Foundations and Sub-areas of AI

Foundations of AI and Sub-areas of AI

• Explores five key areas of Artificial Intelligence (AI): Machine Learning, Expert Systems, Computer Vision, Natural Language Processing (NLP), and Robotics.

Machine Learning

• Focuses on algorithms that enable data-driven decisions.
• Employs computers to identify patterns and relationships in various data types (images, sounds, structured arrays).
• Uses mathematical models and algorithms.
• A critical tool in healthcare, finance, e-commerce, and robotics.

Types of Machine Learning

• Supervised Learning: Uses labeled datasets to train models and predict outcomes for new inputs.
• Unsupervised Learning: Works with unlabeled datasets to discover hidden patterns or anomalies used in customer segmentation, anomaly detection and data exploration.
• Semi-Supervised Learning: Combines labeled and unlabeled data, useful when labeled data is scarce.
• Reinforcement Learning: An agent learns to achieve objectives by interacting with an environment, improving decisions through trial and error, used in robotics and autonomous vehicles.

Expert Systems

• A human expert's decision-making ability is simulated using pre-programmed knowledge and logical rules.
• Provides advice or makes decisions in specific domains.
• Follows a fixed set of rules.
• Relevant where stability and rule-based logic are essential.
• Google's Nest home automation system is an example.
• Other examples include MYCIN (medical diagnosis), DENDRAL (chemical analysis), PXDES (lung diseases), and CADET (military planning).

Computer Vision

• Automates extraction, analysis, and interpretation of information from images/videos.
• Transforms visual data into numerical arrays to enable ML algorithms.
• Enables machines to "see" and understand the world in ways similar to humans.
• Advancements in algorithms (CNNs), GPU resources, distributed architectures, cloud computing, and data availability have made it possible.
• Self-driving cars and automated retail systems like Amazon Go are examples.

Natural Language Processing

• Enables machines to automatically process, analyze, and generate human language.
• Algorithms parse sentences by splitting them into words/letters or reading them left-to-right and right-to-left.
• Numerous use cases across industries: Named Entity Recognition (NER), Part-of-speech tagging, Reading comprehension, Machine translation, Text summarization, Spellcheck and autocomplete.
• Innovations in deep learning have made it faster and easier to train ML models on human language.
• Siri and Alexa are prime examples.

Robotics

• Branch of science designing, constructing, operating, and applying robots to solve problems and automate tasks.
• Robots range in forms and sizes.
• Robotics research has made tremendous strides in technology.
• Many robots have relied on expert systems. The robots of tomorrow integrate ML, CV, and NLP.
• Robots will be able to adapt and improve their performance and handle a wider variety of tasks, with the use of ML.
• With CV, robots will see and understand their environment, and with NLP, they will understand and process human language.

AI Applications

• AI in E-Commerce: Enhances customer experience and operational efficiency through personalized shopping, Al-powered assistants (chatbots), and fraud prevention.
• AI in Navigation: Enhances user experience by providing accurate, timely information.
• AI in Robotics: Al-powered robots can perform tasks like carrying goods in hospitals, cleaning offices, and managing inventory.
• AI in Human Resources: Simplifies the hiring process by analyzing resumes and profiles to filter candidates.
• AI in Healthcare: Utilized to build systems that can detect diseases and identify conditions, aiding in drug discovery.
• AI in Agriculture: Improves crop production and soil health, detects nutrient deficiencies and weeds.
• AI in Gaming: Creates smart, human-like non-playable characters (NPCs).
• AI in Automobiles: Key in self-driving vehicles and offers safety features.
• AI in Social Media: Enhances user experience, recommends posts, translates and filters content.
• AI in Marketing: Delivers targeted and personalized ads and aids in content marketing.