The term AI was introduced by Prof. John McCarthy at a conference at Dartmouth College in 1956. McCarthy defines AI as the “science and engineering of making intelligent machines, especially intelligent computer programs”.
Augmented Intelligence[coined by former IBM CEO Ginni Rometty] :- Helping people and machines work together to create knowledge from data that enhances human expertise.
AI is a sub field of computer science that aims at building computer systems that can perform tasks that normally require human intelligence. AI-based machines are intended to perceive their environment and take actions that optimize their level of success. AI research uses techniques from many fields, such as linguistics, economics, and psychology.
![[Pasted image 20230601210217.png]]
- The first era of computing consisted of single-purpose electromechanical systems that used punched cards to input and store data that instructed the machine what to do.
- These tabulation machines were essentially calculators that counted and summarized information.
- These machines supported the scaling of business and society and were used in government applications, such as processing population census data, and business applications, such as accounting and inventory control.
- Tabulating machines evolved to a class of machines that were known as unit record equipment, which were instrumental in starting the data processing industry.
- This era started with the shift from mechanical tabulators to electronic systems. It began during World War II and was driven by military and scientific needs.
- Following the war, digital “computers” evolved rapidly and moved into businesses and governments. The programmable computing era began.
- They can be reprogrammed to perform different tasks and solve multiple problems in business and society.
- However, they must be programmed and are still constrained in their interaction with human beings.
- Everything we now know as a computing device, such as the mainframe, personal computer, smartphone, and the tablet, is a programmable computer.
- Some experts believe that this era of computing will continue to exist indefinitely.
- AI systems are a natural evolution of programmable computing.
- AI systems exhibit characteristics of human intelligence.
- They can understand and reason and learn and improve their performance overtime.
Dr. Kelly in the white paper Computing, cognition and the future of knowing defines cognitive computing as follows:
“Cognitive computing refers to systems that learn at scale, reason with purpose and interact with humans naturally. Rather than being explicitly programmed, they learn and reason from their interactions with us and from their experiences with their environment.”
Researchers and developers are adopting one of two approaches of thinking while building AI systems:
-
Strong AI: It builds systems that can act and think intelligently like people do. This approach simulates human reasoning and cognition in general tasks and is called strong AI. Most of these projects are in the small-to-medium size range. The three largest projects are DeepMind, the Human Brain Project (an academic project that is based in Lausanne, Switzerland), and OpenAI.
-
Weak AI: It is not concerned about whether the AI systems display human-like cognitive functions; the focus is on AI systems that perform specific tasks accurately and correctly. This approach is called weak AI. An example of weak AI is a chatbot, which is trained to answer a specific set of questions about a domain. A chatbot does not have self-awareness or genuine intelligence.
Current AI is somewhere between strong and weak AI. The current most advanced AI research uses human reasoning as a model but the AI field has not modeled human reasoning yet. A good example of current AI is advanced natural language generation.
In terms of scope, Ray Kurzweil, who is a futurist that is working for Google, identifies three levels of AI:
- ANI, which belongs to the weak AI school, refers to applying AI techniques to narrow problems.
- It focuses on a specific task.
- The strength of ANI is that it focuses on doing something extremely well, sometimes exceeding a human’s capabilities.
- ANI is a good fit for automating simple and repetitive tasks.
- Examples of ANI are bots and virtual assistants, such as Siri, Microsoft Cortana, Amazon book recommendations that are customized for each user, restaurant recommendations, weather updates, Watson DeepQA, and customer services chatbots for answering simple and repetitive customer inquiries.
- AGI belongs to the strong AI school.
- It refers to computer systems that exhibit capabilities of the human brain.
- AGI refers to systems or machines that can generally perform any intellectual task that a human can do.
- ASI refers to machines that surpass humans in general intelligence.
- Nick Bostrom, a leading AI thinker, in the paper How Long Before Superintelligence?, defines ASI as “an intellect that is much smarter than the best human brains in practically every field, including scientific creativity, general wisdom and social skills.”
- Many prominent scientists and technologists have ethical concerns about the future of humanity and intelligent life.
- The unique capabilities of the human brain are the reason why humans have a dominant position over other species.
- Super-intelligent machines might surpass the human brain in general intelligence.
According to Ray Kurzweil, AGI will be there around the year 2040 and by 2060 we will have ASI.
Intelligence or the cognitive process is composed of three main stages:
- Observe and input the information or data in the brain.
- Interpret and evaluate the input that is received from the surrounding environment.
- Make decisions as a reaction towards what you received as an input and interpreted and evaluated.
AI researchers are simulating the same stages in building AI systems or models. This process represents the main three layers or components of AI systems.
Because AI is the science of simulating human thinking, it is possible to map the human thinking stages to the layers or components of AI systems.
In the first stage, humans acquire information from their surrounding environments through human senses, such as sight, hearing, smell, taste, and touch, through human organs, such as eyes, ears, and other sensing organs, for example, the hands. In AI models, this stage is represented by the sensing layer, which perceives information from the surrounding environment. This information is specific to the AI application. For example, there are sensing agents such as voice recognition for sensing voice and visual imaging recognition for sensing images. Thus, these agents or sensors take the role of the hearing and sight senses in humans.
The second stage is related to interpreting and evaluating the input data. In AI, this stage is represented by the interpretation layer, that is, reasoning and thinking about the gathered input that is acquired by the sensing layer.
The third stage is related to taking action or making decisions. After evaluating the input data, the interacting layer performs the necessary tasks. Robotic movement control and speech generation are examples of functions that are implemented in the interacting layer.
Consumer-oriented AI systems such as Google Assistant, Siri, and Microsoft Cortana are adopting the three AI components or layers: Sensing, interpretation, and interacting.
- These systems are composed of multiple components and approaches for thinking and reasoning, and simulating human thinking.
- These combinations represent the identity of the AI system and distinguish one AI system from others.
- For AI systems to be able to assess data, there must be enough data with which to work.
- Using a large volume of data can support the learning process of even a weak AI algorithm to identify the relevant data and discard the “noise” or irrelevant data.
- Availability, variety, and volume of data are factors that contribute to enhancing the learning process and produce more accurate results.
- Many AI systems focus on prediction. To predict an outcome, AI systems use data mining or machine learning algorithms.
- There are different techniques that are used for prediction, such as regression analysis, which is a set of statistical processes for estimating the relationship among variables.
- AI systems build rules that can be used to predict the events of interest before they take place.
According to the type of predicted outcome, AI systems can be categorized into the following two categories:
- Deterministic systems - A deterministic system is a system in which a given initial state or condition will always produce the same results. There is no randomness or variation in the ways that inputs get delivered as outputs.
- Probabilistic systems - A probabilistic system is a system in which the occurrence of events cannot be perfectly predicted. Though the behavior of such a system can be described in terms of probability, a certain degree of error is always attached to the prediction of the behavior of the system.
- Big data refers to huge amounts of data. Big data characteristics are defined by the three Vs:- Variety, Volume, Velocity. Later, Veracity and Visibility were added.
- Big data requires innovative forms of information processing to draw insights, automate processes, and help decision making.
- Big data can be structured data that corresponds to a formal pattern, such as traditional data sets and databases.
- Also, big data includes semi-structured and unstructured formats, such as word-processing documents, videos, images, audio, presentations, social media interactions, streams, web pages, and many other kinds of content.
![[Pasted image 20230602195435.png]]
- Significant advancements in computer processing and memory speeds enable us to make sense of the information that is generated by big data more quickly.
- In recent years, big data and the ability to process a large amount of data at high speeds have enabled researchers and developers to access and work with massive sets of data.
- Processing speeds and new computer chip architectures contribute to the rapid evolution of AI applications.
- The meaning of big data expanded beyond the volume of data after the release of a paper by Google on MapReduce and the Google File System (GFS), which evolved into the Apache Hadoop open source project.
- The Hadoop file system is a distributed file system that may run on a cluster of commodity machines, where the storage of data is distributed among the cluster and the processing is distributed too.
- This approach determines the speed with which data is processed. This approach includes an element of complexity with the introduction of new, structured, unstructured, and multi-structured data types.
- Large manufacturers of computer chips such as IBM and Intel are prototyping “brain-like” chips whose architecture is configured to mimic the biological brain’s network of neurons and the connections between the called synapses.
-
Cloud computing is a general term that describes delivery of on-demand services, usually through the internet, on a pay-per-use basis.
-
Companies worldwide offer their services to customers over cloud platforms.
-
These services might be data analysis, social media, video storage, e-commerce, and AI capabilities that are available through the internet and supported by cloud computing.
-
In general, application programming interfaces (APIs) expose capabilities and services.
-
APIs enable software components to communicate with each other easily.
-
The use of APIs as a method for integration injects a level of flexibility into the application lifecycle by making the task easier to connect and interface with other applications or services.
-
APIs abstract the underlying workings of a service, application, or tool, and expose only what a developer needs, so programming becomes easier and faster.
-
AI APIs are usually delivered on an open cloud-based platform on which developers can infuse AI capabilities into digital applications, products, and operations by using one or more of the available APIs.
-
All the significant companies in the AI services market deliver their services and tools on the internet through APIs over cloud platforms:
- IBM delivers Watson AI services over IBM Cloud.
- Amazon AI services are delivered over Amazon Web Services (AWS).
- Microsoft AI tools are available over the MS Azure cloud.
- Google AI services are available in the Google Cloud Platform.
These services benefit from cloud platform capabilities, such as availability, scalability, accessibility, rapid deployment, flexible billing options, simpler operations, and management.
-
Transportation Autonomous transportation refers to self-driven vehicles, such as driverless cars and unmanned ground vehicles (UGVs), which are vehicles that can sense their environment and navigate without human input. Autonomous transportation will soon be commonplace, and likely include cars, trucks, flying vehicles, and personal robots.
-
Home services and robots Home services and robots have already entered people’s homes in the form of vacuum cleaners and personal assistants. Better chips, low-cost 3D sensors, cloud-based machine learning, and advances in speech understanding will enhance future robots’ services and their interactions with people. Special purpose robots will clean offices, and enhance security. Drones are already being used to deliver packages.
-
Healthcare In healthcare, there has been a huge forward leap in collecting useful data from personal monitoring devices and mobile apps, electronic health records (EHRs) in clinical settings, surgical robots that assist with medical procedures, and service robots that support hospital operations. AI applications can generate new insights that speed up matching subjects to clinical studies, identify promising targets for research, reduce drug discovery times, and provide virtual assistance to patients.
-
Education Even when quality education will always require active engagement by human teachers, AI can enhance education at all levels, especially by providing personalization at scale. Interactive machine tutors are now being matched to students for teaching science, math, language, and other disciplines. Natural Language Processing, machine learning, and crowdsourcing have boosted online learning and enabled teachers in higher education to multiply the size of their classrooms while addressing individual students’ learning needs and styles.
-
Public safety and security AI technologies are already in use in North America for border administration and law enforcement. By 2030, the government will rely heavily upon them, including improved cameras and drones for surveillance, algorithms to detect financial fraud, and predictive policing.
-
Employment and workplace AI is poised to replace people in certain jobs, such as driving taxis and trucks. However, in many realms, AI will likely replace tasks rather than jobs in the near term, and also create new jobs. But the new jobs that emerge are harder to imagine in advance than the existing jobs that will likely be lost.
-
Entertainment Entertainment has been transformed by social networks and other platforms for sharing and browsing blogs, videos, and photos, which rely on techniques actively developed in Natural Language Processing, information retrieval, image processing, and machine learning. Some traditional sources of entertainment have also embraced AI to compose music, create stage performances, and even to generate 3D scenes from natural language text. AI will increasingly enable entertainment that is more interactive, personalized, and engaging.
Other examples of AI solutions for industry vertical sectors include the following ones:
- Agriculture
- Banking, Financial Services, and Insurance (BFSI)
- Manufacturing
- Oil and Gas
- Machine learning is a methodology of taking large amounts of data and finding a reasonably “optimal” mathematical solution that can be used to predict results given future data.
- This methodology is not seen as “proper” AI because it is more a mathematical and statistical technique or algorithm. However, this technique can produce impressive results when it is used to train computer systems that “learn” from training data.
- Machine learning is used to support the other AI subfields, including natural language processing (NLP) and computer vision (CV).
- NLP is the subfield of AI that applies computational techniques to analyze and synthesize human natural language and speech.
- NLP is the study of the computational treatment of natural (human) language. It is the branch of science that deals with analyses of text, whether they are in emails, human speech, social media and more.
- It allows machines to understand human communication to extract different information that was so far, the exclusive privilege of humans.
- NLP is found today in the following types of applications:- Machine translation, Search engines, such as Google and Baidu, Spell checkers, IBM Watson(where it famously won Jeopardy! against the best human contestants), Natural language assistants, such as Siri, Translation systems, such as Google translate, News digest, such as Yahoo
- CV is the subfield of AI that applies computational techniques to understand and process images in pictures and videos.
- CV uses deep learning algorithms to analyze images for scenes, objects, faces, and other content in videos, photos and pictures in general.
- Business analytics solutions provide organizations with actionable insights for smarter decisions and better business outcomes