AI Foundations
- [The University of Chicago]
- Overview
Historically, the following disciplines have focused on artificial intelligence (AI) as their ultimate outcome: philosophy, mathematics, economics, neuroscience, psychology, computer engineering, control theory and cybernetics, and linguistics.
The foundations of AI are the fundamental areas that AI sits at the intersection of: logic, computation, probability, programming, data science, mathematical reasoning, creative problem solving, ethics, and applied experiences.
The four basic concepts of AI are: machine learning, natural language processing, computer vision, and robotics. The Foundations of AI focuses on developing algorithms that use data and and statistical tools to solve complex human tasks.
AI is a technology that uses intelligent systems to perform tasks that usually require human intelligence. The four main components of AI are: learning, reasoning and decision making, problem solving, and perceptions.
Some branches of AI include: computer vision, fuzzy logic, expert systems, robotics, machine learning, neural networks/deep learning, and natural language processing.
- Reasoning
Reasoning in AI is the process of using logical rules and principles to derive new information from existing information. It's a key component of AI applications like natural language processing, machine learning (ML), and expert systems.
Reasoning in AI aims to create machines that can reason like humans, using logic, common sense, and intuition. It's essential in the development of AI because it allows robots to reason and act like humans.
A reasoning system is a software system that uses logical techniques like deduction and induction to generate conclusions from a knowledge base (KB). Reasoning systems play a large role in the implementation of AI and in knowledge-based systems.
Classification algorithms like logistic regression perform well at inductive reasoning. They have a target variable and use specific features to draw a larger conclusion.
- Inference
Inference is the critical moment for an AI model, a test of its ability to apply information learned during training to predict or solve a task. Can it accurately mark incoming emails as spam, transcribe conversations, or summarize reports?
During inference, the AI model begins processing live data, comparing the user's query to information processed during training and stored in its weights or parameters.
The response the model sends back depends on the task, whether it's identifying spam, converting speech to text, or distilling a long document into its key takeaways. The goal of AI inference is to calculate and output actionable results.
- Reasoning vs. Inference in AI
Reasoning plays a huge role in the process of AI. Therefore, reasoning can be defined as the logical process of drawing conclusions, making predictions, or constructing approaches to a specific idea with the help of existing knowledge.
In AI, reasoning is very important because to understand the human brain, how the brain thinks, and how to draw conclusions about specific things, all these tasks require the help of reasoning.
Inference is a process by which new sentences are derived from existing sentences in knowledge base (KB). E.g. Modus tollens is a rule of inference which derives new knowledge. Reasoning is the act of deriving a conclusion from certain premise using given methodology.
- Learning and Adaption
One of the fundamental characteristics of AI is its ability to learn and adapt. Machine learning is a subset of AI that enables systems to continuously improve their performance through continuous learning. AI systems are able to examine extensive data sets, identify recurring trends, and formulate predictions or choices based on this data.
For example, in image recognition, AI models can better recognize objects or faces each time they are exposed to new material. This learning and adapting characteristic is important in the following applications:
- Recommendation system: AI algorithms analyze user behavior and preferences to recommend products, movies or content that suit personal tastes.
- Natural Language Processing (NLP): Language models such as GPT-3 can produce human-like text by learning from large amounts of text data.
- Self-driving cars: Self-driving cars use machine learning to adapt to changing traffic conditions and make driving decisions
- Problem Solving
Artificial intelligence excels at solving complex problems involving large amounts of data and decisions. Unlike traditional programming that encodes explicit rules, the system can find solutions by analyzing the data and adjusting its internal parameters. It enables AI to handle complex tasks in various fields:
- Medical diagnosis: AI can analyze medical data and help diagnose diseases by identifying patterns and anomalies.
- Optimization: AI can optimize supply chains, transportation routes, and resource allocation to increase efficiency and reduce costs.
- Playing games: AI has achieved superhuman performance in chess, Go and video games by using advanced problem-solving techniques.
- Perception
Artificial intelligence systems can sense the environment through various sensors. This is critical in the following applications:
- Computer Vision: AI can interpret visual data to enable tasks such as facial recognition, object detection, and autonomous vehicle navigation.
- Speech recognition: AI can convert spoken words into text, which is valuable for transcription services and voice assistants.
- Environmental monitoring: AI can analyze sensor data to detect environmental changes, such as air quality or seismic activity.
- Autonomous Decision-making
AI systems can make autonomous decisions based on data and predefined rules. This is common in applications such as self-driving cars, where AI algorithms make on-the-fly decisions on driving behavior based on sensor data and safety rules. Independent decision-making also extends to robotics, which can perform tasks independently.
