. EEG data rarely consists of recordings obtained for more than 100 patients with some brain disorder. This makes objective, reliable clinical diagnosis of many mental disorders, such as schizophrenia, very difficult. The standard approach splits the frequency of EEG oscillations focusing on five classical bands: delta, theta, alpha, beta, and gamma. In this paper, we investigate three main questions: (i) Are there certain frequencies that will allow for a reliable diagnosis of schizophrenia, and (ii) if they are, is there a connection to what is known about their neural basis? (iii) How long segments of EEG recordings are sufficient for reliable classification? We filter EEG signals, obtaining a set of 64 very narrow 1Hz frequency bands. A genetic algorithm, with a simple k-NN classifier, finds an optimal combination of these bands. The method is sufficiently simple to be used in clinical settings. A public schizophrenia EEG data set, containing 60 seconds of EEG recordings in the resting state, with only 16 electrodes, for 45 adolescent patients and 39 healthy controls, was used for testing. Signal duration of about 30 seconds enables to reach an accuracy of over 96% in 5-fold cross-validation tests. We compare our results to much more sophisticated state-of-the- art methods and discuss insights gained by our analysis into the brain basis of schizophrenia.

Identifying mental disorders through EEG-based biomarkers holds significant clinical value. However, the small number of patients and the high variability of individual data pose challenges in their development. Existing solutions are either not sufficiently accurate or overly complicated for practical clinical use. Topological Data Analysis (TDA) offers a potential method for deriving robust features from intricate time series. This study explores various EEG signal embeddings, generating TDA features for the classification of cases of schizophrenia in adolescents. Just two features were sufficient to achieve an accuracy of 80% in a 5-fold cross-validation, offering a straightforward and insightful model for diagnosing schizophrenia.

Spectral fingerprints (SFs) are unique power spectra signatures of human brain regions of interest (ROIs, Keitel & Gross, 2016). SFs allow for accurate ROI identification and can serve as biomarkers of differences exhibited by non-neurotypical groups. At present, there are no open-source, versatile tools to calculate spectral fingerprints. We have filled this gap by creating a modular, highly-configurable MATLAB Toolbox for Frequency-based Fingerprinting (ToFFi). It can transform magnetoencephalographic and electroencephalographic signals into unique spectral representations using ROIs provided by anatomical (AAL, Desikan-Killiany), functional (Schaefer), or other custom volumetric brain parcellations. Toolbox design supports reproducibility and parallel computations.

A brain–computer interface (BCI) establishes a direct communication pathway between the human brain and a computer. It has been widely used in medical diagnosis, rehabilitation, education, entertainment, and so on. Most research so far focuses on making BCIs more accurate and reliable, but much less attention has been paid to their privacy. Developing a commercial BCI system usually requires close collaborations among multiple organizations, e.g., hospitals, universities, and/or companies. Input data in BCIs, e.g., electroencephalogram (EEG), contain rich privacy information, and the developed machine learning model is usually proprietary. Data and model transmission among different parties may incur significant privacy threats, and hence, privacy protection in BCIs must be considered. Unfortunately, there does not exist any contemporary and comprehensive review on privacy-preserving BCIs. This article fills this gap, by describing potential privacy threats and protection strategies in BCIs. It also points out several challenges and future research directions in developing privacy-preserving BCIs.

Companies and even institutions have recently tried to form a reasonable opinion about their products.One way companies can achieve this is to prepare and distribute advertisements.Some advertising media, such as TV commercials, are costly and tiresome, so understanding and modeling some aspects of disseminating product information might be helpful.The article presents a multi-agent model of spreading trust in the product among agents based on fuzzy aggregation norms.When people with a similar opinion about a product discuss it with others or listen to a commercial, their trust increases, so the optimistic fuzzy aggregation norm is applied.When people with different opinions meet, their faith decreases, so the pessimistic norm is applied.In addition, the paper presents a theoretical example of this model application, namely, showing the results of a multi-agent model of spreading confidence in the product and presenting the results of the NetLogo simulation.

Classification methods with linear computational complexity O(nd) in the number of samples n and their dimensionality d often give results that are better or at least statistically not significantly worse that slower algorithms. This is demonstrated here for many benchmark datasets downloaded from the UCI Machine Learning Repository. Results provided in this paper should be used as a reference for estimating usefulness of new learning algorithms: higher complexity methods should provide significantly better results to justify their use.

Is science of human experience, aimed at explaining phenomenology of mental events accessible through introspection, possible? What do we really known about ourselves and how do we know it? Psychology and neural sciences have turned away from such questions and experimentally oriented philosophers discovered formidable obstacles in attempts to answer even simple questions about the nature of conscious experience. To talk in a meaningful way about subjective mental processes a new level of description is needed, resulting from neurodynamics but connected to inner experience. Visualization of neurodynamics may lead to geometrical, continuous models of mental events. It should allow to view brains and artificial cognitive systems form mental perspective.

Support Vector Machines (SVMs) with various kernels have played dominant role in machine learning for many years, finding numerous applications. Although they have many attractive features interpretation of their solutions is quite difficult, the use of a single kernel type may not be appropriate in all areas of the input space, convergence problems for some kernels are not uncommon, the standard quadratic programming solution has O(m3) time and O(m2) space complexity for m training patterns. Kernel methods work because they implicitly provide new, useful features. Such features, derived from various kernels and other vector transformations, may be used directly in any machine learning algorithm, facilitating multiresolution, heterogeneous models of data. Therefore Support Feature Machines (SFM) based on linear models in the extended feature spaces, enabling control over selection of support features, give at least as good results as any kernel-based SVMs, removing all problems related to interpretation, scaling and convergence. This is demonstrated for a number of benchmark datasets analyzed with linear discrimination, SVM, decision trees and nearest neighbor methods.

This book identifies the short-term and long-term challenges facing the field of computational intelligence (CI). After an introductory chapter, the next six chapters describe challenges resulting from attempts to model cognitive and neurocognitive processes. The next three chapters concern foundations of CI, while the last six chapters describe some theoretical problem and special subareas of CI. The book provides new ideas, helpful discussions of new directions of research, and non-trivial challenges for CI. It should be interesting for both established scientists and young researchers.

Acquisition and representation of semantic concepts is a necessary requirement for the understanding of natural languages by cognitive systems. Word games provide an interesting opportunity for semantic knowledge acquisition that may be used to construct semantic memory. A task-dependent architecture of the knowledge base inspired by psycholinguistic theories of human cognition process is introduced. The core of the system is an algorithm for semantic search using a simplified vector representation of concepts. Based on this algorithm a 20 questions game has been implemented. This implementation provides an example of an application of the semantic memory, but also allows for testing the linguistic competence of the system. A web portal with Haptek-based talking head interface facilitates acquisition of a new knowledge while playing the game and engaging in dialogs with users.

In the past few years, various attempts to define the problem of consciousness from engineering or computational perspective have emerged. This book explores the topic of what will make a robot conscious. It outlines an engineering approach that should lead toward cognitive and conscious machines. Some of the topics covered include: a vector representation of information as a substitute for nonnumerical information processing by brains; neuron models and associate neural networks; circuit assemblies; machine perception; motor control; machine cognition; machine emotions; natural language understanding; cognitive architecture; and machine consciousness. The book should be interesting to people working in computational intelligence and cognitive robots. Researchers in computational neuroscience, who usually focus on details, may also benefit from the integrative approach presented in this book. While the book introduces a number of interesting points of view, no breakthrough technologies are proposed and no software or hardware demonstration is provided.

A significant part of medical data remains stored as unstructured texts. Semantic search requires introduction of markup tags. Medical concepts discovered in hospital discharge summaries are used to create several feature spaces. Experts use their background knowledge to categorize new documents, and knowing category of the document disambiguate words and acronyms. A model of document similarity to reference sources that captures some intuitions of an expert is introduced. Parameters of the model are evaluated using linear programming techniques. This approach is applied to categorization of the medical discharge summaries providing simpler and more accurate model than alternative text categorization approaches

Putative brain processes responsible for understanding language are based on spreading activation in semantic networks, providing enhanced representations that involve concepts not found directly in the text. Approximation of this process is of great practical and theoretical interest. Vector model should reflect activations of various concepts in the brain spreading through associative network. Medical ontologies are used to select concepts of specific semantic type and add to each of them related concepts, providing expanded vector representations. The process is constrained by selection of useful extensions for the classification task. Short hospital discharge summaries are used to illustrate how this process works on a real, very noisy data. Results show significantly improved clustering and classification accuracy. A practical approach to mapping of associative networks of the brain involved in processing of specific concepts is presented.

Many language-oriented problems cannot be solved without semantic memory containing descriptions of concepts at different level of details. Automatic creation of semantic memories is a great challenge even for the simplest knowledge representation methods based on relations between concepts and keywords. Semantic memory based on such simple knowledge representation facilitates implementation of quite interesting linguistic competences that have not yet been demonstrated by more sophisticated rule or frame-based knowledge bases, for example CYC. These linguistic abilities include word games, such as the twenty questions game, that may be implemented using semantic memory built on relational model for knowledge representation. Creation of large-scale knowledge base for semantic memory involves mining structured information sources (ontologies, dictionaries, encyclopedic entries) and free texts (textbooks and Internet sources). Quality of this knowledge may be improved using collaborative projects in which systems that already possess some linguistic competence actively interact with human users, mining their knowledge. In this article three dialog scenarios for mining human knowledge are introduced, and the data acquired into semantic memory structures through such interaction is described.

Autonomous exploration by a team of robots has many important applications in rescue operations, clearing of mine fields and other military applications, and even space exploration. With limited range of sensors robots have to divide exploration tasks among themselves working under multiple constraints. An optimal covering of two-dimensional area by robot trajectories requires formation of space-filling Peano curves. This may be achieved using self-organizing feature map (SOFM) algorithm. There are two steps involved in the proposed approach: first optimal trajectories are defined generating Peano curves for space of arbitrary shape using the SOFM algorithm, and second, robots are deployed for exploration based on selection of start/end nodes and radius of robot sensors. The same approach may be used to direct people or teams exploring some area in rescue operations. Tests simulations show that this approach achieves better coverage and faster exploration than competing algorithms

Creative thinking is one of the hallmarks of human-level competence. Although it is still a poorly understood subject speculative ideas about brain processes involved in creative thinking may be implemented in computational models. A review of different approaches to creativity, insight and intuition is presented. Two factors are essential for creativity: imagination and selection or filtering. Imagination should be constrained by experience, while filtering in the case of creative use of words may be based on semantic and phonological associations. Analysis of brain processes involved in invention of new words leads to practical algorithms that create many interesting and novel names associated with a set of keywords.

Approximately 57 different types of clinical annotations construct a patient’s medical record. These annotations include radiology reports, discharge summaries, and surgical and nursing notes. Hospitals typically produce millions of text-based medical records over the course of a year. These records are essential for the delivery of care, but many are underutilized or not utilized at all for clinical research. The textual data found in these annotations is a rich source of insights into aspects of clinical care and the clinical delivery system. Recent regulatory actions, however, require that, in many cases, data not obtained through informed consent or data not related to the delivery of care must be made anonymous (as referred to by regulators as harmless), before they can be used. This article describes a practical approach with which Cincinnati Children’s Hospital Medical Center (CCHMC), a large pediatric academic medical center with more than 761,000 annual patient encounters, developed open source software for making pediatric clinical text harmless without losing its rich meaning. Development of the software dealt with many of the issues that often arise in natural language processing, such as data collection, disambiguation, and data scrubbing.

Avatars that show intelligent behavior should have an access to general knowledge about the world, knowledge that humans store in their semantic memories. The simplest knowledge representation for semantic memory is based on the concept description vectors (CDVs) that store, for each concept, an information whether a given property can be applied to this concept or not. Unfortunately large-scale semantic memories are not available. Experiments with automatic creation of concept description vectors from various sources, including ontologies, dictionaries, encyclopedias and unstructured text sources are described. Haptek-based talking head that has an access to this memory has been created as an example of a humanized interface (HIT) that can interact with Web pages and exchange information in a natural way. A few examples of applications of an avatar with semantic memory are given, including the twenty questions game and automatic creation of word puzzles

Probability that a crisp logical rule applied to imprecise input data is true may be computed using fuzzy membership function (MF). All reasonable assumptions about input uncertainty distributions lead to MFs of sigmoidal shape. Convolution of several inputs with uniform uncertainty leads to bell-shaped Gaussian-like uncertainty functions. Relations between input uncertainties and fuzzy rules are systematically explored and several new types of MFs discovered. Multilayered perceptron (MLP) networks are shown to be a particular implementation of hierarchical sets of fuzzy threshold logic rules based on sigmoidal MFs. They are equivalent to crisp logical networks applied to input data with uncertainty. Leaving fuzziness on the input side makes the networks or the rule systems easier to understand. Practical applications of these ideas are presented for analysis of questionnaire data and gene expression data.

A comparison between five feature ranking methods based on entropy is presented on artificial and real datasets. Feature ranking method using /spl chi//sup 2/ statistics gives results that are very similar to the entropy-based methods. The quality of feature rankings obtained by these methods is evaluated using the decision tree and the nearest neighbor classifier with growing number of most important features. Significant differences are found in some cases, but there is no single best index that works best for all data and all classifiers. Therefore to be sure that a subset of features giving highest accuracy has been selected requires the use of many different indices.

In many applications, black-box prediction is not satisfactory, and understanding the data is of critical importance. Typically, approaches useful for understanding of data involve logical rules, evaluate similarity to prototypes, or are based on visualization or graphical methods. This paper is focused on the extraction and use of logical rules for data understanding. All aspects of rule generation, optimization, and application are described, including the problem of finding good symbolic descriptors for continuous data, tradeoffs between accuracy and simplicity at the rule-extraction stage, and tradeoffs between rejection and error level at the rule optimization stage. Stability of rule-based description, calculation of probabilities from rules, and other related issues are also discussed. Major approaches to extraction of logical rules based on neural networks, decision trees, machine learning, and statistical methods are introduced. Optimization and application issues for sets of logical rules are described. Applications of such methods to benchmark and real-life problems are reported and illustrated with simple logical rules for many datasets. Challenges and new directions for research are outlined.

Neural networks are commonly regarded as black boxes performing incomprehensible functions. For classification problems networks provide maps from high dimensional feature space to K-dimensional image space. Images of training vector are projected on polygon vertices, providing visualization of network function. Such visualization may show the dynamics of learning, allow for comparison of different networks, display training vectors around which potential problems may arise, show differences due to regularization and optimization procedures, investigate stability of network classification under perturbation of original vectors, and place new data sample in relation to training data, allowing for estimation of confidence in classification of a given sample. An illustrative example for the three-class Wine data and five-class Satimage data is described. The visualization method proposed here is applicable to any black box system that provides continuous outputs.

A new methodology of extraction, optimization, and application of sets of logical rules is described. Neural networks are used for initial rule extraction, local or global minimization procedures for optimization, and Gaussian uncertainties of measurements are assumed during application of logical rules. Algorithms for extraction of logical rules from data with real-valued features require determination of linguistic variables or membership functions. Contest-dependent membership functions for crisp and fuzzy linguistic variables are introduced and methods of their determination described. Several neural and machine learning methods of logical rule extraction generating initial rules are described, based on constrained multilayer perceptron, networks with localized transfer functions or on separability criteria for determination of linguistic variables. A tradeoff between accuracy/simplicity is explored at the rule extraction stage and between rejection/error level at the optimization stage. Gaussian uncertainties of measurements are assumed during application of crisp logical rules, leading to "soft trapezoidal" membership functions and allowing to optimize the linguistic variables using gradient procedures. Numerous applications of this methodology to benchmark and real-life problems are reported and very simple crisp logical rules for many datasets provided.

The choice of transfer functions may strongly influence complexity and performance of neural networks used in classification and approximation tasks. A taxonomy of activation and output functions is proposed, allowing to generate many transfer functions. Several less-known types of transfer functions and new combinations of activation/output functions are described. Functions parameterize to change from localized to delocalized type, functions with activation based on nonEuclidean distance measures, bicentral functions formed from pairs of sigmoids are discussed.

Multilayer perceptrons (MLPs) use scalar products to compute weighted activation of neurons providing decision borders using combinations of soft hyperplanes. The weighted fan-in activation function corresponds to Euclidean distance functions used to compute similarities between input and weight vector. Replacing the fan-in activation function by non-Euclidean distance function offers a natural generalization of the standard MLP model, providing more flexible decision borders. An alternative way leading to similar results is based on renormalization of the input vectors using non-Euclidean norms in extended feature spaces. Both approaches influence the shapes of decision borders dramatically, allowing to reduce the complexity of MLP networks.

Neural networks are used for initial determination of linguistic variables and extraction of logical rules from data. Hierarchical sets of rules for different accuracy/rejection trade-offs are obtained. Sets of logical rules are optimized by maximization of their predictive power. To avoid global minimization methods for crisp logical rules Gaussian uncertainties of inputs are assumed. Analytical formulas reproducing Monte Carlo results for such inputs are derived, leading to a "soft trapezoidal" membership functions instead of rectangular functions used for crisp logical rules. Such approach increases accuracy, gives probabilities of classification instead of the yes/no answers, and allows one to optimize sets of rules using gradient procedures. A few illustrative applications to benchmark and real life problems show the effectiveness of this approach.

The class of similarity based methods (SBM) covers most neural models and many other classifiers. Performance of such methods is significantly improved if irrelevant features are removed and feature weights introduced, scaling their influence on calculation of similarity. Several methods for feature selection and weighting are described. As an alternative to the global minimization procedures computationally efficient best-first search methods are advocated. Although these methods can be used with any SBM classifier they have been tested using the k-NN method since it is relatively fast and for some databases gives excellent results. A few illustrative examples show significant improvements due to the feature weighting and selection.