TEst presentation for AIML learning.pptx

AIML

1. Hybrid Learning System Hand-Built Classifiers Empirical Learning Artificial Neural Networks (ANNs) Case-Based Reasoning (CBR) RULE BASED SYSTEM Fuzzy System

2. Hybrid Learning System Hybrid learning methods combine theoretical knowledge of a domain with a set of classified examples to develop an approach for accurately classifying new, unseen examples

3. Hybrid Learning System Hybrid learning methods combine theoretical knowledge of a domain with a set of classified examples to develop an approach for accurately classifying new, unseen examples

4. Empirical Learning Empirical learning is a method of acquiring knowledge through direct or indirect observation and experience, where the gathered evidence can be analyzed quantitatively or qualitatively. Empirical learning systems do not require extensive theoretical knowledge of the problem domain; instead, they rely on a large collection of examples. According to A.D. de Groot, empirical learning involves five cycles or stages.

5. Hand-Built Classifiers Hand-built classifiers are non-learning systems that operate strictly based on predefined instructions and do not acquire knowledge through learning (unless manually modified later). Despite their straightforward design, these systems can present significant challenges for those who develop them. They typically assume that their domain theory is comprehensive; however, achieving completeness and accuracy is extremely difficult, if not impossible, for most real-world tasks. Creating a fully accurate domain theory may require writing thousands of interdependent, possibly recursive rules, which can lead to very slow performance. Additionally, such theories can be difficult to modify or update .

6. Artificial Neural Networks (ANNs) Artificial neural networks (ANNs) are designed to mimic the way the human brain processes information. As a method for empirical learning, ANNs are the foundation of knowledge-based artificial neural networks (KBANN). They have been shown to be equal to or better than other empirical learning systems across a wide range of domains, particularly when measured by their ability to generalize. ANNs are effective in solving complex, data-intensive problems where the rules or algorithms are either unknown or difficult to articulate. They process information in parallel and are resilient to errors in data. ANNs can be applied to various categories of problems, including pattern classification, clustering, function approximation, prediction, optimization, content-based retrieval, and process control.

7. Case-Based Reasoning (CBR) Case-based reasoning (CBR) addresses new problems by recalling solutions from similar past cases. It relies on a substantial collection of previous cases to adapt their solutions or methods to current problems. CBR operates on the principle that repeated attempts make problems easier to solve, promoting continuous learning. The process involves four steps, as illustrated in Figure 1.4: 1. Retrieve the most relevant past cases from the database. 2. Apply the retrieved case to develop a solution for the new problem. 3. Refine the proposed solution through simulation or test execution. 4. Store the successfully adapted solution for future reference

8. 1. RULE BASED SYSTEM Rule-based systems (RBS) address problems using expert knowledge encoded in rules. These rules consist of conditions and actions, often formulated as "if-then" statements, which are processed by an inference engine. This engine includes a working memory that holds information about the problem, a pattern matcher to determine relevant rules, and a rule applier to execute these rules. When a rule is applied, the resulting new information is added to the working memory, and the process of matching, selecting, and acting on rules continues until no more applicable rules are found. RBS are straightforward to understand, implement, and maintain due to their structured approach to knowledge representation. However, they rely on predefined rules and lack the capability for learning or automatic rule modification. Thus, their effectiveness is contingent upon having complete and accurate knowledge available beforehand.

9. Fuzzy System Fuzzy systems (FS) use fuzzy sets to manage imprecise or incomplete data. Unlike conventional set theory, where an object is either a member of a set or not, fuzzy sets allow membership values to range between 0 and 1. This approach enables fuzzy models to articulate vague or ambiguous statements similar to natural language. Fuzzy systems are adept at managing unclear information, which is a significant advantage over other AI techniques, making them relatively simple to understand and apply. However, fuzzy systems lack learning capabilities and memory. To address these limitations, fuzzy modeling is often combined with other techniques, such as neural networks, to create hybrid systems like Neuro-fuzzy systems. Fuzzy systems are utilized in various applications, including function approximation, classification, clustering, control, and prediction.

10. ARTIFICIAL INTELLIGENCES A Program that can sense, reason, act, and adapt MACHINE LEARNING Algorithms whose performance improve as they are exposed to more data over time DEEP LEARNING Subset of machine learning in which multilayered neural networks learn from vast amount of data

11. Machine learning Machine learning (ML) is the process of using mathematical models of data to help a computer learn without direct instruction. Uses algorithms to identify patterns within data and those patterns are then used to create a data model that can make predictions. With increased data and experience, the results of machine learning are more accurate. Required domain expertise and human engineering to design feature extractors that transformed raw data into suitable representations from which a learning algorithm could detect patterns. microsoft.com

12. Machine learning techniques Supervised learning Algorithms make predictions based on a set of labelled examples from past experience Useful when you know what outcome should be Unsupervised learning Algorithm allows machine to work on unlabeled data and discover pattern on its own that was previously undetected Useful when you do not know what the outcome should look like Reinforcement learning Algorithms that learn from outcomes and decide which action to take next Algorithm receives feedback that helps it determine whether the choice it made was correct, neutral or incorrect

13. AIM