0% found this document useful (0 votes)
10 views20 pages

Knowledge Based System Lecture 2

The document discusses the relationship between data and artificial intelligence (AI), emphasizing how AI processes and analyzes data to enhance decision-making and predictive analytics. It covers various AI learning methods, including supervised, unsupervised, and reinforcement learning, along with examples of their applications. Additionally, it explains knowledge representation in AI through logical representation, semantic networks, frames, and production rules.

Uploaded by

md0746662
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
10 views20 pages

Knowledge Based System Lecture 2

The document discusses the relationship between data and artificial intelligence (AI), emphasizing how AI processes and analyzes data to enhance decision-making and predictive analytics. It covers various AI learning methods, including supervised, unsupervised, and reinforcement learning, along with examples of their applications. Additionally, it explains knowledge representation in AI through logical representation, semantic networks, frames, and production rules.

Uploaded by

md0746662
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 20

KNOWLEDGE BASED

SYSTEM
LECTURE 2
BY Farah Naz Anika
Data
• Data and artificial intelligence have a synergistic relationship. AI requires
a massive scale of data to learn and improve decision-making processes
and big data analytics leverages AI for better data analysis.
• AI's Ability to Process and Analyze Data
• Data Processing: AI systems can process enormous amounts of data far
more efficiently than humans. This ability allows organizations to gain
insights from big data, which would otherwise be impossible to analyze
manually.
• Real-Time Analysis: AI can analyze data in real time, enabling
applications like recommendation systems (e.g., Netflix, Amazon) and
predictive analytics (e.g., stock market forecasts, medical diagnoses).
Continues…
• 1. AI Enhancing Data Utilization
• Pattern Recognition: AI excels at recognizing patterns in data. This
capability is used in various applications, such as fraud detection,
image recognition, and natural language processing (NLP).
• Predictive Analytics: By learning from historical data, AI systems can
predict future trends, customer behavior, or even product demand,
helping businesses make data-driven decisions.
Contiunes…
• 4. AI Requires Continuous Data
• Learning and Adaptation: AI models need continuous data input to
improve over time. This concept is called "model retraining," where
new data helps the AI adapt to changing conditions or new
information.
• Feedback Loop: Many AI applications, such as chatbots and
recommendation engines, use user interactions as new data to refine
their algorithms and improve performance.
Information
• The relationship between AI (Artificial Intelligence) and information can be
explained conceptually by examining how AI relies on information to function
and, in turn, transforms information to deliver valuable insights, predictions, and
automation.
• AI as a Processor of Information
• Concept: At its core, AI is designed to process information. Information can be
structured data (like numbers, tables, or categories) or unstructured data (like
text, images, or videos). AI models use this input information to generate
outputs such as predictions, classifications, or decisions.
• Example: In a weather forecasting system, AI takes in historical weather data
(temperature, humidity, wind speed) and processes this information to predict
future weather conditions.
Information Feedback Loops in
AI
• Concept: Many AI systems operate in feedback loops, where the
information output by the AI feeds back into the system as new input.
This loop allows AI to refine its processes and become more accurate
over time.
• Example: AI-powered recommendation engines, such as those used
by Amazon or Netflix, continuously gather user interaction data
(clicks, views, ratings) and use this information to improve future
recommendations.
knowledge
• In the context of Artificial Intelligence (AI), knowledge refers to the information,
facts, rules, and insights that AI systems acquire, process, and utilize to perform
tasks, solve problems, and make decisions. Knowledge is a foundational aspect
of AI, enabling it to simulate human intelligence and behavior.
• Logical Representation
• Description: AI systems use formal logic (propositional or predicate logic) to
represent facts and rules. This representation allows for precise reasoning.
• Example: “If it rains, the ground will be wet” can be represented as "Rain →
WetGround."
• Use in AI: Logical representation is often used in rule-based systems and expert
systems for clear, rule-driven decision-making.
Semantic Networks

• Description: Knowledge is represented as a network of


interconnected concepts. Each node in the network represents a
concept, and the links define relationships between concepts.
• Example: A semantic network might represent the relationship
between "dog," "animal," and "pet."
• Use in AI: Semantic networks are widely used in natural language
processing (NLP) and knowledge graphs, such as in Google’s
Knowledge Graph.
Frames

• Description: Frames are data structures for representing stereotypical


situations, while scripts represent sequences of events in typical
situations.
• Example: A frame for a "restaurant" might include typical attributes
like tables, menus, and waiters, while a script might detail the steps of
dining at a restaurant.
• Use in AI: Frames and scripts are useful for AI systems that need to
understand context and make inferences, particularly in NLP and
understanding everyday scenarios.
Production Rules:

• AI, especially in the context of rule-based systems and expert


systems. In a production rule system, the AI operates based on if-then
rules (also called condition-action rules), where:
• IF a certain condition or set of conditions is true,
• THEN a specific action or conclusion is taken or inferred.
• Here’s a quick breakdown of Production Rule Systems in AI:
• Components of Production Rules:
• Condition (IF Part): The criteria that need to be met for the rule to be
applied. This is often based on the current state of the system or
specific input data.
• Example: "If the temperature is above 30°C."
• Action (THEN Part): What happens if the condition is satisfied. This
could be a decision, an inference, or triggering another rule.
Continues…
• Example in a Real-Life Scenario: Home Automation
• In a smart home system, production rules might be used to automate
actions based on environmental conditions or user preferences:
• Rule 1:
• IF the time is between 6:00 PM and 6:00 AM AND no one is in the living room,
• THEN turn off the lights.
• Rule 2:
• IF the front door opens AND the outdoor temperature is below 5°C,
• THEN turn on the heater.
learning in AI
• The concept of learning in AI refers to the ability of AI systems to
improve their performance on tasks over time through experience
and data. Rather than being explicitly programmed with all the rules
and knowledge they need, AI systems can learn from the data they
receive, adapt their behavior, and make better decisions based on
past experiences.
• In AI, learning is often categorized into three primary types:
supervised learning, unsupervised learning, and reinforcement
learning. Each of these approaches allows AI systems to learn and
evolve in different ways, based on the kind of data and the desired
outcomes.
learning in AI
• 1. Supervised Learning
• In supervised learning, the AI is trained on a labeled dataset, which
means that each input comes with the correct output. The goal is for
the AI to learn the mapping between inputs and outputs so that it can
predict the output for new, unseen inputs.
Continues
• Real-Life Example: Email Spam Detection
• One of the most common applications of supervised learning is email
spam detection. Email service providers like Gmail use AI systems to
automatically classify incoming emails as "spam" or "not spam" based
on a large dataset of previously classified emails.
• Input: Features of an email, such as the sender's address, subject line,
and email content.
• Output: The label "spam" or "not spam."
Continues
• Learning Process: The AI is trained on a dataset of thousands (or
millions) of emails that have been manually labeled as spam or not
spam. It learns to identify patterns and features (like certain phrases,
URLs, or frequent sender addresses) that are more likely to appear in
spam emails.
• Prediction: After learning from the data, the AI can now examine new
incoming emails and classify them as spam or not spam with high
accuracy.
• 2. Unsupervised Learning
• In unsupervised learning, the AI system is given data without explicit
labels or instructions. The AI must find hidden patterns, groupings, or
structures in the data without knowing in advance what the correct
output should be.
• Real-Life Example: Customer Segmentation in Marketing
• A common use case of unsupervised learning is customer segmentation, where
businesses use AI to group their customers based on similar characteristics for
targeted marketing strategies.
• Input: Customer data, including purchasing habits, age, location, and income.
• Output: Segments or groups of customers with similar behaviors or characteristics.
• Learning Process: The AI system analyzes the data and groups customers based on
patterns, such as those who make frequent purchases, those who prefer specific
types of products, or those who are more sensitive to discounts. The AI does this
without any pre-labeled categories.
• Outcome: Businesses can use these segments to create personalized marketing
campaigns, offering specific promotions to different customer groups.
3. Reinforcement Learning

• 3. Reinforcement Learning
• In reinforcement learning, an AI agent interacts with an environment
and learns to perform tasks by receiving feedback in the form of
rewards or penalties. The agent tries to maximize the cumulative
reward by learning from the consequences of its actions.
Example:
• Real-Life Example: Self-Driving Cars
• A great example of reinforcement learning is self-driving cars. The AI controlling the
car learns to navigate and drive by interacting with the environment, which includes
other cars, pedestrians, traffic lights, and road signs.
• Input: The environment, including sensor data like camera feeds, radar, and GPS.
• Output: The car's actions, such as steering, accelerating, or braking.
• Learning Process: The car receives feedback (rewards or penalties) based on its
actions. For example, successfully stopping at a red light might yield a positive reward,
while causing an accident might result in a penalty. The AI learns which actions
maximize safe driving and minimizes risks or penalties.
• Outcome: Over time, the self-driving car improves its driving performance, making
better decisions based on its previous experiences.

You might also like