Cognitive Intelligence: Deep Learning, Thinking, and Reasoning by Brain-Inspired Systems

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/310613903
Cognitive Intelligence: Deep Learning, Thinking, and Reasoning

by Brain-Inspired Systems
Article in International Journal of Cognitive Informatics and Natural Intelligence · November 2016

DOI: 10.4018/IJCINI.2016100101
CITATIONS READS
40 4,560
10 authors, including:
Yingxu Wang Newton Howard

The University of Calgary University of Oxford
706 PUBLICATIONS 9,972 CITATIONS 181 PUBLICATIONS 2,200 CITATIONS
SEE PROFILE SEE PROFILE
Sally Wood Virendrakumar C. Bhavsar

Santa Clara University University of New Brunswick
56 PUBLICATIONS 502 CITATIONS 131 PUBLICATIONS 937 CITATIONS
SEE PROFILE SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Autonomous Systems (AS) View project
Image Pattern Recognition and Voronoi Diagrams View project
All content following this page was uploaded by Yingxu Wang on 01 February 2018.
The user has requested enhancement of the downloaded file.

International Journal of Cognitive Informatics and Natural Intelligence
Volume 10 • Issue 4 • October-December 2016
Cognitive Intelligence:
Deep Learning, Thinking, and Reasoning
by Brain-Inspired Systems
Yingxu Wang, Schulich School of Engineering and Hotchkiss Brain Institute, University of Calgary, Calgary, Canada
Bernard Widrow, Stanford University, Stanford, CA, USA
Lotfi A. Zadeh, UC Berkeley, Berkeley, CA, USA
Newton Howard, University of Oxford, Oxford, UK
Sally Wood, Santa Clara University, Santa Carla, CA, USA
Virendrakumar C. Bhavsar, University of New Brunswick, Fredericton, Canada
Gerhard Budin, Vienna University, Vienna, Austria
Christine Chan, University of Regina, Regina, Canada
Rodolfo A. Fiorini, Politecnico di Milano University, Milano, Italy
Marina L. Gavrilova, University of Calgary, Calgary, Canada
Duane F. Shell, University of Nebraska-Lincoln, Lincoln, NE, USA
ABSTRACT
The theme of IEEE ICCI*CC’16 on Cognitive Informatics (CI) and Cognitive Computing (CC) was
on cognitive computers, big data cognition, and machine learning. CI and CC are a contemporary
field not only for basic studies on the brain, computational intelligence theories, and denotational
mathematics, but also for engineering applications in cognitive systems towards deep learning, deep
thinking, and deep reasoning. This paper reports a set of position statements presented in the plenary
panel (Part I) in IEEE ICCI*CC’16 at Stanford University. The summary is contributed by invited
panelists who are part of the world’s renowned scholars in the transdisciplinary field of CI and CC.
Keywords
Applications, Artificial Intelligence, Brain-Inspired Systems, Cognitive Computers, Cognitive Engineering,
Cognitive Informatics, Cognitive Robotics, Cognitive Systems, Computational Intelligence, CWW, Deep
Learning, Deep Reasoning, Deep Thinking, Denotational Mathematics, Knowledge Learning
1. INTRODUCTION
Cognitive Informatics (CI) is a transdisciplinary enquiry of the internal information processing

mechanisms and processes of the brain and abstract intelligence, as well as their applications in
cognitive computing and cognitive engineering (Wang, 2002a, 2003, 2006, 2007a; Wang et al., 2002,
2009a/b, 2010, 2011a). CI is a contemporary field spanning across computer science, information
science, cognitive science, brain science, neuroscience, intelligence science, knowledge science,
robotics, cognitive linguistics, cognitive philosophy, and cognitive engineering. Cognitive Computing
(CC) is a novel paradigm of intelligent computing platforms of cognitive methodologies and systems
based on CI, which embodying computational intelligence by cognitive and autonomous systems
mimicking the mechanisms of the brain (Wang, 2009b, 2012e, 2015b, 2016a; Wang et al., 2006).
IEEE ICCI*CC’16 on Cognitive Informatics and Cognitive Computing has been held at Stanford
DOI: 10.4018/IJCINI.2016100101
Copyright © 2016, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.
1
University during Aug. 22-23, 2016. The theme of ICCI*CC’16 was on cognitive computing, big
data cognition, and machine learning (Widrow, 2016; Zadeh, 2016; Wang et al., 2016b).
The IEEE series of International Conferences on Cognitive Informatics and Cognitive Computing
(ICCI*CC) has been established since 2002 (Wang, 2002a). The inaugural ICCI event in 2002 was held
at University of Calgary, Canada (ICCI’02) (Wang et al., 2002), followed by ICCI’03 in London, UK
(Patel et al., 2003); ICCI’04 in Victoria, Canada (Chan et al., 2004); ICCI’05 in Irvine, USA (Kinsner
et al., 2005); ICCI’06 in Beijing, China (Yao et al., 2006); ICCI’07 in Lake Tahoe, USA (Zhang et al.,
2007); ICCI’08 at Stanford University, USA (Wang et al., 2008); ICCI’09 in Hong Kong (Baciu et
al., 2009); ICCI’10 at Tsinghua University, Beijing, China (Sun et al., 2010); ICCI*CC’11 in Banff,
Canada (Wang et al., 2011a); ICCI*CC’12 in Kyoto, Japan (Sugawara et al., 2012), ICCI*CC’13 in
New York, USA (Hsu et al., 2013), ICCI*CC’14 in London, UK (Patel et al., 2014), ICCI*CC’15 at
Tsinghua University, Beijing, China (Ge et al., 2012), and ICCI*CC’16 at Stanford University, USA
(Wang et al., 2016). Since its inception, the ICCI*CC series has been growing steadily in its size, scope,
and depth. It has attracted worldwide researchers from academia, government agencies, and industry
practitioners. The conference series provides a main forum for the exchange and cross-fertilization
of ideas in the new research field of CI toward revealing the cognitive mechanisms and processes of
human information processing and the approaches to mimic them in cognitive computing. A wide
range of breakthroughs have been recognized and a wide range of applications has been developed in
CI and CC in the last decade. The representative paradigms and technologies developed in cognitive
informatics include cognitive computers, abstract intelligence, cognitive learning engines, cognitive
knowledge bases, denotational mathematics, and applied cognitive systems.
Fundamental theories of CI and CC cover the Matter-Energy-Information-Intelligence (MEII)
model (Wang, 2015c), the Layered Reference Model of the Brain (LRMB) (Wang et al., 2006), the
Object-Attribute-Relation (OAR) model of internal information representation in the brain (Wang,
2007c), the Cognitive Functional Model of the Brain (CFMB) (Wang & Wang, 2006), Abstract
Intelligence (Wang, 2009a, 2012c), the Abstract Intelligence Model of the Brain (AIMB) (Wang,
2012c), Neuroinformatics (Wang, 2013a; Wang and Fariello, 2012), Denotational Mathematics
(Wang, 2007b, 2008, 2012a, 2015a), Cognitive Linguistics (Wang and Berwick, 2012, 2013), the
Spike Frequency Modulation (SFM) Theory of neural signaling (Wang, 2016e), the Neural Circuit
Theories (Wang, in press; Wang and Fariello, 2012), and cognitive systems (Wang, 2010a, 2014a).
Recent studies on LRMB in CI reveal an entire set of cognitive functions of the brain and their
cognitive process models, which explain the cognitive mechanisms and processes of the natural
intelligence with 48 cognitive processes at seven layers known as the sensation, action, memory,
perception, cognitive, inference, and intelligence layers (Wang et al., 2006).
This paper is a summary of the position statements of invited panellists presented in the Plenary
Panel on Perspectives on Cognitive Computing, Big Data Cognition, and Machine Learning (Part I),
which was held in IEEE ICCI*CC 2016 (Wang et al., 2016b/c) at Stanford University, USA, on Aug.
23, 2016. It is noteworthy that the individual statements and opinions included in this paper may not
necessarily be shared by all panellists.
2. A THEORETIC FRAMEWORK OF KEY COGNITIVE INTELLIGENCE

OF DEEP LEARNING, DEEP THINKING, AND DEEP REASONING
Learning is a cognitive process of inductive knowledge acquisition, behavior generation, and

intelligence aggregation (Dietterich et al., 2008; Kober & Peters, 2010; Morimoto, 2010; Mehryar
et al., 2012; Widrow, 2016; Widrow & Aragon, 2013; Widrow et al., 2013, 2015, Lecun et al., 2015;
Schmidhuber, 2015; Zadeh, 2016; Wang, 2013, 2014, 2015, 2016a/c; Wang & Berwick, 2013; Wang
et al., 2001, 2016). The relationship between learning and other key cognitive processes in the brain
is explained by the Layered Reference Model of the Brain (LRMB) (Wang et al., 2006).
2
The taxonomy of machine learning methodologies can be classified into six categories (Wang,
2015b, 2016a/b), i.e.:
 Li (x, P | x ∈ P )  x = P.x // 1) Object identification


 Lc ( X, P)  X ⊂ P // 2) Cluster classificattion
 L ( X, P)  X = P
 p // 3) Pattern recognition
 (1)
 Lr ( X, P)  X ⇒ P( X) // 4) Functional regression
 L ( X, P)  X ⇒ b(P( X)) // 5) Behavior generation
 b
 Lk ( X, P)  X ⇒ k (P( X))  K // 6) Knowledge acquisition
where X, P, K, and ⊎ denote a target vector, a pattern vector, a framework of existing knowledge,
and knowledge composition, respectively.
Classic machine learning deals with object identification (Li), cluster classification (Lc), pattern
recognition (Lp), functional regression (Lr), and behavior (game) generation (Lb) (Dietterich et al.,
2008; Maulik & Bandyopadhyay, 2010; Mehryar et al., 2012; Wang, 2015b, 2016a). It is recognized
that classic machine learning technologies cannot sufficiently solve the fundamental problems in
the sixth category of Equation (1) for common knowledge learning (Lk) (Wang, 2016a-c), which is
a fundamental challenge to convey human learning mechanisms into machines.
Knowledge learning is the largest category of learning for knowledge acquisition beyond classic
machine learning for data regression, image recognition, or game playing. Knowledge learning is
not only dependent on precise concept manipulations as those of thinking threads in short-term
memory (STM), but also supported by a cognitive knowledge base embodying existing knowledge
in long-term memory (LTM) (Wang et al., 2006a). This learning mechanism of the brain reveals the
main challenges to traditional machine learning approaches embodied at the low neural level to deep
knowledge acquisition and reinforced learning.
The fundamental challenges to knowledge learning have led to the emergence of the field of
cognitive machine learning. In recent basic studies, it is recognized that the theoretical framework
of general cognitive systems and brain-inspired systems, such as cognitive robot (Wang, 2010a) and
cognitive computers (Wang, 2009b, 2012b/e), requires a significant cognitive ability for deep thinking
and reasoning (Reisberg, 2001; Wang, 2016a; Wang et al., 2006) beyond classic machine learning.
Cognitive knowledge learning as a central ability of cognitive systems cannot be implemented merely
at the neural level. Instead, it is deeply dependant on higher functions such as a cognitive knowledge
base (Wang, 2014a) and a cognitive learning engine (Wang, 2015b, 2016c) to facilitate deep learning
mimicking the brain.
Typical mathematical means for classic machine learning include predicate logic (Poole et al.,
1997), statistical methods (Bender, 1996), neural networks (Hinton et al., 1995), genetic algorithms
(Goldberg, 1989), fuzzy logic (Zadeh, 1975), regression functions (Widrow et al., 2015; Gilat
& Subramaniam, 2010), semantic computing (Sheu, 2007; Wang, 2010a, 2013b), support vector
machines (Karasuyama & Takeuchi, 2010), and recurrent neural networks (Gers & Schmidhuber, 2001;
Lecun et al., 2015). However, cognitive machine learning is underpinned by recent breakthroughs
in denotational mathematics (Wang, 2008a, 2009c, 2012a) and mathematical engineering (Wang,
2015a). The challenges to cognitive knowledge learning are deeply rooted in the fact that almost all
AI problems are out of the domain of ℝ designated in classic mathematics (Wang, 2012). As new
types of problems require new forms of mathematics, denotational mathematics (Wang, 2009c, 2012)
is created for knowledge learning, deep reasoning, and abstract intelligence such as concept algebra
(Wang, 2015d), semantic algebra (Wang, 2013b), real-time process algebra (Wang, 2002b), inference
3
algebra (Wang, 2011b), fuzzy probability algebra (Wang, 2015e), and big data algebra (Wang, 2016d),
as well as basic theories of software science (Wang, 2014b) and system science (Wang, 2015f).
It was recently discovered that the basic structural model of human knowledge is a formal concept
in cognitive linguistics (Wang, 2015d; Wang & Berwick, 2013), and the basic unit of knowledge
is a binary relation (bir) (Wang, 2016b) as a counterpart of bit for information and data (Shannon,
1948). Supplementary to human empirical learning mechanisms, machine created knowledge bases
can be mutually shared to facilitate a new approach to cloned knowledge learning (Wang, 2016a/c).
These basic studies have led to breakthroughs on: a) The architectural and behavioral reference
models of cognitive robots (Wang, 2010) and cognitive systems (Wang, 2010a, 2012b, 2016c; Wang
& Ruhe, 2007); b) A cognitive model of brain functions (Wang & Wang, 2006); c) Cognitive models
of knowledge science (Wang, 2016b); and d) Abstract intelligence theories (Wang, 2009a, 2012c)
towards novel technologies of cognitive computers (Wang, 2009b, 2012b/e) and cognitive systems
(This section is contributed by Prof. Yingxu Wang.)
3. HEBBIAN LEARNING AND THE LMS ALGORITHM
Hebbian learning is one of the fundamental premises of neuroscience. The LMS (least mean square)
algorithm of Widrow and Hoff is the world’s most widely used learning algorithm. Hebbian learning
is unsupervised. LMS learning is supervised. However, a form of LMS can be constructed to perform
unsupervised learning and to implement Hebbian learning. Combining the two paradigms creates a
new unsupervised learning algorithm that has practical engineering applications and provides insight
into learning in living neural networks.
A fundamental question is, how does learning take place in living neural networks? The learning
algorithm practiced by nature at the neuron and synapse level may well be the Hebbian-LMS algorithm
(Widrow, 2016, Widrow & Aragon, 2013; Widrow et al., 2013, 2015). (This section is contributed
by Prof. Bernard Widrow.)
4. A KEY ISSUE OF SEMANTICS OF INFORMATION
In his epoch-making work on information theory, Shannon defended information in terms of entropy.
Entropy-based definitions of information relate to quantity of information, but not to its meaning.
Subsequent attempts to introduce semantics into information theory have made some progress but
fell short of having a capability to deal with information described in natural language. This section is
aimed that of laying information for the theory which has this capability, call it a theory of semantic
information (TSI).
TSI is centered on a concept which plays a key role in human intelligence – A concept whose
basic importance has long been and continues to be unrecognized – The concept of a restriction is
pervasive in human cognition. Restrictions underlie the remarkable human ability to reason and make
rational decisions in an environment of imprecision, uncertainty and incompleteness of information.
Such environments are the norm in the real-world. Such environments have the traditional logical
systems that become dysfunctional. There are many applications in which semantics of information
plays an important role. Among such applications are: machine translation, summarization, search
and decision-making under uncertainty. Informally, a restriction on a specified (focal) variable, X,
written as R (X), is a statement which is a carrier of information about the values which X can take.
Typically, restrictions are described in natural language. Example. X = length of time it takes to drive
from Berkeley to SF Airport; R(X) = usually it takes about 90 minutes to drive from Berkeley to SF
Airport. In adverse weather it may take close to 2 hours. An important issue in TSI is computation
with restrictions. TSI opens the door to modes of computation in which approximation is accepted.
Acceptance of approximate computations takes the calculus of restrictions (CR) into uncharted
territory (Zadeh, 1965, 1975, 1997, 2004, 2016). (This section is contributed by Prof. Lotfi A. Zadeh.)
4
5. COGNITIVE COMPUTING IN MEDICAL IMAGING:

OPPORTUNITIES AND CHALLENGES
Cognitive computing applied to medical images has great potential in both diagnostic and research
contexts, but it also faces significant challenges. The wide range of possible beneficial results includes
assisting physicians, complementing physicians, improving the quality of medical care and medical
outcomes, reducing the costs of delivering care, and developing new understanding of disease processes
and variability across diverse populations. Physicians could be assisted by fast and comprehensive
screening of large image data sets to rapidly and accurately identify the most significant elements.
This task is well suited to cognitive computing since visualization of 3D and 4D data sets is not a
normal human visual experience, and reviewing large data sets can be time consuming and require both
effective visualization tools and experienced reviewers. In addition, cognitive computing could be used
for large scale longitudinal comparisons as well as correlations with other medical data not derived
from images that would otherwise be difficult or impossible to do at scale. The results might detect
disease conditions at earlier and more manageable stages and allow wider correlation of information
over time and across “similar” patients. The potential for improved care and outcomes would be
possible due to more pervasive monitoring and integration of information from diverse data sets.
The size and variability of medical imaging data sets (Shaw, 2010) can be a significant challenge.
There is a wide range of modalities and protocols for functional and anatomical imaging from two
dimensional projection X-rays and surface skin photographs, to three dimensional volume images
from computed tomography (CT) and magnetic resonance imaging (MRI), to higher dimensional
image sets which include time variation for alternate protocols, contrast studies, motion studies, or
longitudinal comparisons. When multimodal comparisons or data fusion are considered, for example
using both structural MRI and functional MRI with electroencephalogram data (Adali et al., 2015),
the dimensionality is further expanded.
Challenges to successful application of cognitive computing to medical imaging include the
data set sizes and variability and the acquisition of sufficient training data. Image data sets are large
compared to text data sets and many types of physiological measurements such as blood chemistry
and EKGs. For example, a single 3D breast CT image set may be (1024)3 voxels (2 GB), and 100 sets
might be used for one examination using contrast (200 GB) (Shaw, 2010). To make inferences for
single patient using the full context of all medical examinations for that patient, many different image
data sets using multiple modalities, e.g. CT and PET, or multiple scales, e.g. blood vessels and full
torso, may be considered. Often clinical images for a patient are acquired for the specific local area
of most concern, not a wider context. Extracting information for such diverse data sets which may
image different anatomical and functional characteristics of nonrigid volumes with different protocols
using different scales and orientation presents challenges for training. Large coordinated systems
of data access are needed to provide enough reliable data for training and validation. Recently the
IBM Watson Health Medical Imaging Collaborative has brought together a large number of health
systems and academic medical systems to train a potentially personalized medical analysis system
(IBM, 2003). (This section is contributed by Prof. Sally Wood.)
6. BIG DATA COGNITION: GRAPH SIMILARITY AND

NATURAL LANGUAGE PROCESSING
Big Data in the world is composed of structured, semi-structured as well as unstructured data. In Big
Data cognition, one of the fundamental aspect is data representation. The data may be represented
starting from simple discrete structures such as sets to the complex structures such as trees and graphs.
Trees and graphs allow representation of complex concepts.
Similarity plays a central role in many cognitive capabilities and applications (see for example
Wang et al. 2011, Wang et al. 2015). Our main research hypothesis is the following: The use of weighted
5
and highly structured object representations will result in a more precise similarity computation
between objects compared to the unweighted and less structured representations. The meaning of a
weight on an edge (directed or undirected) from say node A to node B depends on the application
domain. For example, the weight could mean: (a) the frequency of occurrence of node B given A, (b)
information content of node B, (c) strength of a relationship between the two nodes, (d) the degree
of preference of a user for B, or (f) the probability of B given A. We have developed novel similarity
algorithms for weighted structures: sets (Joshi M. et al. 2010a, 2010b), trees (Bhavsar V.C. et al.
2004; Yang. L. et al. 2005, 2008; Boley H. et al. 2005) and weighted directed acyclic graphs (DAGs)
and attributed generalized trees (AGTs) (Kiani et al. 2014a, 2014b, 2015, 2016). The weights for
nodes and edges may represent relative importance, information content, probabilities, strength of
relationships, user preferences, or other attributes as appropriate for applications. The weights may
be crisp or fuzzy. AGTs are generalized hierarchical tree representations that allow representations
complex relationships in many domains, e.g. insurance underwriting (Kiani et al. 2014a) and electronic
medical records (EMRs) (Kiani et al. 2013, 2014b). We have introduced the concept of simplicity
(opposite of complexity) in the computation of similarity (Bhavsar et al. 2004, Kiani et al. 2016). Our
precise and structured representations give more precise similarity compared to mostly unweighted
representations proposed by others. To be able to process Big Data sets, we have implemented some
of the similarity computations on computing clusters and general purpose graphics processing units
(GPGPUs) and have obtained good speedups (Lu, Y. et al. 2016). Recently, we are working on graph
data bases for storing AGTs as well as general graphs, and their parallel/distributed implementations.
Big Data analytics often involves the use machine learning techniques for natural language
processing of textual data arising for example in the Web, databases and social media. The identification
and understanding of Multiword Expressions (MWEs) are very important problems in parsing
and generation of natural languages, as well as many other language technology applications (e.g.
machine translation). We have used deep learning techniques (LeCun et al. 2015) for identifying
token instances of Verb–noun idiomatic combinations. Our proposed supervised and unsupervised
approaches (Gharbieh et al. 2016) perform better than state-of-the-art approaches. At present, we
are experimenting with more complex deep learning models to obtain high performance results on
other types of data sets. Visualization of textual data can be useful in many applications. We have
designed a visualizer, named Vishit, for visualizing texts in the Hindi language used in many states
of India (Jain et al. 2013, 2014). Finally, we have also carried out work on the parallel/distributed
implementations of natural language processing systems on high performance computing platforms,
such as multicore grid, cluster platforms Hadoop (Tomar et al. 2013a, 2013b, 2014).
There are many challenges and opportunities in Big Data cognition using graph similarity
(e.g. domain-specific similarity measures and scalability), natural language processing (e.g. natural
language understanding and visualization), and their efficient implementations on peta- and exa-scale
supercomputing and high performance computing platforms. Further, many efforts are underway to
overcome limitations of deep learning, for example in provenance and class generalization. (This
section is contributed by Prof. Virendrakumar C. Bhavsar.)
7. COGNITIVE COMPUTING IN MODELING THE CARBON

DIOXIDE CAPTURE PROCESS SYSTEM
The research objective is to better understand how to operate the post-combustion carbon dioxide
capture process (CDCP) system, which is implemented at the Clean Energy Technology Institute at
University of Regina, Saskatchewan, Canada. The carbon capture technology is widely recognized to
be an important approach for mitigating greenhouse gas emissions from fossil-fueled power plants. In
this study, different machine learning approaches are used for deriving insights from data generated
by the CDCP system.
6
Cognitive computing can be viewed as developing software that mimics functioning of the human
brain. Applying cognitive computing to modeling data from the CDCP system involves deriving
insights that mimic human expertise in operation of the carbon capture system. The approach adopted
is to develop predictive models that describe the relationships among key parameters of the CDCP
system. These predictive models can be used for problem solving in system operations. In developing
these models, important contextual knowledge was captured from experts when they identified the
important predictor parameters, discussed ways to segment the dataset, and refined the input parameter
set after the initial modeling. The experts were also involved in validating the discovered rules.
Applying cognitive computing to the project also requires the modeling results to be adaptive to
user needs. The knowledge discovered from the CDCP system needs to adapt its format and content
to the specific context of problem solving. For example, when the statistical regression approach
was adopted, the regression equations generated were explicit but their accuracy was not satisfactory.
When the artificial neural network or ANN approach was adopted, the predictive models generated
had satisfactory predictive accuracies, but they were black boxes. When the adaptive-neural inference
system or ANFIS approach was adopted, the predictive models generated have satisfactory accuracies,
but the models are not explicit (Zhou et al., 2002). Since some scenarios of problem solving require
explicit knowledge on relationships among the key parameters of the process system, the integrated
approach of ANN and rule extraction was adopted. By adopting the decompositional approach of
rule-extraction, which opens up the black box, explicit equations from the highly accurate ANN
predictive models generated from the data can be produced (Chan, 2016). The explicit equations then
can be validated by the experts. (This section is contributed by Prof. Christine Chan.)
8. THE LOGICAL RELATIONSHIP BETWEEN EXPERIENCE

AND KNOWLEDGE EXTRACTION
Cognitive Informatics (CI) (Wang, 2012a) is a transdisciplinary enquiry of computer science,

information sciences, cognitive science, and intelligence science that investigates into the internal
information processing mechanisms and processes of the brain and natural intelligence, as well
as their engineering applications in cognitive computing (CC) and cognitive robotics (CR). In CI,
Denotational Mathematics (DM) and the LRMB (Layered Reference Model of the Brain) (Wang,
2012e; Wang et al., 2006) provide an integrated framework for modeling the brain and the mind.
LRMB also enables future extension and refinement of the CPs (Cognitive Processes) within the
same hierarchical framework. LRMB can be applied to explain a wide range of physiological,
psychological, and cognitive phenomena in CI. Particularly the relationships and interactions between
the inherited and the acquired life functions, those of the subconscious and conscious CPs, as well as
the dichotomy between two modes of thought: “System 1”, fast, instinctive and emotional; “System
2”, slower, more deliberative, and more logical (Kahneman, 2011), can be focused quite well. LRBM
is a powerful and flexible framework and its potential has yet to be fully exploited by current and
near future computer systems.
In fact, the real performance of current, applied cognitive systems is still very far from the
requirements of effectiveness, robustness, compactness and autonomy, necessary for a meaningful and
skilled interaction with the real world. The latest example comes from Novartis where deep machine
learning tools are vastly applied. On August 2016, Novartis dissolved its high-profile cell and gene
therapy unit operating under the guiding hand of Usman “Oz” Azam (Reuters, 2016). Usually, this is
the final result of research planning based on assumed trustworthy mathematical tools that they are
not, unfortunately. According to physicist Jose G. Vargas “All physical concepts are to be viewed as
emergent. Emergence of this type and reductionism are as inseparable as the two faces of an ordinary
surface. In advanced research, failure comes from lacking of effective and reliable mathematical
model and tools to deal with system emergent behavior” (Vargas & Torr, 2015).
7
How does it come that “science 1.0” (reductionist statisticians) are still in business without having
worked out a definitive solution to the problem of the logical relationship between experience and
knowledge extraction? There is a strong need to develop more effective and reliable experimental
observation correlates by the correct arbitrary multiscale system (AMS) modelling approach for
complex system understanding and automatic knowledge extraction (Fiorini, 2015). To grasp a more
reliable representation of reality and to get stronger cognitive, biological and physical system correlates,
researchers and scientists need to raise the traditional statistical veil in advanced AMS modeling.
As a matter of fact, scientific, computational and simulation current classic systemic resources and
most sophisticated instrumentation system (developed under the positivist reductionist paradigm and
the “continuum hypothesis”, CH for short) are still totally unable to capture and to discriminate so
called “random noise” from any combinatorically optimized encoded message, called “deterministic
noise”, as evidenced by the computational information conservation theory (CICT), developed at
Politecnico di Milano University since the 1990s (Fiorini, 2014). This is the information double-
bind (IDB) problem in current science and scientific community (but nobody likes to talk about it).
Cognitive ambiguity emphasizes this major IDB problem in most advanced research laboratory and
instrumentation system, just at the inner core of human knowledge extraction by experimentation in
current science (Fiorini, 2014). At IEEE ICCI*CC’16 we presented a paper showing sound examples
of the IDB problem in science (Fiorini, 2016). We showed that in any random process the presence
of deterministic computational structures can never be ruled out, even in Gaussian noise. Therefore,
complete randomness is an information theory abstraction only, paired by no physical process. Usually,
the deterministic contribute (developed under the combinatorical paradigm and the “discreteness
hypothesis”, DH for short) is completely overwhelmed by the random ones which are more numerous;
but CICT can provide us with the right tools to let it emerge out of the digital pseudorandom noise
components (Fiorini, 2014, 2015).
To address effectively the above modelling issue, we need to become deeply aware that human
beings and most traditional, advanced computational and scientific measurement systems have quite
a limited capacity to extract reliable information from noisy sources and generators currently. To
solve the IDB problem in science, we need two intelligently articulated hands: both statistic (CH)
and combinatorical (DH) approaches synergistically articulated by natural coupling. We need a new,
fresh “Science 2.0” approach. Therefore, to achieve reliable system intelligence outstanding results,
current computational system modelling and simulation community has to face and to overcome two
orders of issues at least, immediately:
a. To minimize the traditional limitation of current digital computational resources that are unable
to capture and to manage even the full information content of a single Rational Number Q exactly,
leading to information dissipation and opacity.
b. To develop stronger, more effective and reliable correlates by correct AMS modelling approach
to complex system.
The CICT framework addresses both issues at the same time offering simple and competitive
solutions to them. This is the main reason CICT can offer a strong and effective contribution to the
whole cognitive community, to Abstract Intelligence (AI), CI, CC, CR and to the emerging cognitive
engineering (CE), exponential arithmetic (EA) and mathematical engineering (ME) fields. Main
goal of the present note is to disseminate the awareness and education on the current availability of
new ways and computational tools to elicit applied learning to increase innovation and creativity
in all traditional disciplines. To find new competitive solution to old, current cognitive problems.
The reader interested in deeper computational details is referred to (Fiorini, 2015). (This section is
contributed by Prof. Rodolfo A. Fiorini.)
8
9. IMPACT OF SOCIAL BIOMETRICS AND COGNITIVE

INTELLIGENCE ON PRIVACY AND SECURITY
Our society continues to undergo tremendous growth with respect to all aspects of information access
and sharing. It had a profound effect on the way we, humans, and the whole society lives, works and
interacts in business and social settings. The terabytes of information being shared through social
networks, on-line communities, games, software development tools, e-mails, blogs, posts, etc., is
enormous. It also ranges in type: text, images, hyperlinks, likes, network connections, etc. What’s
more, human social, behavioral and even cognitive traits are becoming more and more visible through
interlinking of heterogeneous communications in on-line and off-line settings. This phenomenon gave
rise to the rise of a new concept: Social Biometrics, that attempts to understand and extrapolate trends
related to all aspects of human social interactions. This position paper is devoted to definitions, case
studies and state-of-the-art multidisciplinary research in social biometrics.
More and more we find ourselves being part of a social network, being it a professional, personal
or recreational endeavors. For majority of human population, professional skills, interests, hobbies,
finances and even travel profiles become not only integrated, but publicly available to a vast majority
of population. Moreover, even such personal items as prescriptions or favorite brands become
inadvertently part of big data, and thus are subject to all the processing, data mining and pattern
recognition. This is particularly visible in the cyberworld context, where human online identities
prominently showcase themselves. Not surprising, such areas as big data analytics, decision-making,
decision fusion, artificial intelligence, pattern recognition and biometrics, now have a permanent
presence in the online domain. The particularly notable examples of such research are cognitive
multi-modal security system architecture (Wang et. al. 2012), establishing identity of Twitter users
through social networks analysis (Sultana et. al. 2014b), and gender recognition of Flickr users based
on human aesthetic preferences (Azam & Gavrilova 2016).
Over the past few years, there has been a steady rise in research focused on analysis of users
social networks (Facebook, Twitter etc.), with the goal of understanding user preferences and
interactions based on the spheres of their interests or their affiliation (Paul et. al. 2014, Sourin et.
al. 2016). A new domain of particular interest for social biometrics is computational aesthetics
– understanding someone’s preferences and interests with the goal of building aesthetic profiles
(Sagalin et.al. 2014, Azam & Gavrilova 2016). In big data analytics domain, web browsing histories,
buyers preferences and online communications of millions of users are being collected and analyzed
on a daily basis. On the other hand, in security research, social behavioral biometric authentication
as well as artificial biometrics emerged as powerful research tools, based on the carefully selected
features extracted using machine intelligence and deep learning from recorded history of user online
interactions (Yampolskiy and Gavrilova 2012, Sultana et.al. 2014a). Traditional scope of biometric
research usually assumes recognizing someone’s identity from biometric data (Jain et al. 2004),
that traditionally was limited to physiological and behavioural biometrics. Physiological features
(face, iris, retina, ear) can be often collected by specialized image or video capturing devices, such
as infrared sensors, remote temperature measuring devices, and so on. Behavioural characteristics,
which traditionally include the way a person walks (gait), the way a person talks (voice), the way
a person writes (typing patterns, keystroke pressure), or the way a person authenticates documents
(signature), can be obtained from variety of sensors, such as KINECT sensor or signature tablets.
Soft biometrics include easily collected but not as unique as the previous two biometric data, i.e. age,
gender, height, weight, eye colour or hair colour of a person. Even the way a person socially interacts
or expressed their preferences can become a rich source of authentication information. Finally, a
social biometric area of research was proposed as a new way to obtain supplementary but sometimes
crucial for authentication information (Bocenna et al 2014, Sultana et.al. 2014b). In this domain,
social behavioural features are extracted from the way users interact through various social on-line
and off-line networks. This includes user’s online presence patterns (time, day, month), the nature
9
of interaction (tweets, blogs, chats), the content of interaction (topics, likes, opinions), online game
playing strategies, virtual world avatar preferences, etc. (Sultana et al. 2014a, Sultana et. al 2015).
One of the main features, which is crucial for research on the social behavioural biometrics, is the
communication patters in the networks of users and the composition of such networks themselves.
It is generating a lot of interest and getting traction in biometric research, as well as in related fields
looking into human interaction, physiological studies, user profiling, pattern recognition, authorship
identification, and collective intelligence (Sultana et al. 2014a, Gavrilova and Yampolskiy 2011). The
idea can be transferred to the real world as well. For instance, in a given social context some social
behavioural patterns such as friends and acquaintances, daily routine, psychological states, style
of conversation, gestures and emotions during conversations, preferences, spatial information, and
activity logs can play important role in recognizing a person. Such patterns can also provide a unique
set of tools to promote better understanding of collaborative processes, workflow dynamics and risk
analysis in collaborative/social environments in real and virtual worlds. With staggering amount of
information gathered and analyzed every day, it is impossible to underestimate value of privacy of
the data. Thus, another aspect of recent research is focused on protecting the data and the biometric
template from unauthorized access, forgery or tampering (Paul et. al. 2014).
The emergence of social biometric testifies to the increasing importance of the degree of inter-
connectedness we observe in the social online settings. Recent research shows that utilizing cognitive
informatics and machine intelligence can reveal hidden patterns of behavior, which allow to identify
the person or its gender from such soft preferences as favorite picture, music, movies, favorite
places to visit, hobbies, style of texting or a circle of friends. Combining this social information
with physiological or behavioral data as part of multi-modal biometric system can provide an
unparalleled degree of precision. The challenge remaining is protecting this crucial information and
thus in developing privacy-preserving policies, such as template protection, to be a required part or
any security system. (This section is contributed by Prof. Marina L. Gavrilova.)
10. CHUNKING AND AUTOMATICITY: HANDLING BIG

DATA IN HUMAN COGNITION AND THE BRAIN
How does a human drive a car? If one takes into account the speed at which neural signals are
transmitted and the time it takes to process sensory information, attend to it, and consciously process
it, it should be impossible to drive a car, especially at high speeds. Perhaps the real question isn’t
“why are there so many accidents?” rather “why aren’t there a lot more accidents?”.
These questions are central not just to whether “self-driving” cars are possible; but also to the
fundamental nature of how humans learn and behave. They are also central to the ICCI*CC*16
conference themes of “Perspectives on Cognitive Computing, Big Data Cognition, and Machine
Learning.” Perhaps the “biggest” of all big data cognition questions, is how does the human brain
manage the vast, almost overwhelming, amount of data that is input continuously through the senses
and from internal proprioceptive and neural feeds to make meaning and effectively take action in
the world.
Psychological and neurological theory has answered these questions by proposing two parallel
mechanisms: chunking, for sensory, and conceptual or declarative knowledge, and automaticity, for
behavior or procedural knowledge (Gobet & Lane, 2012; Moors, 2016). Both processes involve data
compression. Chunking being the reduction of sensory information into smaller units of meaning
and automaticity being the linking of behaviors to specific sensory patterns in ways that eliminate
intermediate steps between the input and behavior; thus producing the stimulus – response (S – R)
connections studied in classic behavior psychology. Through this data compression, chunking allows
fast retrieval of large amounts of information beyond what would be available given the storage
limitations of working memory, and automaticity allows rapid behavioral reaction to immediate
sensory data without the need for conscious deliberation or problem solving.
10
In the Unified Learning Model (Shell et al., 2010), we have advanced the theory that chunking
and automaticity are accomplished in the human brain through statistical learning. In the ULM,
each perceived sensory input in working memory constitutes a sample of the possible attributes of
the objects or events in the physical world. Neural assemblies act as frequency counters, recording
the repetition or non-repetition of attributes across multiple sensory experiences. As the number of
sensory experiences (samples) increases, the frequency counts of attributes in the neural assemblies
will move closer and closer to the actual frequencies of attributes in the object or event via the “law of
large numbers.” Thus, the neural assembly is a data compression mechanism that reduces the massive
amount of experienced sensory data to a probabilistic frequency distribution that provides the best
representation of the information that was in the original sensory data. When applied to hierarchical
neural structures, the neural strengthening and pruning via Hebbian neural plasticity (Hebb, 1948)
means that the neural assemblies in each hierarchical level are tuned to provide the feeds of the
highest probability frequencies to the next higher level, such that each level moves toward the most
parsimonious data compression.
In our studies using the Computational-Unified Learning Model (C-ULM), we have shown that
computational implementation of the ULM statistical learning processes of motivated attention,
repetition, and connection in a multiagent simulation can produce effective learning of concepts
required for task solution and application of this knowledge to solving tasks, with this learning
mirroring the learning curves observed in human learning (Chiriacescu, Soh, & Shell, 2013). We
have also shown that self-efficacy or confidence mechanisms can provide rapid modulation of
knowledge retrieval which would be important for automaticity (Shell, Soh, & Chiriacescu, 2015).
In our paper at the ICCI*CC*16 conference, we showed that chunking vastly increases efficiency
of learning and successful task solution within the known limits of human working memory (Shell,
Soh, & Chiriacescu, 2016).
These chunking and automaticity mechanismw are what allow humans with very limited working
memory and very slow processing speeds to perform behaviors and attain levels of knowledge and
expertise that equal and in many cases exceed what computer systems with vastly more power can
achieve. Although reinforcement based machine learning (Mohri, Rostamizadeh, & Talwalkar,
2012) utilizes repetition during training and problem solving and the results are used to select which
solutions produce the best, or at least acceptable, outcomes, these methods do not do the extensive
data compression of chunks or reductions of processing steps characteristic of automaticity. Deep
learning and other neural-net variants used in cognitive informatics, come the closest to representing
true chunking (LeCun, Bengio, & Hinton, 2015). These contain hierarchical layers that can and mirror
the structural chunking seen within the hierarchical structure of the brain. But, as yet, they do not
produce the same type of data compression to statistical regularities or chunk within layers. Computers
still rely primarily on brute force via speed and massively parallel working memory.
The ability of the human brain to compress vast amounts of information into a relatively small
neural assembly via chunking and automaticity is a true marvel. The remarkable storage capacity
of the human brain of about 100 billion neurons, with each having about 1,000 connections to other
neurons, has been widely noted. But this storage capacity is even greater when one realizes that neural
connections represent not a single piece of information, but rather chunks. Given the processing
power advantages of computers, it may not be important for future cognitive computers that they
be able to chunk or achieve automaticity of behavior. If Moore’s Law persists, they may not need
to as brute force may be sufficient. On the other hand, there are limits. Watson may have bested
its human competitors at Jeopardy; but, its human competitors did something that is perhaps even
more impressive-they walked to their podiums. They could smoothly, and easily walk because of
automaticity. No robotic or computer-based system has yet been able to do this, so maybe there are
some advantages to be gained from incorporating chunking and automaticity into future cognitive
computing systems. (This section is contributed by Prof. Duane F. Shell.)
11
11. CONCLUSION
This paper has summarized a set of position statements presented in the plenary panel (Part I) of IEEE
ICCI*CC’16 on Cognitive Informatics: Perspectives on Cognitive Computing, Big Data Cognition,
and Machine Learning contributed by invited panelists who are part of the world’s renowned
scholars in the field of cognitive informatics and cognitive computing. It has been elaborated that
the theoretical foundations underpinning cognitive computing are cognitive informatics, the science
of cognitive information and knowledge processing, as well as denotational mathematics. A wide
range of theoretical breakthroughs and engineering applications have been reported such as cognitive
informatics theories, cognitive computing methodologies, cognitive robots, computing with words,
deep learning machines, cognitive learning engines, cognitive systems, cognitive knowledge bases,
cognitive engineering, and cognitive self-driving cars.
12
REFERENCES
Anderson, J. R. (1983). The Architecture of Cognition. Cambridge, MA: Harvard Univ. Press.
Azam, S., & Gavrilova, M. (2016). Soft Biometric: Give Me Your Favorite Images and I Will Tell Your Gender.
Proceedings of ICCI*CC ’16, Stanford, USA.
G. Baciu, Y. Wang, Y. Yao, K. Chan, W. Kinsner, & L. A. Zadeh (Eds.). (2009, June). Proceedings of the 8th
IEEE International Conference on Cognitive Informatics (ICCI ‘09), Hong Kong. Los Alamitos, CA, USA:
IEEE Computer Society Press.
Bender, E. A. (1996). Mathematical Methods in Artificial Intelligence. Los Alamitos, CA: IEEE CS Press.
Bhavsar, V. C., Boley, H., & Yang, L. (2004). A Weighted-Tree Similarity Algorithm for Multiagent Systems in
e-Business Environments. Computational Intelligence, 20(4), 584–602. doi:10.1111/j.0824-7935.2004.00255.x
Boley, H., Bhavsar, V. C., Hirtle, D., Singh, A., Sun, Z., & Yang, L. (2005). A Match-Making System for Learners
and Learning Objects. International Journal of Interactive Technology and Smart Education, 2(3), 171–178.
doi:10.1108/17415650580000042
Boucenna, S., Gaussier, P., & Hafemeister, L. (2014). Development of first social referencing skills: Emotional
interaction as a way to regulate robot behavior. IEEE Trans. Autonom. Mental Develop, 6(1), 42–55. doi:10.1109/
TAMD.2013.2284065
Byrd, R. J., Calzolari, N., Chodorow, M. S., Klavans, J. L., Neff, M. S., & Rizk, O. (1987). Tools and Methods
for Computational Linguistics. Computational Linguistics, 13(3/4), 219–240.
Chan, C., & Kinsner, W., Wang, Y., & Miller, D. (Eds.), (2004, July). Proceedings of the 3rd IEEE International
Conference on Cognitive Informatics (ICCI’04), Victoria, Canada. Los Alamitos, CA, USA: IEEE Computer
Society Press.
Chan, V., & Chan, C. W. (2016). Development and application of an algorithm for extracting multiple linear
regression equations from artificial neural networks for nonlinear regression problems.Proceedings of 15th
International IEEE Conference of Cognitive Informatics and Cognitive Computing (ICCI*CC ’16), Palo Alto
USA (pp. 479-488).
Chiriacescu, V., Soh, L. K., & Shell, D. F. (2013). Understanding human learning using a multi-agent simulation
of the Unified Learning Model. International Journal of Cognitive Informatics and Natural Intelligence, 7(4),
1–25. doi:10.4018/ijcini.2013100101
Dietterich, T. G., Domingos, P., Getoor, L., Muggleton, S., & Tadepalli, P. (2008). Structured Machine Learning:
The Next Ten Years. Machine Learning, 73(1), 3–23. doi:10.1007/s10994-008-5079-1
Fiorini, R.A. (2015). Computerized tomography noise reduction by CICT optimized exponential cyclic sequences
(OECS) co-domain. Fundamenta Informaticae, 141(2-3), 115–134. doi:10.3233/FI-2015-1267
Fiorini, R.A. (2015, November 15-30). A Cybernetics Update for Competitive Deep Learning System. Proc.
2nd International Electronic Conference on Entropy and Its Applications. Retrieved from http://sciforum.net/
conference/ecea-2/paper/3277
Fiorini, R. A. (2015). More effective biomedical experimentation data by CICT advanced ontological uncertainty
management techniques. Int. J. of Biol. and Biomed. Eng., 9, 29–41.
Fiorini, R.A. (2016, August 22-23). Deep Learning and Deep Thinking New Application Framework by CICT
IEEE ICCI’CC’16, Proc. 15th IEEE International Conference on Cognitive Informatics & Cognitive Computing,
Stanford, CA, USA.
Fiorini, R.A. (2014). How random is your tomographic noise? A number theoretic transform (NTT) approach.
Fundamenta Informaticae, 135(1-2), 135–170.
Ge, N., Lu, J., Wang, Y., Howard, N., Chen, P., Tao, X., . . . Zadeh, L. A. (Eds.). (2015, July). Proceedings of
the 14th IEEE International Conference on Cognitive Informatics and Cognitive Computing (ICCI*CC ‘15),
London, UK. Los Alamitos, CA, USA: IEEE Computer Society Press.
13
Gers, F. A., & Schmidhuber, J. (2001). LSTM Recurrent Networks Learn Simple Context Free and Context
Sensitive Languages. IEEE Transactions on Neural Networks, 12(6), 1333–1340. doi:10.1109/72.963769
PMID:18249962
Gharbieh, A., Bhavsar, V. C., & Cook, P. (2016, August 7-12). A Word Embedding Approach to Identifying
Verb-Noun Idiomatic Combinations. Proc. of the 12th Workshop on Multiword Expressions (MWE 2016), Berlin,
Germany (pp. 112-118). doi:10.18653/v1/W16-1817
Gilat, A., & Subramaniam, V. (2010). Numerical Methods for Engineers and Scientists (2nf ed.). John Wiley.
Gobet, F., & Lane, P. C. R. (2012). Chunking mechanisms and learning. In M. Seel (Ed.), Encyclopedia of the
Sciences of Learning (pp. 541–544). New York, NY: Springer.
Goldberg, D. (1989). Genetic Algorithms in Search, Optimization and Machine Learning. Reading, MA:
Addison-Wesley.
Hebb, D. O. (1949). The organization of behavior: A neuropsychological theory. New York: Wiley and Sons.
Hinton, G. E., Dayan, P., Frey, B. J., & Neal, R. (1995). The Wake-sleep Algorithm for Unsupervised Neural
Networks. Science, 268(5214), 1158–1161. doi:10.1126/science.7761831 PMID:7761831
Hsu, D. F., Wang, Y., Rao, A. R., Zhang, D., Kinsner, W., Pedrycz, W., . . . Zadeh, L. A. (Eds.). (2013, July).
Proceedings of the 12th IEEE International Conference on Cognitive Informatics and Cognitive Computing
(ICCI*CC ‘13), New York, USA. Los Alamitos, CA, USA: IEEE Computer Society Press.
IBM. (2016). Press release. Retrieved from https://www-03.ibm.com/press/us/en/pressrelease/49985.wss
Jain, A. K., Ross, A., & Prabhakar, S. (2004). An introduction to biometric recognition. IEEE Transactions on
Circuits Syst. Video Technol., 14(1), 420.
Jain, P., Darbari, H., & Bhavsar, V. C. (2013, June 28-29). Text Visualization as an Aid to Language Learning
Disability. Proc. National Conference on e-Learning and e-Learning Technologies, (ELETECH ‘13), Hyderabad,
India, Center for Development of Advanced Computing (C-DAC).
Jain, P., Darbari, H., & Bhavsar, V. C. (2014, April 7-9). Vishit: A Visualizer for Hindi Text. Proc. of IEEE
Fourth International Conference on ‘Communication Systems and Network Technologies (CSNT), Bhopal, MP,
India (pp. 886-890). IEEE. doi:10.1109/CSNT.2014.183
Joshi, M., Bhavsar, V. C., & Boley, H. (2010a), Knowledge Representation in Matchmaking Applications. In
P.S. Sajja & R. Akerkar (Eds.), ‘Advanced Knowledge-Based Systems: Models, Applications and Research (Ch.
3, pp. 29-49). India: TMRF.
Joshi, M., Bhavsar, V. C., & Boley, H. (2010b, August 2-4). Matchmaking in P2P e-Marketplaces: Soft Constraints
and Compromise Matching. Proc. of the 12th International Conference on e-Commerce (ICEC ‘10), Hawaii,
USA (pp. 148-154).
Kahneman, D. (2011). Thinking, Fast and Slow. New York: Macmillan.
Karasuyama, M., & Takeuchi, I. (2010). Multiple Incremental Decremental Learning of Support Vector Machines.
IEEE Transactions on Neural Networks, 21(7), 1048–1059. doi:10.1109/TNN.2010.2048039 PMID:20550990
Kiani, M., Bhavsar, V. C., & Boley, H. (2013). Combined structure-weight graph similarity and its application in
e-health, in CEUR Workshop. Proceedings of the 4th Canadian Semantic Web Symposium, Concordia University,
Montreal, Canada (pp. 12–18).
Kiani, M., Bhavsar, V. C., & Boley, H. (2014a, June 16-18). Structure Similarity of Attributed Generalized
Trees. Proc. of the Eighth IEEE International Conference on Semantic Computing (ICSC), Newport Beach,
CA, USA (pp. 100-107).
Kiani, M., Bhavsar, V. C., & Boley, H. (2014b, August 18-20). A fuzzy-structure similarity algorithm for
attributed generalized trees. Proc. of the 2014 IEEE 13th International Conference on ‘Cognitive Informatics and
Cognitive Computing (ICCI*CC ‘14), London South Bank University, London, UK (pp. 203-210). doi:10.1109/
ICCI-CC.2014.6921461
14
Kiani, M., Bhavsar, V. C., & Boley, H. (2015, October 12-14). Similarity of Attributed Generalized Tree
Structures: A Comparative Study. Proc. of the 8th International Similarity Search and Applications (SISAP
‘15) conference, Glasgow, UK, LNCS (Vol. 9371, pp. 150-161). Switzerland: Springer International Pub. , .
doi:10.1007/978-3-319-25087-8_14
Kiani, M., Bhavsar, V. C., & Boley, H. (2016, August 22-23). Semantic Computing of Simplicity in Attributed
Generalized Trees. Proc. of 2016 IEEE International Conference on ‘Cognitive Informatics and Cognitive
Computing (ICCI*CC’16), Y. Wang, N. Howard, B. Widrow, K. Plataniotis, L.A. Zadeh (Eds.), Stanford
University, 2016, pp. 244-251.
Kober, J., & Peters, J. (2010). Imitation and Reinforcement Learning. IEEE Robotics and Automation, 17(2),
55–62. doi:10.1109/MRA.2010.936952
LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature, 521(7553), 436–444. doi:10.1038/
nature14539 PMID:26017442
Lu, Y., Yang, L., Kumar, N., & Bhavsar, V. C. (2016). (in preparation). Tree Structured Data Processing for
Weighted Tree Similarity Algorithm on GPUs.
Maulik, U., Bandyopadhyay, S., & Saha, I. (2010). Integrating Clustering and Supervised Learning for Categorical
Data Analysis[Part A]. IEEE Transactions on Systems, Man, and Cybernetics, 40(4), 664–675. doi:10.1109/
TSMCA.2010.2041225
Mehryar, M., Rostamizadeh, A., & Talwalkar, A. (2012). Foundations of Machine Learning. MA: MIT Press.
Mohri, M., Rostamizadeh, A., & Talwalkar, A. (2012). Foundations of machine learning. Cambridge, MA:
The MIT Press.
Moors, A. (2016). Automaticity: Componential, Causal, and Mechanistic Explanations. Annual Review of
Psychology, 67(1), 263–287. doi:10.1146/annurev-psych-122414-033550 PMID:26331343
Morimoto, J. (2010). Robot Learning in Practice. IEEE Robotics & Automation Magazine, 17(2), 17–18.
doi:10.1109/MRA.2010.937374
S. Patel, D. Patel, & Y. Wang (Eds.). (2003, July). Proceedings of the 2nd IEEE International Conference on
Cognitive Informatics (ICCI’03). London: IEEE Computer Society Press.
Patel, S., Wang, Y., Patel, D., Kinsner, W., & Zadeh, L. A. (Eds.). (2014, August). Proceedings of the 13th IEEE
International Conference on Cognitive Informatics and Cognitive Computing(ICCI*CC ‘14), London, UK. Los
Alamitos, CA, USA: IEEE Computer Society Press.
Paul, P.P., Sultana, M., Matei, S.A., & Gavrilova, M. (2015). Editing Behavior to Recognize Authors of
Crowdsourced Content. In Systems, Man, and Cybernetics (pp. 1676-1681).
Paul, P. P., Gavrilova, M., & Klimenko, S. (2014). Situation awareness of cancelable biometric system. The
Visual Computer, 30(9), 1059–1067. doi:10.1007/s00371-013-0907-0
Poole, D., Mackworth, A., & Goebel, R. (1997). Computational Intelligence: A Logical Approach. Oxford, UK:
Oxford University Press.
Reisberg, D. (2001). Cognition: Exploring the Science of the Mind. Norton & Company, Inc.
Reuters Health News. (2016, August 31). Novartis to disband cell & gene therapy unit, 120 jobs to go. Health
News. Retrieved from http://www.reuters.com/article/us-novartis-therapy-idUSKCN1162NX
Schmidhuber, J. (2015). Deep Learning in Neural Networks: An Overview. Neural Networks, 61, 85–117.
doi:10.1016/j.neunet.2014.09.003 PMID:25462637
Segalin, C., Perina, A., & Cristani, M. (2014). Personal aesthetics for soft biometrics: A generative multi-resolution
approach. Proceedings of the16th International Conference on Multimodal Interaction (ICMI ‘14) (pp. 180-187).
Shannon, C. E. (1948). A Mathematical Theory of Information. The Bell System Technical Journal, 27(3),
379–423. doi:10.1002/j.1538-7305.1948.tb01338.x
Shaw, C. C. (2010). Dimensions in Medical Imaging: The More the Better? Proceedings of the IEEE, 98(1),
2–5. doi:10.1109/JPROC.2009.2035158
15
Shell, D. F., Brooks, D. W., Trainin, G., Wilson, K., Kauffman, D. F., & Herr, L. (2010). The Unified Learning
Model: How motivational, cognitive, and neurobiological sciences inform best teaching practices. Netherlands:
Springer. doi:10.1007/978-90-481-3215-7
Shell, D. F., Soh, L.-K., & Chiriacescu, V. (2015). Modeling self-efficacy as a dynamic cognitive process with
the Computational-Unified Learning Model (C-ULM): Implications for cognitive informatics and cognitive
computing. International Journal of Cognitive Informatics and Natural Intelligence, 9(3), 1–24. doi:10.4018/
IJCINI.2015070101
Shell, D. F., Soh, L.-K., & Chiriacescu, V. (2016). Modeling chunking effects on learning and performance using
the Computational-Unified Learning Model (C-ULM): A multiagent cognitive process model. Proceedings of
the 15th IEEE International Conference on Cognitive Informatics & Cognitive Computing (ICCI*CC ‘16) (pp.
77-85). Stanford University, Palo Alto: IEEE Computer Society.
Sheu, P. (2007). Semantic Computing. Intl Journal of Semantic Computing, 1(1), 1–9. doi:10.1142/
S1793351X07000068
Sourin, A., Earnshaw, R., Gavrilova, M., & Sourina, O. (2016). Problems of Human-Computer Interaction in
Cyberworlds. Transactions on Computational Science, XXVIII, 1–22.
Sugawara, K., Wang, Y., Hattori, F., Fujita, S., Nishida, T., Kinsner, W., & Zadeh, L. A. (Eds.). (2012, August).
(ICCI*CC’12), Kyoto, Japan. Los Alamitos, CA, USA: IEEE Computer Society Press.
Sultana, M., Paul, P. P., & Gavrilova, M. (2014a). A concept of social behavioral biometrics: motivation, current
developments, and future trends. Proceedings of theInt. Conf. on Cyberworlds (pp. 271-278). doi:10.1109/
CW.2014.44
Sultana, M., Paul, P. P., & Gavrilova, M. (2014b). Mining Social Behavioral Biometrics in Twitter. Proceedings
of theInt. Conf. on Cyberworlds (pp. 293-299).
Sultana,M., Paul, P.P., and Gavrilova, M. (2015). Social behavioral biometrics: An emerging trend. International
Journal of Pattern Recognition and Artificial Intelligence, 29(8).
Sultana, M., Paul, P. P., & Gavrilova, M. (2016). Identifying users from online interactions in Twitter. In
Transactions on Computational Science XXVI (pp. 111–124). Springer Berlin Heidelberg. doi:10.1007/978-3-
662-49247-5_7
Sun, F., Wang, Y., Lu, J., Zhang, B., & Kinsner, W. (Eds.), (2010), Proceedings of the 9th IEEE International
Conference on Cognitive Informatics(ICCI’10), Tsinghua University, Beijing. Los Alamitos, CA, USA: IEEE
Computer Society Press.
Tomar, A. A., Bodhankar, J., Kurariya, P., Anarase, P., Jain, P., Lele, A., & Bhavsar, V. C. et al. (2013a, February
21-22). Parallel Implementation of Machine Translation using MPJ Express. Proc. National Conference on
Parallel Computing Technologies (PARCOMPTECH), Bengaluru, India, Center for Development of Advanced
Computing (C-DAC) (pp. 223-233). doi:10.1109/ParCompTech.2013.6621391
Tomar, A. A., Bodhankar, J., Kurariya, P., Anarase, P., Jain, P., Lele, A., & Bhavsar, V. C. et al. (2013b, February
21-23). High Performance Natural Language Processing Services in the GARUDA Grid. Proc. National
Conference on Parallel Computing Technologies (PARCOMPTECH), Bengaluru, India, Center for Development
of Advanced Computing (C-DAC) (pp. 249-261). doi:10.1109/ParCompTech.2013.6621407
Tomar, A. A., Bodhankar, J., Kurariya, P., Jain, P., Lele, A., Darbari, H., & Bhavsar, V. C. (2014, August 27-29).
Translation Memory for a Machine Translation System using the Hadoop Framework. Proceedings of the4th
IEEE International Conference on Advances in Computing and Communications (ICACC ‘14), Rajagiri School
of Engineering and Technology, Cochin, Kerala, India (pp. 203-207). doi:10.1109/ICACC.2014.56
Tülay, A., Levin-Schwartz, Y., & Calhoun, V. D. (2015). Multimodal Data Fusion Using Source Separation:
Application to Medical Imaging. Proceedings of the IEEE, 103(9), 1494–1506. doi:10.1109/JPROC.2015.2461601
Vargas, J. G., & Torr, D. G. (2015). The Cartan-Einstein-Kähler Unification Theory Based on Teleparallelism.
Retrieved from http://www.physical-unification.com/
16
Wang, Y. (2002a). Keynote: On Cognitive Informatics, Proc. 1st IEEE International Conference on Cognitive
Informatics (ICCI’02), Calgary, Canada, IEEE CS Press, August, 34-42. doi:10.1109/COGINF.2002.1039280
Wang, Y. (2002b). The Real-Time Process Algebra (RTPA). Annals of Software Engineering, 14(1/4), 235–274.
doi:10.1023/A:1020561826073
Wang, Y. (2003). On Cognitive Informatics. Brain and Mind: A Transdisciplinary Journal of Neuroscience and
Neurophilosophy, 4(2), 151-167.
Wang, Y. (2006, July). Keynote: Cognitive Informatics - Towards the Future Generation Computers that Think
and Feel. Proc. 5th IEEE International Conference on Cognitive Informatics(ICCI’06), Beijing, China (pp. 3-7).
IEEE CS Press. doi:10.1109/COGINF.2006.365666
Wang, Y. (2007a). The Theoretical Framework of Cognitive Informatics. International Journal of Cognitive
Informatics and Natural Intelligence, 1(1), 1–27. doi:10.4018/jcini.2007010101
Wang, Y. (2007b, July). Software Engineering Foundations: A Software Science Perspective, CRC Series in
Software Engineering (Vol. 2). NY, USA: Auerbach Publications.
Wang, Y. (2007c). The OAR Model of Neural Informatics for Internal Knowledge Representation in the
Brain. International Journal of Cognitive Informatics and Natural Intelligence, 1(3), 64–75. doi:10.4018/
jcini.2007070105
Wang, Y. (2008), On Contemporary Denotational Mathematics for Computational Intelligence. In Transactions
of Computational Science. Springer.
Wang, Y. (2009a). On Abstract Intelligence: Toward a Unified Theory of Natural, Artificial, Machinable, and
Computational Intelligence. International Journal of Software Science and Computational Intelligence, 1(1),
1–18. doi:10.4018/jssci.2009010101
Wang, Y. (2009b). On Cognitive Computing. International Journal of Software Science and Computational
Intelligence, 1(3), 1–15. doi:10.4018/jssci.2009070101
Wang, Y. (2009c). Paradigms of Denotational Mathematics for Cognitive Informatics and Cognitive Computing.
Fundamenta Informaticae, 90(3), 282–303.
Wang, Y. (2010a). Cognitive Robots: A Reference Model towards Intelligent Authentication. IEEE Robotics
and Automation, 17(4), 54–62. doi:10.1109/MRA.2010.938842
Wang, Y. (2010b). On Formal and Cognitive Semantics for Semantic Computing. International Journal of
Semantic Computing, 4(2), 203–237. doi:10.1142/S1793351X10000833
Wang, Y. (2011a). Towards the Synergy of Cognitive Informatics, Neural Informatics, Brain Informatics, and
Cognitive Computing. International Journal of Cognitive Informatics and Natural Intelligence, 5(1), 75–93.
doi:10.4018/jcini.2011010105
Wang, Y. (2011b). Inference Algebra (IA): A Denotational Mathematics for Cognitive Computing and Machine
Reasoning (I). International Journal of Cognitive Informatics and Natural Intelligence, 5(4), 62–83. doi:10.4018/
jcini.2011100105
Wang, Y. (2012a). In Search of Denotational Mathematics : Novel Mathematical Means for Contemporary
Intelligence, Brain, and Knowledge Sciences. Journal of Advanced Mathematics and Applications, 1(1), 1–31.
doi:10.1166/jama.2012.1001
Wang, Y. (2012b). On the Denotational Mathematics Foundations for the Next Generation of Computers:
Cognitive Computers for Knowledge Processing. Journal of Advanced Mathematics and Applications, 1(1),
101–112. doi:10.1166/jama.2012.1009
Wang, Y. (2012c). On Abstract Intelligence and Brain Informatics: Mapping Cognitive Functions of the Brain
onto its Neural Structures. International Journal of Cognitive Informatics and Natural Intelligence, 6(4), 54–80.
doi:10.4018/jcini.2012100103
Wang, Y. (2012d). Contemporary Mathematics as a Metamethodology of Science, Engineering, Society, and
Humanity. Journal of Advanced Mathematics and Applications, 1(1), 1–3. doi:10.1166/jama.2012.1001
17
Wang, Y. (2012e, June 18-21). Keynote: Towards the Next Generation of Cognitive Computers: Knowledge
vs. Data Computers. Proceedings of the12th International Conference on Computational Science and
Applications(ICCSA’12), Salvador, Brazil (pp. 3). Springer.
Wang, Y. (2013a). Neuroinformatics Models of Human Memory: Mapping the Cognitive Functions of Memory
onto Neurophysiological Structures of the Brain. International Journal of Cognitive Informatics and Natural
Intelligence, 7(1), 98–122. doi:10.4018/jcini.2013010105
Wang, Y. (2013b). On Semantic Algebra: A Denotational Mathematics for Cognitive Linguistics, Machine
Learning, and Cognitive Computing. Journal of Advanced Mathematics and Applications, 2(2), 145–161.
doi:10.1166/jama.2013.1039
Wang, Y. (2014a). On a Novel Cognitive Knowledge Base (CKB) for Cognitive Robots and Machine Learning.
International Journal of Software Science and Computational Intelligence, 6(2), 42–64.
Wang, Y. (2014b). Software Science: On General Mathematical Models and Formal Properties of Software.
Journal of Advanced Mathematics and Applications, 3(2), 130–147. doi:10.1166/jama.2014.1060
Wang, Y. (2015a, July). Keynote: Cognitive Robotics and Mathematical Engineering. Proceedings of the 15th
IEEE Int’l Conf. on Cognitive Informatics & Cognitive Computing (ICCI*CC ‘15), Tsinghua Univ. (pp. 3).
IEEE CS Press.
Wang, Y. (2015b). Cognitive Learning Methodologies for Brain-Inspired Cognitive Robotics. International
Journal of Cognitive Informatics and Natural Intelligence, 9(2), 37–54. doi:10.4018/IJCINI.2015040103
Wang, Y. (2015c). On Mathematical Theories and Cognitive Foundations of Information. International Journal
of Cognitive Informatics and Natural Intelligence, 9(3), 41–63. doi:10.4018/IJCINI.2015070103
Wang, Y. (2015d). Concept Algebra: A Denotational Mathematics for Formal Knowledge Representation
and Cognitive Robot Learning. Journal of Advanced Mathematics & Applications, 4(1), 1–26. doi:10.1166/
jama.2015.1066
Wang, Y. (2015e). Fuzzy Probability Algebra (FPA): A Theory of Fuzzy Probability for Fuzzy Inference and
Computational Intelligence. Journal of Advanced Mathematics and Applications, 4(1), 38–55. doi:10.1016/j.
jmaa.2014.06.082
Wang, Y. (2015f). Towards the Abstract System Theory of System Science for Cognitive and Intelligent Systems.
Springer Journal of Complex and Intelligent Systems, 1(1), 1–22. doi:10.1007/s40747-015-0001-5
Wang, Y. (2016a), Keynote: Deep Reasoning and Thinking beyond Deep Learning by Cognitive Robots and
Brain-Inspired Systems. Proceedings of the 15th IEEE International Conference on Cognitive Informatics and
Cognitive Computing (ICCI*CC ‘16), Stanford University, Stanford, CA (pp. 3). IEEE CS Press.
Wang, Y. (2016b). On Cognitive Foundations and Mathematical Theories of Knowledge Science. International
Journal of Cognitive Informatics and Natural Intelligence, 10(2), 1–24. doi:10.4018/IJCINI.2016040101
Wang, Y. (2016c, November 28-30). Keynote: Brain-Inspired Deep Machine Learning and Cognitive Learning
Systems. Proceedings of the8th International Conference on Brain Inspired Cognitive Systems (BICS’16),
Beijing (pp. 2).
Wang, Y. (2016d). (in press). Big Data Algebra: A Denotational Mathematics for Big Data Science and
Engineering. Journal of Advanced Mathematics and Applications, 5(1).
Wang, Y. (2016e, November 21-22). Keynote: Cognitive Neuroscience and the Spike Frequency Modulation
(SFM) Theory for Neural Signaling Systems. Proceedings of the9th Global Neuroscience Conference (GNSC’16),
Melbourne, Australia, pp 3..
Wang, Y. (2017). Keynote: Latest Advances in Neuroinformatics and the Neural Circuit Theory. Proceedings
of the 8th World Congress of NeuroTalk (NeuroTalk’17), Barcelona, Spain.
Wang, Y., & Berwick, R. C. (2012). Towards a Formal Framework of Cognitive Linguistics. Journal of Advanced
Mathematics and Applications, 1(2), 250–263. doi:10.1166/jama.2012.1019
18
Wang, Y., & Berwick, R. C. (2013). Formal Relational Rules of English Syntax for Cognitive Linguistics,
Machine Learning, and Cognitive Computing. Journal of Advanced Mathematics and Applications, 2(2),
182–195. doi:10.1166/jama.2013.1042
Wang Y., Berwick, R.C., Haykin, S., Pedrycz, W., Kinsner, W., Baciu, G., Zhang, D., Bhavsar, V.C., & Gavrilova,
M. (2011, October-December). Cognitive Informatics and Cognitive Computing in Year 10 and Beyond.
International Journal of Cognitive Informatics and Natural Intelligence, 5(4), 1-21.
Wang, Y., Berwick, R.C., Haykin, S., Pedrycz, W., Kinsner, W., Baciu, G., Zhang, D., & Gavrilova, M. (2012).
Cognitive informatics and cognitive computing in year 10 and beyond. In Cognitive Informatics for Revealing
Human Cognition: Knowledge Manipulations.
Wang, Y., Celikyilmaz, A., Kinsner, W., Pedrycz, W., Leung, H., & Zadeh, L. A. (Eds.). (2011a, August).
(ICCI*CC ‘11), Banff, Canada. Los Alamitos, CA, USA: IEEE Computer Society Press.
Wang, Y., & Fariello, G. (2012). On Neuroinformatics: Mathematical Models of Neuroscience and Neurocomputing.
Journal of Advanced Mathematics and Applications, 1(2), 206–217. doi:10.1166/jama.2012.1015
Wang, Y., Howard, N., Plataniotis, K., Widrow, B., & Zadeh, L. A. (Eds.). (2016b). Proceedings of the 15th
IEEE International Conference on Cognitive Informatics and Cognitive Computing (ICCI*CC’16), Stanford,
CA., IEEE Computer Society Press, Los Alamitos, CA., Aug.
Y. Wang, R. H. Johnston, & M. R. Smith (Eds.). (2002). Proceedings of the 1st IEEE International Conference
on Cognitive Informatics (ICCI’02).IEEE Computer Society Press.
Wang, Y., & Kinsner, W. (2006). Recent Advances in Cognitive Informatics. IEEE Transactions on Systems,
Man and Cybernetics. Part C, Applications and Reviews, 36(2), 121–123. doi:10.1109/TSMCC.2006.871120
Wang, Y., Kinsner, W., Anderson, J. A., Sheu, P., Tsai, J., Pedrycz, W., & Zadeh, L. A. et al. (2009a). A Doctrine
of Cognitive Informatics. Fundamenta Informaticae, 90(3), 203–228.
Wang, Y., Kinsner, W., & Zhang, D. (2009b). Contemporary Cybernetics and its Faces of Cognitive Informatics
and Computational Intelligence[Part B]. IEEE Transactions on Systems, Man, and Cybernetics, 39(4), 1–11.
Wang, Y., Rolls, E. T., Howard, N., Raskin, V., Kinsner, W., & Murtagh, F., Bhavsar, V.C., Patel, S., Patel, DShell,
D. F. et al. (2015). Cognitive Informatics: From Information Revolution to Intelligence Revolution. International
Journal of Software Science and Computational Intelligence, 7(2), 52–71. doi:10.4018/IJSSCI.2015040103
Wang, Y., & Ruhe, G. (2007). The Cognitive Process of Decision Making. International Journal of Cognitive
Informatics and Natural Intelligence, 1(2), 73–85. doi:10.4018/jcini.2007040105
Wang, Y., Tian, Y., & Hu, K. (2011b). Semantic Manipulations and Formal Ontology for Machine Learning
Based on Concept Algebra. International Journal of Cognitive Informatics and Natural Intelligence, 5(3), 1–29.
doi:10.4018/IJCINI.2011070101
Wang, Y., Valipour, M., & Zatarain, O. A. (2016a). (in press). Quantitative Semantic Analysis and Comprehension
by Cognitive Machine Learning. International Journal of Cognitive Informatics and Natural Intelligence, 10(3).
Wang, Y., & Wang, Y. (2006). Cognitive Informatics Models of the Brain. IEEE Transactions on Systems, Man
and Cybernetics. Part C, Applications and Reviews, 36(2), 203–207. doi:10.1109/TSMCC.2006.871151
Wang, Y., Wang, Y., Patel, S., & Patel, D. (2006). A Layered Reference Model of the Brain (LRMB)[Part C].
IEEE Transactions on Systems, Man, and Cybernetics, 36(2), 124–133. doi:10.1109/TSMCC.2006.871126
PMID:16602600
Wang, Y., Zadeh, L. A., Widrow, B., Howard, N., Wood, S., Patel, S., & Zhang, D. et al. (2016c). (in press).
Abstract Intelligence: Embodying and Enabling Cognitive Systems by Mathematical Engineering. International
Journal of Software Science and Computational Intelligence, 8(4).
Wang, Y., Zhang, D., & Kinsner, W. (Eds.), (2010). Advances in Cognitive Informatics and Cognitive Computing.
Springer. doi:10.1007/978-3-642-16083-7
Y. Wang, D. Zhang, J.-C. Latombe, & W. Kinsner (Eds.). (2008). Proceedings of the 7th IEEE International
Conference on Cognitive Informatics (ICCI ‘08). Los Alamitos, CA: IEEE Computer Society Press.
19
Widrow, B. (2016). Hebbian Learning and the LMS Algorithm [Keynote speech]. Proceedings of the IEEE
15th International Conference on Cognitive Informatics and Cognitive Computing (ICCI*CC’16), Stanford
University, CA, USA (pp. 2). IEEE Press.
Widrow, B., & Aragon, J. C. (2013). Cognitive Memory. Neural Networks, 41, 3–14. doi:10.1016/j.
neunet.2013.01.016 PMID:23453302
Widrow, B., Greenblatt, A., Kim, Y., & Park, D. (2013). The no-prop algorithm: A new learning algorithm for
multilayer neural networks. Neural Networks, 37, 182–188. doi:10.1016/j.neunet.2012.09.020 PMID:23140797
Widrow, B., Kim, Y., & Park, D. (2015). The Hebbian-LMS Learning Algorithm. IEEE Computational
Intelligence Magazine, 10(11), 37–53. doi:10.1109/MCI.2015.2471216
Widrow, B., Kim, Y., & Park, D. (2015). The Hebbian-LMS Learning Algorithm. IEEE Computational
Intelligence Magazine, 10(Nov), 37–53. doi:10.1109/MCI.2015.2471216
Wilson, R. A., & Keil, F. C. (2001). The MIT Encyclopedia of the Cognitive Sciences. MIT Press.
Yampolskiy, R., & Gavrilova, M. (2012). Artimetrics: Biometrics for artificial entities. IEEE Robotics &
Automation Magazine, 19(4), 48–58. doi:10.1109/MRA.2012.2201574
Yang, L., Ball, M., Boley, H., & Bhavsar, V. C. (2005). Weighted Partonomy-Taxonomy Trees with Local
Similarity Measures for Semantic Buyer-Seller Match-Making. Journal of Business and Technology, 1(1), 42–52.
Yang, L., Sarker, B. K., Bhavsar, V. C., & Boley, H. (2008). Range Similarity and Satisfaction Measures
for Buyers and Sellers in e-Marketplaces. Journal of Intelligent Systems, 17(1-3), 247–266. doi:10.1515/
JISYS.2008.17.1-3.247
Y. Yao, Z. Shi, Y. Wang, & W. Kinsner (Eds.). (2006). Proceedings of the 5th IEEE International Conference
on Cognitive Informatics (ICCI’06).Los Alamitos, CA: IEEE Computer Society Press.
Zadeh, L. A. (1965). Fuzzy Sets. Information and Control, 8(3), 338–353. doi:10.1016/S0019-9958(65)90241-X
Zadeh, L. A. (1975). Fuzzy Logic and Approximate Reasoning. Syntheses, 30(3-4), 407–428. doi:10.1007/
BF00485052
Zadeh, L. A. (1997). Towards a Theory of Fuzzy Information Granulation and its Centrality in Human Reasoning
and Fuzzy Logic. Fuzzy Sets and Systems, 19(2), 111–127. doi:10.1016/S0165-0114(97)00077-8
Zadeh, L. A. (1999). From Computing with Numbers to Computing with Words - From Manipulation of
Measurements to Manipulation of Perception. IEEE Transactions on Circuits and Systems, 45(1), 105–119.
doi:10.1109/81.739259
Zadeh, L. A. (2004). Precisiated Natural Language (PNL). AI Magazine, 25(3), 74–91.
Zadeh, L. A. (2016). A Key Issue of Semantics of Information [Keynote speech]. Proceedings of the IEEE
15th International Conference on Cognitive Informatics and Cognitive Computing (ICCI*CC’16), Stanford
University, CA, USA (pp. 1). IEEE Press.
D. Zhang, Y. Wang, & W. Kinsner (Eds.). (2007). Proceedings of the 6th IEEE International Conference on
Cognitive Informatics (ICCI’07).Los Alamitos, CA: IEEE Computer Society Press.
Zhou, Q., Wu, Y., Chan, C. W., Tontiwachwuthikul, P., Idem, R., & Gelowitz, D. (2012). Application of data
modeling and analysis techniques to the carbon dioxide capture process system. Carbon Management Journal,
3(1), 81–94. doi:10.4155/cmt.11.77
20
View publication stats

Cognitive Intelligence: Deep Learning, Thinking, and Reasoning by Brain-Inspired Systems

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Cognitive Intelligence: Deep Learning, Thinking, and Reasoning by Brain-Inspired Systems

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Cognitive Intelligence: Deep Learning, Thinking, and Reasoning

Article in International Journal of Cognitive Informatics and Natural Intelligence · November 2016

Yingxu Wang Newton Howard

SEE PROFILE SEE PROFILE

Sally Wood Virendrakumar C. Bhavsar

SEE PROFILE SEE PROFILE

Autonomous Systems (AS) View project

Image Pattern Recognition and Voronoi Diagrams View project

The user has requested enhancement of the downloaded file.

Cognitive Informatics (CI) is a transdisciplinary enquiry of the internal information processing

2. A THEORETIC FRAMEWORK OF KEY COGNITIVE INTELLIGENCE

Learning is a cognitive process of inductive knowledge acquisition, behavior generation, and

 Li (x, P | x ∈ P )  x = P.x // 1) Object identification

3. HEBBIAN LEARNING AND THE LMS ALGORITHM

4. A KEY ISSUE OF SEMANTICS OF INFORMATION

5. COGNITIVE COMPUTING IN MEDICAL IMAGING:

6. BIG DATA COGNITION: GRAPH SIMILARITY AND

7. COGNITIVE COMPUTING IN MODELING THE CARBON

8. THE LOGICAL RELATIONSHIP BETWEEN EXPERIENCE

Cognitive Informatics (CI) (Wang, 2012a) is a transdisciplinary enquiry of computer science,

9. IMPACT OF SOCIAL BIOMETRICS AND COGNITIVE

10. CHUNKING AND AUTOMATICITY: HANDLING BIG

You might also like