ملخص.pdf

What is AI? Artificial intelligence is the science of making computers do things that require intelligence when done by humans (Russell & Norvig, 2020). What is AI? Cont. Sensors Perceive environment...

What is AI? Artificial intelligence is the science of making computers do things that require intelligence when done by humans (Russell & Norvig, 2020). What is AI? Cont. Sensors Perceive environment : e.g. sensors and cameras Agent System takes actions – to make decision – based on Decision- perception Actuators Acting – e.g. making robotic arm Decides what or a text-to- actions to speech take system What is AI? Cont. AI = Ability to learn (intelligence) AI = Ability to plan (rationality) AI = Ability to communicate in natural language (abstraction) AI = Ability to reason about morality (humanity) Search Algorithms Uninformed Search: explore the search space without having any information about the problem other than its definition. Informed Search: know whether one particular state is better than another. They use heuristics to guide the search process Local Search: move from solution to solution in the space of candidate solutions by making local changes Adversarial Search in AI Adversarial search: various players compete and the result of the search depends on the actions of all the players. (e.g. AI games). Constraint Satisfaction Problems (CSPs) CSP Solving Techniques: Backtracking Forward Checking Constraint Propagation. Knowledge Representation Knowledge Representation (KR) is the way we organize and structure information in artificial intelligence. It's how an AI system understands and uses information to solve problems. Uncertainty in AI Uncertainty in AI refers to situations where the state of the world is not completely known, or the outcomes of actions are not completely predictable. This can be due to inherent randomness, incomplete or noisy data, or the complexity of the world. Uncertainty in AI Cont. Types of Uncertainty: Aleatory uncertainty (inherent randomness) Flipping a coin and predicting either HEADS or TAILS ? Epistemic uncertainty (lack of knowledge) Incomplete knowledge of the phenomena ! Uncertainty in AI Cont. Methods for dealing with uncertainty Probabilistic reasoning Fuzzy logic Bayesian networks Learning in AI Learning in AI is a process that allows artificial intelligence systems to improve their performance on specific tasks over time. This is achieved by modifying the system's internal parameters based on past experiences and feedback, thereby allowing the AI to 'learn' from its mistakes and successes. Learning in AI Cont. Types of Learning Supervised Reinforcement Learning Learning Unsupervised Learning Learning in AI Cont. Supervised Learning: The AI is provided with input-output pairs, where the output is the 'correct' answer for each input. The AI learns a function that maps inputs to correct outputs. Unsupervised Learning: The AI is given inputs but no explicit outputs. The goal is to find structure in the inputs, such as clustering or dimensionality reduction. Reinforcement Learning: The AI learns how to perform actions based on reward feedback. It's about making a sequence of decisions, with the goal of maximizing a reward signal. Learning in AI Cont. Importance of Learning To allows AI systems to adapt to new circumstances, generalize from previous experiences, and improve over time. Communicating in AI Cont. Natural Language Processing (NLP): It focuses on the interaction between computers and humans through language. (More in Lecture 5) Communicating in AI Cont. Multi-Agent Systems: Multiple AI systems need to coordinate their behavior. End of Lecture 1 What is Data Mining? Cont. A type of supervised learning approach where the goal is to predict the categorical class labels of new instances, based on past observations. Classification These class labels are discrete, unordered values that can be understood as the group memberships of the instances What is Data Mining? Cont. A type of unsupervised learning technique used in machine learning and data mining. Clustering The goal of clustering is to group a set of objects in such a way that objects in the same group (called a cluster) are more similar to each other than to those in other groups (clusters). Mathematical Preliminaries Cont. Vectors “Vectors are built from components, which are ordinary numbers. You can think of a vector as a list of numbers, and vector algebra as operations performed on the numbers in the list” - Page 69, No Bullshit Guide To Linear Algebra, 2017 e.g. x = (3.2, -9.1, 0.1) Mathematical Preliminaries Cont. Matrices We obtain matrices by arranging a collection of vectors by columns or rows. A matrix is a two-dimensional array that has a fixed number of rows and columns and contains a number at the intersection of each row and column. A matrix is usually delimited by square brackets. Example Here is an example of a matrix having two rows and two columns: Mathematical Preliminaries Cont. If a matrix has rows and columns, we say it has dimension or is a matrix. Example The matrix A matrix has rows and columns. So, we say that it is a matrix. Data Representation Categorical data discrete categories (colour of a flower petal) Numerical data integer values (number of petals in a flower) real values (length of a petal) String/textual data words in a document time series data / sequential data continuous, chronological, flows in one direction  Data Representation Cont. Method 1: By a list of words? [“the”, “burger”, “i”, “ate”, “was”, “an”, “awesome”, “burger”] Method 2: By the set of words? {“the”, “burger”, “i”, “ate”, “was”, “an”, “awesome”} Method 3: By a vector of word frequency? (“the”:1, “burger”:2, “i”:1, “ate”:1, “was”:1, “an”:1, “awesome”:1) Method 4: By a vector of letter frequency? {‘a’: 3, ' ': 7, 'b': 2, 'e': 6, 'g': 2, 'i': 2, 'h': 1, 'm': 1, 'o': 1, 'n': 1, 's': 2, 'r': 4, 'u': 2, 't': 2, 'w': 1} Data Pre-processing Cont. Data Discretization: This involves converting continuous data into discrete buckets or intervals. It's useful for reducing the complexity of data and improving the efficiency of certain algorithms. Data Reduction: The purpose here is to reduce the volume but produce the same or similar analytical results. Data reduction can be achieved by various methods including binning, histograms, clustering, and Principal Component Analysis (PCA). Feature Selection: This step involves selecting the most important features to use in data analysis. The aim is to reduce the number of input variables when developing a predictive model. It can be done using various techniques like statistical tests, selection based on models, or iterative algorithms. Dimensionality Reduction Dimensionality reduction in data mining refers to the process of reducing the number of random variables under consideration by obtaining a set of principal variables. It is a critical step in the pre-processing of high-dimensional data sets to simplify models, speed up computation, and reduce the risks of overfitting. Dimensionality Reduction Cont. Principal Component Analysis (PCA) PCA transforms correlated variables into a smaller number of uncorrelated variables called principal components. The first principal component captures the most variance, with each subsequent component orthogonal to the last and capturing the next most variance. PCA is used for dimensionality reduction in exploratory data analysis and predictive modeling. Information Retrieval Cont. 1. Query Processing This involves interpreting the user's query and possibly transforming it into a more effective query that can retrieve the desired results. It can include parsing, tokenization, and sometimes query expansion to include synonyms or alternative terms. 2. Indexing To efficiently retrieve information, data needs to be indexed. Indexing involves creating data structures that allow for quick searching, such as inverted indices that map keywords to their locations in a set of documents. 3. Search and Ranking Search algorithms take a user's query and return a set of results. Ranking algorithms then order these results by relevance, which can be based on various factors such as keyword matches, the context of the query, the user's personal profile, or the page's authority (in the case of web pages). Graph Mining A Graph G can be defined as a set of vertices (nodes) V connected by a set of edges (links) E A graph G(V,E) is fully defined by specifying the two sets V and E Sequential Data It is a type of data that is arranged in a sequence, often based on time. This data is characterized by the order in which the data points occur, which is critical to the analysis. Sequential data can be found in various domains such as retail (customer purchase histories), finance (stock prices), weather (temperature readings over time), and many others. Sequential Data Cont. Time Series Data: Data points collected or recorded at successive points in time. For example, daily stock market Types of prices. Sequence Data: Data where the order of items is Sequential important but not necessarily tied to a specific time interval. For example, a sequence of web pages visited by a user. Data Event Data: Data where the timestamp of each event is recorded. For example, logs of user actions in an application. Sequential Data Cont. Sequential Pattern Discovery: Identifying frequently occurring patterns or subsequences within the data. For example, finding common sequences of symptoms in medical records. Data Anomaly Detection: Finding unusual or unexpected sequences which can indicate errors, fraud, or other significant events. Prediction: Using historical sequential data to predict future Mining events or values. For example, predicting the next purchase of a customer based on their transaction history. Tasks Segmentation: Dividing the data into segments based on similar characteristics or behaviors within the sequence.