This post was originally published by Dr. Jorge Garza-Ulloa at Medium [AI]
Classically, a “physician” is defined as a professional who possesses special knowledge and skills derived from rigorous education, training, and experience, in other words “medical education remains based on information acquisition and application”. Currently, the amount of available medical knowledge now exceeds the organizing capacity of the human mind. In addition, the skills required of practicing physicians and healthcare personnel and by consequences in Biomedical Engineering will increasingly in two areas : “Collaborating with and managing Artificial Intelligence (AI) applications”, and “the need for more sophisticated mathematical understanding”. My goal on this book is facilitate the learning and implementation of new smart systems through the relationship between three different multidiscipline engineering branches: “Biomedical Engineering”, “Cognitive Science” and “Computer science”; through different “Artificial Intelligent models” to analyze “human illness, diseases and disorders”, with special emphasis in “mental processes of the information during cognition” when disorders of “neurologic diseases are present in the human body”, with the purpose to “evaluate their non-motors symptoms will help to find solution for treatments, following-up and by consequence improve their quality of life ”
This book has seven chapters, covering the following topics:
Chapter I Biomedical Engineering and the evolution of Artificial Intelligence
Study of the interactions on different injuries, illness, and diseases with special emphasis in Neurology with Cognitive Science in Biomedical Engineering solutions based in the evolution of Artificial intelligence (AI) through Machine Learning (ML), Deep Learning (DL) and Cognitive Computing (CC). Introduction to the general framework architecture for AI- Cognitive Computing Agent Systems (AI-CCAS) to help in the detection of cognitive human-like abilities with the objective of develop AI methods for medicine and healthcare through the analysis of numeric data, images, speech, and text to help in the detection and diagnostic of illness or determine health conditions, making special emphasis in neurologic diseases.
Chapter II Introduction to Cognitive science, cognitive computing, and human cognitive relation to help in the solution of AI Biomedical engineering problems
Introduction for the analysis body injuries, diseases, and neurological disorders separated basically as: “motors symptoms (related to movement disorders)” and “non-motor symptoms (related to cognition and no related to movement disorders)”. “Human cognitive development stages” and their relation to “neurons and neural pathways”. “Cognition and its integration with multidiscipline sciences”. “Natural Language Processing applications”, “NLP Text to speech”, “NLP Speech to Text”, “Audio Labeler for Machine Learning”, “NLP analysis for: Sentiment, Emotion, Keywords, Entities, Categories, Concept and Semantic Roles” with MATLAB™ and API as a set of functions and procedures allowing the creation of applications that access the features or data of an operating system, application, or other service through IBM Cloud™ services.
Chapter III Artificial Intelligence models applied to Biomedical Engineering
Introduction to apply Artificial Intelligence algorithms to resolve Biomedical Engineering problems through Evolutionary Algorithms using the evolution of the species, trying to emulate the natural evolution as: “Genetic Algorithms”, “Swarm Algorithms”, “traditional search methods”, “optimization of numeric value problems in 2D and 3D”, “visual analysis of Biomedical Engineering datasets of different diseases from different Bioinstruments to analyze relation between their attributes” applying “AI tools”.
Chapter IV Machine learning models applied to Biomedical Engineering
“Machine Learning” is studied as a subset of “AI” following the steps to obtain a “prediction model” based in “pattern recognition” in data using: “clustering”, “classifiers” and “regression” models. Some advices to follow, select, find, and implement the best “ML Models”, accordingly of the type of machine learning problem. Generally, this can be: “Unsupervised Learning”, “Supervised Learning”, “Reinforcement Learning”, “Survival Models”, “Association Rules” and others. Study of different “ML Models Families” applying “IBM Watson™ SPSS Modeler Flow / IBM Watson Machine Learning™ applications” and “MATLAB™ ML solution under the Statistics and Machine Learning Toolbox™”
Chapter V Deep Learning Models Principles applied to Biomedical Engineering
The underlying principle of “Deep Learning” is a compositional nature of “neural network” inspired by the biological elements that forms the “human brain”, as a collection of “nodes” emulating “brain neurons”, and their “neuron synapses” connections as primary elements of a net, that combined form mid-level elements identified as “Artificial Neural Networks (ANN)”, which in turn are combined with different architectures to form more complex networks. In this book “ANN” are organized based on their architectural type, and the way of their different components are connected to one another to define the specific learning goal as different types as: “Feed Forward Neural Network”, “Backpropagation Neural Networks”, “Recurrent Neural Networks”, “Memory Augmented Neural Networks”, “Modular Neural Networks” and “Evolutive Neural Networks”.
Chapter VI Deep Learning Models Evolution Applied to Biomedical Engineering
In this chapter we focus to study “Deep Learning Models Evolution” that combined mid-level elements with different connections “ANN” to form more complex networks family types organized as: “Recurrent Neural Networks”, “Memory Augmented Neural Networks”, “Modular Neural Networks” and “Evolutive Neural Networks”.
The “ANN” studied in this chapter are: “Recurrent Neural Network (RNN) vanilla”, “Long/Short Term Memory (LSTM)”, “Long/Short Term Memory (LSTM)”, “Gated Recurrent Unit (GRU)”, “Recurrent convolutional neural networks (RCNN)”, “Regional-Convolutional Neural Network (R-CNN)”, “Hopfield Network (HN)”, “Boltzmann Machine (BM)”, “Restricted Boltzmann Machine (RBM)”, “Liquid State Machine (LSM)”, “Echo State Network (ESN)”, “Korhonen Network (KN) also knows as Self Organizing Map (SOM)”, “Neural Turning Machine (NTM)”, “Differentiable Neural Computers (DNC)”, “Deep Belief Network (DBN)”, “Capsule Networks (CapsNet)”, “Attention network (AN)”, and others.
Chapter VII Cognitive Learning & Reasoning models applied to Biomedical Engineering
In this chapter we will focus on many “pre-studies and pre-analysis of different Biomedical Engineering problems that need to be develop with specialized research projects applying Cognitive Learning & Reasoning (CL&R ) algorithm, that can be integrated to the Proposed General Architecture framework of a Cognitive Computing Agents System (AI-CCAS)” with special emphasis at “Cognitive Learning and its relationship with neuroscience of reasoning using Cognitive Learning- Reasoning (CL&R) under Cognitive Computing (CC)” . The complexity for the analysis needed is bondless, and only be analyzed through the multidisciplinary sciences, where interactions of science as “Biomedical Engineering”, “Neurology”, “Cognitive Sciences” and “Computer Science” using tools with “exponential technologies” as “Artificial intelligence (AI)” and others through its “continuous exponential evolution” that includes Machine Learning (ML), Deep Learning (DL) and Cognitive Computing (CC). With the main purpose of obtain useful “AI Models” that can help analyze human health problems. Now it is the time and place to apply them and many others in research challenge projects.
The AI cognitive models to be used for the AI-CCAS are studied in this chapter are: “inference engine to extract information needed from knowledge storage AI storage”, “Attention network for NLP applying Long Short-Term Memory (LSTM) model to process information extracted by the AI-CCAS inference engine”, “Cognitive Learning & Reasoning” using “deductive reasoning”, “inductive reasoning”, “abductive reasoning”, “metaphoric reasoning”, “neuro-fuzzy logic reasoning”, “visual-spatial relational reasoning”, “Inferences Fuzzy Systems for fuzzy reasoning”, “Cognitive sentiments analysis”, “reasoning evaluation for neurologic diseases”, and others
This book includes a companion website with datasets with MATLAB® examples and IBM Watson® tutorials.
As explained in this book, Cognitive Science and Neuroscience has evolved over the years, and recently they have started to intersect, and this is very useful for the development of new Biomedical Engineering solutions applying tools based: on Artificial Intelligence-Machine Learning-Deep Learning-Cognitive Computing (AI-ML-DL-CC). Their evolution seems to point in the direction of revolutionaries’ new software technologies based in “new AI learning methods” and “new auto-configurable hardware” as “Neural network computers”, “Bio-molecular computers needed to develop”, “living bio-hybrid systems” and many more in continuous research and development, where the concept of “Cognitive Learning and its relationship with neuroscience of reasoning proposed as Cognitive Learning- Reasoning (CL&R)” will be easier and faster to implement..
Book/ ebook is available worldwide:
This post was originally published by Dr. Jorge Garza-Ulloa at Medium [AI]