Abstract
This chapter brings out the perspective outcomes of combining three terminologies: artificial intelligence, cloud, and internet of things. The relation between artificial intelligence, machine learning, and deep learning is also emphasized. Intelligence, which is the capability to attain and apply knowledge in addition to skills, is analysed in the following sections of the chapter along with its categories that include natural intelligence, artificial intelligence, and hybrid intelligence. Analysis of artificial intelligence-based internet of things system is deliberated on two approaches, namely criterion-based analysis and elemental analysis. Criterion-based analysis covers the parameter-based investigation to highlight the relation between machine learning and deep learning. Elemental analysis involves four main components of artificial intelligence-based internet of things system, such as device, data, algorithm, and computation. Research works done using deep learning and internet of things are also discussed.
TopIntroduction
Artificial intelligence is a part of computer science that deals with empowering skills and knowledge to the inhuman things in the world. It mainly encompasses two noteworthy thrill terms, namely, machine learning and deep learning. A profound scrutiny of three reckoning terms such as Artificial intelligence, machine learning and deep learning are laid out. A streamlined implementation of deep learning algorithm in Raspberry Pi with Beings classification algorithm of accuracy 98% is achieved.
Machine learning is an outlet of artificial intelligence that enables prediction of the actions and enrich the learning capability of a physical system. It depends on formerly educated features from the training data (Xin et al., 2018). Multi layered function in machine learning is achieved with the help of deep learning; which is a novel machine learning technique; which works with neural networks that are similar to the neuron structures of the human brain. The alliance between three terms, namely, Artificial Intelligence, Machine learning and Deep Learning indicates the benefit that deep learning has use of unsupervised or semi-supervised feature learning and hierarchical feature extraction for automatic and resourceful swapping of features (Deng et al., 2014). The relationship and differences between three terms artificial intelligence, machine learning and deep learning is put up in Figure 1.
Figure 1. Relation and Differences among Artificial Intelligence, Machine Learning and Deep Learning
TopIntelligence
Intelligence is the most fascinating factor for a human as well as a machine in the present state. It can be described as the capability to estimate, reason, distinguish relations and analogies, study from practice, store and retrieve information from memory, resolve difficulties, comprehend multifaceted ideas, practice natural language effortlessly, categorize, simplify, and adapt to new situations. This intelligence can be demarcated into three varied types, such as, natural, artificial and hybrid. Intelligence along with its detailed subtypes is shown as taxonomy in Figure 2.
Natural Intelligence
Natural intelligence is generally present in humans and animals. It can be the extraordinary smelling power of a dog or the capability to have clear view in night for some animals. It can be categorized into nine types, namely, linguistic natural intelligence, melodic natural intelligence, logical and mathematical natural intelligence, spatial natural intelligence, kinaesthetic’ natural intelligence, intrapersonal natural intelligence, interpersonal natural intelligence, memory natural intelligence and sensory natural intelligence (Wang et al., 2009) (Gardner et al., 1983). Linguistic natural intelligence is the ability to speak with the use of phonology and semantics. Melodic natural intelligence is the capability to create communication with a variety of sounds, pitch and rhythm. Logical and mathematical natural intelligence involves the creation of relationships between various factors. Spatial natural intelligence refers to the perception of visual information and reconstruction of visual images. Kinaesthetic’ natural intelligence refers to the use of the body parts to solve problems along with the use of motor skills of the body. Intrapersonal natural intelligence denotes to influence on others feelings, motivations and decisions. Interpersonal natural intelligence is used in the identification and making judgements on other persons’ decisions. Memory natural intelligence relates to remembering of facts, figures over a long time. Sensory natural intelligence is the extraordinary ability of the sensory parts of the body.
Key Terms in this Chapter
IoT: Internet of things is grid of internet linked objects with capability to analyse, gather and exchange data.
Feature Engineering: The study which involves obtaining precise features from huge data sets.
Cloud: Is a computing technique which provides on- demand services to store, process and compute databases.
Big Data: Indicates the huge volume of data sets used computationally to expose beneficial patterns or trends.
Problem Solving Method: The way of dealing with a problem for obtaining the best outcome.
Execution Time: The time taken for effecting an algorithm for production.
Intelligence: Is the capability to attain and apply knowledge in addition skills.
Interpretability: The ability to detect future aspects with available input parameters.
Extracted Feature Dependence: The level of dependency on the specific feature obtained from the data through feature engineering.