My Literature Resources
Content Overview
Cognitive Architecture
Introduction
Cognitive architecture constitutes the generic blueprint for general intelligent agents, which provides an infrastructure for developing computational models that exhibit human-level intelligence and cover a wide range of domains. Cognitive architectures broadly fall into three types: symbolic, connectionist and hybrid. Symbolic architectures typically operate on a set of expert rules, the design philosophy of which is based on the analogy that mind is like a computer program. Connectionist architectures, on the other hand, specify no such rules a priori and relies instead on the emergent properties of many processing units. Most of these architectures employ artificial neural networks in their implementations. Given the respective strengths of the symbolic and connectionist approaches, a promising venue would therefore be to integrate symbolic and connectionist paradigms in order to build a complete framework for cognition. The need for such hybrid architectures have been steadily growing, with the aim to gain the merits of both symbolic and connectionist approaches as well as their synergy.
Scientists
Cognitive Neuroscience
Introduction
Neuroscience is a field of study that deals with the structure, function, genetics, biochemistry, physiology, pharmacology, and pathology of the nervous system. One primary focus is the study of the human brain at various levels of abstraction, ranging from the molecular level of individual neurons, the level of relatively small assembly of neurons like cortical columns, up to large systems including the cerebral cortex or cerebellum, and finally the nervous system as a whole. At this highest level, neuroscientific approaches combine with cognitive science to create cognitive neuroscience, a discipline first populated by cognitive psychologists and currently becoming a specialty of its own. Cognitive neuroscience provides a bottom-up approach to understanding (human) cognition, and is complementary to the top-down approach of psychology.
Scientists
Neural Network
Introduction
Artificial neural network (ANN) or simply neural network, is an interconnected group of artificial neurons that capture the parallel, distributed notion of the human cognition. ANN is an adaptive system that can change its structure based on external or internal information flowing through the network. From an engineering standpoint, ANNs can be viewed as non-linear statistical data modeling tools. They can be used to model complex relationships between inputs and outputs or to find patterns in data. These networks are also similar to the biological neural networks in the sense that functions are performed collectively and in parallel by the units, rather than there being a clear delineation of sub-tasks to which various units are assigned. While complete description of biological neural systems has not been attained, progress is being made towards a better understanding of their mechanisms.
Scientists
Fuzzy System
Introduction
Fuzzy logic is a mathematical approach to represent vagueness in linguistics, and can be considered as a generalization of classical set theory. The main difference between crisp and fuzzy sets is that the former always have unique membership functions, whereas every fuzzy set has an infinite number of membership functions that may represent it. One key feature of fuzzy logic is its ability to express the degree of ambiguity in human thinking and subjectivity (including natural language) in a comparatively elegant manner. With these properties, fuzzy systems find their applications in a wide range of areas including: control system, pattern recognition (e.g., image, audio, signal processing), quantative analysis (e.g., operational research), logical inference (e.g., expert systems for diagnosis, planning, and prediction; natural language processing; intelligent interface; intelligent robots; software engineering, and information retrieval (e.g., medical diagnosis).
Scientists
Evolutionary Computation
Introduction
Evolutionary computation (EC) is a subfield of computational intelligence involving search, learning, optimization, and modeling methods loosely inspired by biological evolution. The popular methods include genetic algorithm (GA), evolutionary strategy (ES), and evolutionary programming (EP). All such methods start with a population of randomly generated candidate solutions (termed individuals), and perform fitness-based selection and random variation to create a new population. The two main sources of genetic variation in this case include the crossover and mutation operations, which mimic the biological sexual recombination and mutation respectively. EC methods have been widely applied, with major examples such as numerical parameter and combinatorial optimization, automatic generation of computer programs, bioengineering, financial prediction, robot learning, evolving production systems for artificial intelligence applications, as well as design and training of neural networks.
