The document outlines the objectives, outcomes, and learning outcomes of a course on artificial intelligence. The objectives include conceptualizing ideas and techniques for intelligent systems, understanding mechanisms of intelligent thought and action, and understanding advanced representation and search techniques. Outcomes include developing an understanding of AI building blocks, choosing appropriate problem solving methods, analyzing strengths and weaknesses of AI approaches, and designing models for reasoning with uncertainty. Learning outcomes include knowledge, intellectual skills, practical skills, and transferable skills in artificial intelligence.
HML: Historical View and Trends of Deep LearningYan Xu
The document provides a historical view and trends of deep learning. It discusses that deep learning models have evolved in several waves since the 1940s, with key developments including the backpropagation algorithm in 1986 and deep belief networks with pretraining in 2006. Current trends include growing datasets, increasing numbers of neurons and connections per neuron, and higher accuracy on tasks involving vision, NLP and games. Research trends focus on generative models, domain alignment, meta-learning, using graphs as inputs, and program induction.
Knowledge representation In Artificial IntelligenceRamla Sheikh
facts, information, and skills acquired through experience or education; the theoretical or practical understanding of a subject.
Knowledge = information + rules
EXAMPLE
Doctors, managers.
The document discusses different types of knowledge that may need to be represented in AI systems, including objects, events, performance, and meta-knowledge. It also discusses representing knowledge at two levels: the knowledge level containing facts, and the symbol level containing representations of objects defined in terms of symbols. Common ways of representing knowledge mentioned include using English, logic, relations, semantic networks, frames, and rules. The document also discusses using knowledge for applications like learning, reasoning, and different approaches to machine learning such as skill refinement, knowledge acquisition, taking advice, problem solving, induction, discovery, and analogy.
The document discusses procedural versus declarative knowledge representation and how logic programming languages like Prolog allow knowledge to be represented declaratively through logical rules. It also covers topics like forward and backward reasoning, matching rules to facts in working memory, and using control knowledge to guide the problem solving process. Logic programming represents knowledge through Horn clauses and uses backward chaining inference to attempt to prove goals.
Deep learning is a type of machine learning that uses neural networks inspired by the human brain. It has been successfully applied to problems like image recognition, speech recognition, and natural language processing. Deep learning requires large datasets, clear goals, computing power, and neural network architectures. Popular deep learning models include convolutional neural networks and recurrent neural networks. Researchers like Geoffry Hinton and companies like Google have advanced the field through innovations that have won image recognition challenges. Deep learning will continue solving harder artificial intelligence problems by learning from massive amounts of data.
Non-monotonic reasoning allows conclusions to be retracted when new information is introduced. It is used to model plausible reasoning where defaults may be overridden. For example, it is typically true that birds fly, so we could conclude that Tweety flies since Tweety is a bird. However, if we are later told Tweety is a penguin, we would retract the conclusion that Tweety flies since penguins do not fly despite being birds. Non-monotonic reasoning resolves inconsistencies by removing conclusions derived from default rules when specific countervailing information is received.
Knowledge Representation in Artificial intelligence Yasir Khan
This document discusses different methods of knowledge representation in artificial intelligence, including logical representations, semantic networks, production rules, and frames. Logical representations use formal logics like propositional logic and first-order predicate logic to represent facts and relationships. Semantic networks represent knowledge graphically as nodes and edges to model concepts and their relationships. Production rules represent knowledge as condition-action pairs to model problem-solving. Frames represent stereotyped situations as templates with slots to model attributes and behaviors. Choosing the right knowledge representation method is important for building successful AI systems.
The document discusses capsule networks, a type of neural network proposed by Geoff Hinton in 2017 as an alternative to convolutional neural networks (CNNs) for computer vision tasks. Capsule networks aim to address some limitations of CNNs, such as their inability to capture spatial relationships and pose information. The key concepts discussed include dynamic routing between capsules, which allows for parts-based representation, and equivariance, where capsules can learn transformation properties like position and orientation. The document provides an overview of a capsule network architecture and routing algorithm proposed in a 2017 paper by Sabour et al.
HML: Historical View and Trends of Deep LearningYan Xu
The document provides a historical view and trends of deep learning. It discusses that deep learning models have evolved in several waves since the 1940s, with key developments including the backpropagation algorithm in 1986 and deep belief networks with pretraining in 2006. Current trends include growing datasets, increasing numbers of neurons and connections per neuron, and higher accuracy on tasks involving vision, NLP and games. Research trends focus on generative models, domain alignment, meta-learning, using graphs as inputs, and program induction.
Knowledge representation In Artificial IntelligenceRamla Sheikh
facts, information, and skills acquired through experience or education; the theoretical or practical understanding of a subject.
Knowledge = information + rules
EXAMPLE
Doctors, managers.
The document discusses different types of knowledge that may need to be represented in AI systems, including objects, events, performance, and meta-knowledge. It also discusses representing knowledge at two levels: the knowledge level containing facts, and the symbol level containing representations of objects defined in terms of symbols. Common ways of representing knowledge mentioned include using English, logic, relations, semantic networks, frames, and rules. The document also discusses using knowledge for applications like learning, reasoning, and different approaches to machine learning such as skill refinement, knowledge acquisition, taking advice, problem solving, induction, discovery, and analogy.
The document discusses procedural versus declarative knowledge representation and how logic programming languages like Prolog allow knowledge to be represented declaratively through logical rules. It also covers topics like forward and backward reasoning, matching rules to facts in working memory, and using control knowledge to guide the problem solving process. Logic programming represents knowledge through Horn clauses and uses backward chaining inference to attempt to prove goals.
Deep learning is a type of machine learning that uses neural networks inspired by the human brain. It has been successfully applied to problems like image recognition, speech recognition, and natural language processing. Deep learning requires large datasets, clear goals, computing power, and neural network architectures. Popular deep learning models include convolutional neural networks and recurrent neural networks. Researchers like Geoffry Hinton and companies like Google have advanced the field through innovations that have won image recognition challenges. Deep learning will continue solving harder artificial intelligence problems by learning from massive amounts of data.
Non-monotonic reasoning allows conclusions to be retracted when new information is introduced. It is used to model plausible reasoning where defaults may be overridden. For example, it is typically true that birds fly, so we could conclude that Tweety flies since Tweety is a bird. However, if we are later told Tweety is a penguin, we would retract the conclusion that Tweety flies since penguins do not fly despite being birds. Non-monotonic reasoning resolves inconsistencies by removing conclusions derived from default rules when specific countervailing information is received.
Knowledge Representation in Artificial intelligence Yasir Khan
This document discusses different methods of knowledge representation in artificial intelligence, including logical representations, semantic networks, production rules, and frames. Logical representations use formal logics like propositional logic and first-order predicate logic to represent facts and relationships. Semantic networks represent knowledge graphically as nodes and edges to model concepts and their relationships. Production rules represent knowledge as condition-action pairs to model problem-solving. Frames represent stereotyped situations as templates with slots to model attributes and behaviors. Choosing the right knowledge representation method is important for building successful AI systems.
The document discusses capsule networks, a type of neural network proposed by Geoff Hinton in 2017 as an alternative to convolutional neural networks (CNNs) for computer vision tasks. Capsule networks aim to address some limitations of CNNs, such as their inability to capture spatial relationships and pose information. The key concepts discussed include dynamic routing between capsules, which allows for parts-based representation, and equivariance, where capsules can learn transformation properties like position and orientation. The document provides an overview of a capsule network architecture and routing algorithm proposed in a 2017 paper by Sabour et al.
An introduction to quantum machine learning.pptxColleen Farrelly
Very basic introduction to quantum computing given at Indaba Malawi 2022. Overviews some basic hardware in classical and quantum computing, as well as a few quantum machine learning algorithms in use today. Resources for self-study provided.
The subject of this study is to show the application of fuzzy logic in image processing with a brief introduction to fuzzy logic and digital image processing.
1. Planning involves finding a sequence of actions that achieves a goal starting from an initial state. It uses a set of operators that define the possible actions and their effects.
2. A plan is a sequence of operator instances that transforms the initial state into a goal state. Classical planning assumes fully observable, deterministic environments.
3. Planning problems can be represented using a logical language that describes states, goals, actions and their preconditions and effects. This representation allows planning algorithms to operate over problems.
The document discusses pattern recognition including defining a pattern and pattern class, examples of pattern recognition applications, and the statistical and machine learning approaches used. It provides details on the human and machine perception of patterns and the typical pattern recognition process of data acquisition, preprocessing, feature extraction, classification, and post processing. It also presents a case study on using pattern recognition for fish classification to sort sea bass and salmon.
Computer vision is the goal of writing programs that can interpret images, such as video sequences or medical scans. It involves acquiring images, preprocessing them, extracting features, detecting/segmenting objects, and recognizing/interpreting the images. Computer vision draws from fields like calculus, linear algebra, and statistics. It has applications in areas like robotics, navigation, inspection, and medical imaging. While computer vision has improved, it still lacks the subtlety and versatility of human vision.
Neural networks are inspired by biological neural networks and are composed of interconnected processing elements called neurons. Neural networks can learn complex patterns and relationships through a learning process without being explicitly programmed. They are widely used for applications like pattern recognition, classification, forecasting and more. The document discusses neural network concepts like architecture, learning methods, activation functions and applications. It provides examples of biological and artificial neurons and compares their characteristics.
The document provides an introduction to artificial intelligence, outlining key concepts such as characteristics of AI programs, categories of AI systems, foundations of AI in areas like mathematics and neuroscience, applications of AI, and sub-areas within the field such as game playing, machine learning, and computer vision.
This document discusses object recognition by computers. It notes that while object recognition is easy for humans, it is difficult for computers because they cannot rely on appearance alone. Key challenges for computers include variations in scale, shape, occlusion, lighting and background clutter. The document then discusses techniques used for object recognition, including feature detection methods like SIFT and SURF that extract keypoints, descriptors that describe regions around keypoints, and feature matching to identify corresponding regions between images. It also covers bag-of-words models, visual vocabularies and inverted indexing to allow large scale image retrieval. Finally, it lists applications of object recognition like digital watermarking, face detection and robot navigation.
Knowledge representation is a field of artificial intelligence that represents information about the world in a way that a computer system can understand to perform complex tasks. It simplifies complex systems through modeling human psychology and problem-solving. Examples of knowledge representation include semantic nets, frames, rules, and ontologies. Knowledge representation allows for automated reasoning about represented knowledge and asserting new knowledge. While first-order logic provides powerful and compact representation, it lacks ease of use and practical implementation for real-world problems. Effective knowledge representation requires balancing expressive power with practical considerations like execution efficiency.
Residual neural networks (ResNets) solve the vanishing gradient problem through shortcut connections that allow gradients to flow directly through the network. The ResNet architecture consists of repeating blocks with convolutional layers and shortcut connections. These connections perform identity mappings and add the outputs of the convolutional layers to the shortcut connection. This helps networks converge earlier and increases accuracy. Variants include basic blocks with two convolutional layers and bottleneck blocks with three layers. Parameters like number of layers affect ResNet performance, with deeper networks showing improved accuracy. YOLO is a variant that replaces the softmax layer with a 1x1 convolutional layer and logistic function for multi-label classification.
COM2304: Introduction to Computer Vision & Image Processing Hemantha Kulathilake
At the end of this lesson, you should be able to;
Describe image.
Describe digital image processing and computer vision.
Compare and Contrast image processing and computer vision.
Describe image sources.
Identify the array of imaging application under the EM Image source.
Describe the components of Image processing system and computer vision system.
Deep learning uses neural networks, which are systems inspired by the human brain. Neural networks learn patterns from large amounts of data through forward and backpropagation. They are constructed of layers including an input layer, hidden layers, and an output layer. Deep learning can learn very complex patterns and has various applications including image classification, machine translation, and more. Recurrent neural networks are useful for sequential data like text and audio. Convolutional neural networks are widely used in computer vision tasks.
Introduction to artificial intelligence lecture 1REHAN IJAZ
This document provides an introduction to artificial intelligence. It defines intelligence as the ability to solve problems, think, plan, learn, recognize patterns, and handle ambiguous situations. The document then asks if machines can exhibit these intelligent behaviors such as searching meshes, solving sequences, developing plans, diagnosing issues, answering questions, recognizing fingerprints, understanding concepts, and perceiving the world. It states that the goal of AI is to create systems that can learn, think, perceive, analyze and act like humans. Early work in AI included the development of programs that used logic to solve problems like humans. Current areas of AI research include computer vision, natural language processing, expert systems, robotics, and creating human-like robots.
This is the slideshow for a presentation I gave as part of my graduate coursework at the Institute for Innovation and Public Purpose at University College London (UCL IIPP). Drawing on the work of IIPP professors including Carlota Perez (techno-economic paradigms), Mariana Mazzucato (“The Entrepreneurial State”), and Tim O’Reilly, I evaluate the innovation trajectory of Deep Neural Networks as a method of machine learning. I trace the history of machine learning to its present-day and conclude that while Deep Neural Networks have not yet reached technological maturity, they are already starting to encounter barriers to exponential growth and innovation. These slides were designed to be read independently from the spoken portion. If you found this useful or interesting, please message me on LinkedIn! - Justin Beirold
Performance analysis of automated brain tumor detection from MR imaging and CT scan using basic image processing techniques based on various hard and soft computing has been performed in our work. Moreover, we applied six traditional classifiers to detect brain tumor in the images. Then we applied CNN for brain tumor detection to include deep learning method in our work. We compared the result of the traditional one having the best accuracy (SVM) with the result of CNN. Furthermore, our work presents a generic method of tumor detection and extraction of its various features.
The document discusses ontology engineering and provides details about:
1. Ontology engineering is the process of developing ontologies for a particular domain by defining concepts, arranging them hierarchically, and defining their properties and relationships.
2. Ontology engineering is analogous to object-oriented database design but ontologies reflect the structure of the world using open world assumptions.
3. Popular ontology engineering tools include Protégé, which supports ontology development and knowledge modeling.
1) This document discusses semantic networks, which are a knowledge representation technique used in artificial intelligence. Semantic networks represent knowledge through nodes and links, where nodes represent concepts or objects, and links represent relationships between the nodes.
2) As an example, a simple semantic network is presented representing facts about a cat named Jerry - that Jerry is a cat, a mammal, owned by Jay, white in color, and likes cheese.
3) The document outlines different types of semantic networks including definitional, assertional, implicational, and learning networks. It also discusses advantages such as being a natural representation of knowledge, and disadvantages including lack of quantifiers and lack of intelligence.
human activity recognization using machine learning with data analysisVenkat Projects
Human activity recognition, or HAR for short, is a broad field of study concerned with identifying the specific movement or action of a person based on sensor data.
The sensor data may be remotely recorded, such as video, radar, or other wireless methods. It contains data generated from accelerometer, gyroscope and other sensors of Smart phone to train supervised predictive models using machine learning techniques like SVM , Random forest and decision tree to generate a model. Which can be used to predict the kind of movement being carried out by the person, which is divided into six categories walking, walking upstairs, walking down-stairs, sitting, standing and laying?
MLM and SVM achieved accuracy of more than 99.2% in the original data set and 98.1% using new feature selection method. Results show that the proposed feature selection approach is a promising alternative to activity recognition on smart phones.
This document provides a summary of artificial intelligence including definitions, history, and whether computers can perform certain tasks. It discusses four approaches to defining AI: (1) thinking like humans through cognitive science, (2) thinking rationally using logic, (3) acting like humans as in the Turing test, and (4) acting rationally to achieve the best outcomes. The document also summarizes key events in the history of AI and whether computers can beat humans at games, recognize speech, understand language, learn, see, plan, and more.
The document provides an overview of artificial intelligence (AI), including its main areas of study, progress made, applications, and ongoing challenges. It discusses how AI involves automated perception, learning, reasoning and planning. While recognition and learning have advanced, planning and general reasoning remain challenging. The document outlines applications in industries like finance, medicine and transportation, but notes that many problems remain unsolved, making AI an active area of research.
An introduction to quantum machine learning.pptxColleen Farrelly
Very basic introduction to quantum computing given at Indaba Malawi 2022. Overviews some basic hardware in classical and quantum computing, as well as a few quantum machine learning algorithms in use today. Resources for self-study provided.
The subject of this study is to show the application of fuzzy logic in image processing with a brief introduction to fuzzy logic and digital image processing.
1. Planning involves finding a sequence of actions that achieves a goal starting from an initial state. It uses a set of operators that define the possible actions and their effects.
2. A plan is a sequence of operator instances that transforms the initial state into a goal state. Classical planning assumes fully observable, deterministic environments.
3. Planning problems can be represented using a logical language that describes states, goals, actions and their preconditions and effects. This representation allows planning algorithms to operate over problems.
The document discusses pattern recognition including defining a pattern and pattern class, examples of pattern recognition applications, and the statistical and machine learning approaches used. It provides details on the human and machine perception of patterns and the typical pattern recognition process of data acquisition, preprocessing, feature extraction, classification, and post processing. It also presents a case study on using pattern recognition for fish classification to sort sea bass and salmon.
Computer vision is the goal of writing programs that can interpret images, such as video sequences or medical scans. It involves acquiring images, preprocessing them, extracting features, detecting/segmenting objects, and recognizing/interpreting the images. Computer vision draws from fields like calculus, linear algebra, and statistics. It has applications in areas like robotics, navigation, inspection, and medical imaging. While computer vision has improved, it still lacks the subtlety and versatility of human vision.
Neural networks are inspired by biological neural networks and are composed of interconnected processing elements called neurons. Neural networks can learn complex patterns and relationships through a learning process without being explicitly programmed. They are widely used for applications like pattern recognition, classification, forecasting and more. The document discusses neural network concepts like architecture, learning methods, activation functions and applications. It provides examples of biological and artificial neurons and compares their characteristics.
The document provides an introduction to artificial intelligence, outlining key concepts such as characteristics of AI programs, categories of AI systems, foundations of AI in areas like mathematics and neuroscience, applications of AI, and sub-areas within the field such as game playing, machine learning, and computer vision.
This document discusses object recognition by computers. It notes that while object recognition is easy for humans, it is difficult for computers because they cannot rely on appearance alone. Key challenges for computers include variations in scale, shape, occlusion, lighting and background clutter. The document then discusses techniques used for object recognition, including feature detection methods like SIFT and SURF that extract keypoints, descriptors that describe regions around keypoints, and feature matching to identify corresponding regions between images. It also covers bag-of-words models, visual vocabularies and inverted indexing to allow large scale image retrieval. Finally, it lists applications of object recognition like digital watermarking, face detection and robot navigation.
Knowledge representation is a field of artificial intelligence that represents information about the world in a way that a computer system can understand to perform complex tasks. It simplifies complex systems through modeling human psychology and problem-solving. Examples of knowledge representation include semantic nets, frames, rules, and ontologies. Knowledge representation allows for automated reasoning about represented knowledge and asserting new knowledge. While first-order logic provides powerful and compact representation, it lacks ease of use and practical implementation for real-world problems. Effective knowledge representation requires balancing expressive power with practical considerations like execution efficiency.
Residual neural networks (ResNets) solve the vanishing gradient problem through shortcut connections that allow gradients to flow directly through the network. The ResNet architecture consists of repeating blocks with convolutional layers and shortcut connections. These connections perform identity mappings and add the outputs of the convolutional layers to the shortcut connection. This helps networks converge earlier and increases accuracy. Variants include basic blocks with two convolutional layers and bottleneck blocks with three layers. Parameters like number of layers affect ResNet performance, with deeper networks showing improved accuracy. YOLO is a variant that replaces the softmax layer with a 1x1 convolutional layer and logistic function for multi-label classification.
COM2304: Introduction to Computer Vision & Image Processing Hemantha Kulathilake
At the end of this lesson, you should be able to;
Describe image.
Describe digital image processing and computer vision.
Compare and Contrast image processing and computer vision.
Describe image sources.
Identify the array of imaging application under the EM Image source.
Describe the components of Image processing system and computer vision system.
Deep learning uses neural networks, which are systems inspired by the human brain. Neural networks learn patterns from large amounts of data through forward and backpropagation. They are constructed of layers including an input layer, hidden layers, and an output layer. Deep learning can learn very complex patterns and has various applications including image classification, machine translation, and more. Recurrent neural networks are useful for sequential data like text and audio. Convolutional neural networks are widely used in computer vision tasks.
Introduction to artificial intelligence lecture 1REHAN IJAZ
This document provides an introduction to artificial intelligence. It defines intelligence as the ability to solve problems, think, plan, learn, recognize patterns, and handle ambiguous situations. The document then asks if machines can exhibit these intelligent behaviors such as searching meshes, solving sequences, developing plans, diagnosing issues, answering questions, recognizing fingerprints, understanding concepts, and perceiving the world. It states that the goal of AI is to create systems that can learn, think, perceive, analyze and act like humans. Early work in AI included the development of programs that used logic to solve problems like humans. Current areas of AI research include computer vision, natural language processing, expert systems, robotics, and creating human-like robots.
This is the slideshow for a presentation I gave as part of my graduate coursework at the Institute for Innovation and Public Purpose at University College London (UCL IIPP). Drawing on the work of IIPP professors including Carlota Perez (techno-economic paradigms), Mariana Mazzucato (“The Entrepreneurial State”), and Tim O’Reilly, I evaluate the innovation trajectory of Deep Neural Networks as a method of machine learning. I trace the history of machine learning to its present-day and conclude that while Deep Neural Networks have not yet reached technological maturity, they are already starting to encounter barriers to exponential growth and innovation. These slides were designed to be read independently from the spoken portion. If you found this useful or interesting, please message me on LinkedIn! - Justin Beirold
Performance analysis of automated brain tumor detection from MR imaging and CT scan using basic image processing techniques based on various hard and soft computing has been performed in our work. Moreover, we applied six traditional classifiers to detect brain tumor in the images. Then we applied CNN for brain tumor detection to include deep learning method in our work. We compared the result of the traditional one having the best accuracy (SVM) with the result of CNN. Furthermore, our work presents a generic method of tumor detection and extraction of its various features.
The document discusses ontology engineering and provides details about:
1. Ontology engineering is the process of developing ontologies for a particular domain by defining concepts, arranging them hierarchically, and defining their properties and relationships.
2. Ontology engineering is analogous to object-oriented database design but ontologies reflect the structure of the world using open world assumptions.
3. Popular ontology engineering tools include Protégé, which supports ontology development and knowledge modeling.
1) This document discusses semantic networks, which are a knowledge representation technique used in artificial intelligence. Semantic networks represent knowledge through nodes and links, where nodes represent concepts or objects, and links represent relationships between the nodes.
2) As an example, a simple semantic network is presented representing facts about a cat named Jerry - that Jerry is a cat, a mammal, owned by Jay, white in color, and likes cheese.
3) The document outlines different types of semantic networks including definitional, assertional, implicational, and learning networks. It also discusses advantages such as being a natural representation of knowledge, and disadvantages including lack of quantifiers and lack of intelligence.
human activity recognization using machine learning with data analysisVenkat Projects
Human activity recognition, or HAR for short, is a broad field of study concerned with identifying the specific movement or action of a person based on sensor data.
The sensor data may be remotely recorded, such as video, radar, or other wireless methods. It contains data generated from accelerometer, gyroscope and other sensors of Smart phone to train supervised predictive models using machine learning techniques like SVM , Random forest and decision tree to generate a model. Which can be used to predict the kind of movement being carried out by the person, which is divided into six categories walking, walking upstairs, walking down-stairs, sitting, standing and laying?
MLM and SVM achieved accuracy of more than 99.2% in the original data set and 98.1% using new feature selection method. Results show that the proposed feature selection approach is a promising alternative to activity recognition on smart phones.
This document provides a summary of artificial intelligence including definitions, history, and whether computers can perform certain tasks. It discusses four approaches to defining AI: (1) thinking like humans through cognitive science, (2) thinking rationally using logic, (3) acting like humans as in the Turing test, and (4) acting rationally to achieve the best outcomes. The document also summarizes key events in the history of AI and whether computers can beat humans at games, recognize speech, understand language, learn, see, plan, and more.
The document provides an overview of artificial intelligence (AI), including its main areas of study, progress made, applications, and ongoing challenges. It discusses how AI involves automated perception, learning, reasoning and planning. While recognition and learning have advanced, planning and general reasoning remain challenging. The document outlines applications in industries like finance, medicine and transportation, but notes that many problems remain unsolved, making AI an active area of research.
This document provides an overview of artificial intelligence and discusses key concepts in AI search. It begins by defining an intelligent agent and its interaction with the environment. It then discusses uninformed search strategies like breadth-first search and depth-first search. It also covers iterative deepening depth-first search, uniform-cost search, searching backwards from the goal, and bidirectional search. The document aims to introduce foundational AI concepts like state spaces, actions, search trees, and strategies for traversing the problem space in an attempt to find a solution.
- The document discusses artificial intelligence, including its history, key areas such as knowledge representation and learning, and applications in areas like consumer marketing, identification technologies, predicting stock markets, and machine translation.
- While progress has been made in areas like recognition and learning, challenges remain in full natural language understanding, human-level planning and decision making. AI is being applied across many industries but remains an active area of research.
The document discusses artificial intelligence and defines it as the intelligence demonstrated by machines, in particular the ability to solve novel problems, act rationally, and act like humans. It covers the history of AI from its beginnings in 1943 to modern applications of machine learning and neural networks. While some problems like chess and math proofs have been solved, full human-level intelligence remains elusive and computers still cannot understand speech, plan optimally, or learn completely on their own without specific programming.
This document provides an overview of an artificial intelligence course, including:
- The course covers introduction to AI history and applications, knowledge representation, problem solving using search and reasoning, machine learning, robotics, and advanced AI topics.
- Required materials include an AI textbook, CLIPS programming guide, and reference books on AI structures and complex problem solving.
- The document then provides definitions and discussions of intelligence, artificial intelligence, applications of AI, and the current capabilities and limitations of AI systems.
This document discusses computational thinking and provides examples of how it can be applied. It defines computational thinking as using logical reasoning and problem-solving skills to solve problems. It gives examples of computational thinking in everyday life, sciences, archaeology, journalism, and more. The document also discusses teaching computational thinking to others using block-based programming languages like Snap, Scratch, and Pictoblox. Key concepts covered include sequences, loops, conditionals, events, parallelism, operators, and data.
Here are three possible interpretations of the phrase "Time flies like an arrow":
1. The passage of time seems to go by very quickly, in the same way that an arrow flies through the air.
2. Certain types of insects that lay their eggs on decaying matter, known as flies, move through the air in a similar way to arrows.
3. The idiom is using "flies" to refer to time passing quickly in an abstract sense, similar to an arrow moving swiftly through space.
The key challenges with natural language understanding are ambiguity and context. Even a short phrase like this one could have multiple meanings without additional context clues. Determining the intended interpretation requires commonsense reasoning abilities that computers still lack
The document discusses various topics related to artificial intelligence including definitions of AI, goals of AI, whether machines can think, the Turing test, types of AI tasks including mundane, formal and expert tasks, technologies based on AI such as machine learning, natural language processing, computer vision, and applications of AI such as in healthcare, gaming, finance, data security, social media, travel and more.
This document provides an introduction and overview of an artificial intelligence course. It outlines the following key points:
- The objectives of the course are to cover many primary AI concepts and ideas but within 15 weeks not everything can be covered.
- Today's lecture will discuss what intelligence is, a brief history of AI including modern successes like Stanley the robot, and how much progress has been made in different aspects of AI.
- The course agenda includes fuzzy logic, propositional logic and expert systems, rough set theory, decision trees, k-nearest neighbors, naive Bayes, and neural networks.
Artificial intelligence (AI) is the ability of machines to think and act intelligently like humans. It involves creating machines that can think and act rationally. While AI does not occur naturally, it is created by humans to enable machines to think, reason, and understand instead of just performing tasks automatically. There are still many challenges to fully achieving human-level artificial general intelligence.
This document provides an overview of an Artificial Intelligence course, including:
- The course covers topics such as strong and weak AI, knowledge representation, problem solving using search techniques, machine learning, and more.
- The learning outcomes are to understand different approaches to AI and implications for cognitive science, expand knowledge of search and learning algorithms, and understand basic planning and reasoning methods.
- Required materials include an AI textbook and reference books, as well as a programming language for AI applications.
This document provides information about the COMPSCI 270: Artificial Intelligence course at Duke University. The course will be taught in the spring of 2019 by Professor Vincent Conitzer. It will cover topics such as search, constraint satisfaction, game playing, logic, knowledge representation, and planning. Assignments will count for 30% of the grade, midterms for 40%, and a final exam for 30%. The course assumes some programming experience and background in algorithms, probability, and discrete mathematics. It aims to cover general AI techniques applied to tasks like solving Rubik's cubes, scheduling meetings, and playing games like chess.
This document provides an introduction to artificial intelligence (AI) by discussing key concepts related to intelligence and different approaches to defining AI. It examines definitions of intelligence as the ability to reason, understand, learn from experience, and plan complex tasks. AI is defined as attempting to build intelligent computer systems that exhibit human-like intelligence. There are no widely agreed upon definitions of AI, but definitions generally fall into four categories: systems that think like humans, act like humans, think rationally, or act rationally. The document also discusses the Turing test, thinking rationally using logic, knowledge representation, and intelligent agents.
The document discusses artificial intelligence and provides information on various AI topics. It includes a list of 9 NPTEL video links on topics related to unit 1 of an AI course, learning outcomes of the course, definitions and descriptions of AI, areas and applications of AI, a brief history of AI, task domains and techniques in AI, and examples of search problems and search methods. Depth-first search is described as a method that exhaustively explores branches in a search tree to the maximum depth until a solution is found.
CS6659 Artificial Intelligence
Slides in the features of Artificial Intelligence, Definition of Artificial Intelligence
Can be used by undergraduate students
Artificial intelligence and machine learning are discussed. AI is defined as making computers intelligent like humans through understanding, reasoning, planning, communication and perception. Machine learning is a subset of AI that allows machines to learn from experience without being explicitly programmed. The document provides background on AI and ML, including definitions, history, and discussions of intelligence and applications.
This document provides an overview of a course on trends and research applications in natural language processing (NLP). It begins with introducing the goals of the course, which are to understand interesting NLP tasks and novel projects through a research-oriented webinar. The document then covers various NLP topics like question answering, machine translation, sentiment analysis, natural language generation applications, and challenges in NLP like grounded language and embodied language. It also provides tips for aspiring NLP researchers.
This document provides an introduction to artificial intelligence (AI) and related concepts. It defines key terms like data, information, knowledge and knowledge base. It discusses different views on defining AI, including thinking humanly, thinking rationally, acting humanly, and acting rationally. The document also covers the history of AI, applications of AI, knowledge representation, and the differences between human and artificial intelligence. It provides examples of expert systems like DENDRAL, MYCIN, PUFF and ELIZA to illustrate the concept.
The document provides an introduction to unsupervised learning and reinforcement learning. It then discusses eigen values and eigen vectors, showing how to calculate them from a matrix. It provides examples of covariance matrices and using Gaussian elimination to solve for eigen vectors. Finally, it discusses principal component analysis and different clustering algorithms like K-means clustering.
Cross validation is a technique for evaluating machine learning models by splitting the dataset into training and validation sets and training the model multiple times on different splits, to reduce variance. K-fold cross validation splits the data into k equally sized folds, where each fold is used once for validation while the remaining k-1 folds are used for training. Leave-one-out cross validation uses a single observation from the dataset as the validation set. Stratified k-fold cross validation ensures each fold has the same class proportions as the full dataset. Grid search evaluates all combinations of hyperparameters specified as a grid, while randomized search samples hyperparameters randomly within specified ranges. Learning curves show training and validation performance as a function of training set size and can diagnose underfitting
This document provides an overview of supervised machine learning algorithms for classification, including logistic regression, k-nearest neighbors (KNN), support vector machines (SVM), and decision trees. It discusses key concepts like evaluation metrics, performance measures, and use cases. For logistic regression, it covers the mathematics behind maximum likelihood estimation and gradient descent. For KNN, it explains the algorithm and discusses distance metrics and a numerical example. For SVM, it outlines the concept of finding the optimal hyperplane that maximizes the margin between classes.
The document provides information on solving the sum of subsets problem using backtracking. It discusses two formulations - one where solutions are represented by tuples indicating which numbers are included, and another where each position indicates if the corresponding number is included or not. It shows the state space tree that represents all possible solutions for each formulation. The tree is traversed depth-first to find all solutions where the sum of the included numbers equals the target sum. Pruning techniques are used to avoid exploring non-promising paths.
The document discusses the greedy method and its applications. It begins by defining the greedy approach for optimization problems, noting that greedy algorithms make locally optimal choices at each step in hopes of finding a global optimum. Some applications of the greedy method include the knapsack problem, minimum spanning trees using Kruskal's and Prim's algorithms, job sequencing with deadlines, and finding the shortest path using Dijkstra's algorithm. The document then focuses on explaining the fractional knapsack problem and providing a step-by-step example of solving it using a greedy approach. It also provides examples and explanations of Kruskal's algorithm for finding minimum spanning trees.
The document describes various divide and conquer algorithms including binary search, merge sort, quicksort, and finding maximum and minimum elements. It begins by explaining the general divide and conquer approach of dividing a problem into smaller subproblems, solving the subproblems independently, and combining the solutions. Several examples are then provided with pseudocode and analysis of their divide and conquer implementations. Key algorithms covered in the document include binary search (log n time), merge sort (n log n time), and quicksort (n log n time on average).
What is an Algorithm
Time Complexity
Space Complexity
Asymptotic Notations
Recursive Analysis
Selection Sort
Insertion Sort
Recurrences
Substitution Method
Master Tree Method
Recursion Tree Method
This document provides an outline for a machine learning syllabus. It includes 14 modules covering topics like machine learning terminology, supervised and unsupervised learning algorithms, optimization techniques, and projects. It lists software and hardware requirements for the course. It also discusses machine learning applications, issues, and the steps to build a machine learning model.
The document discusses problem-solving agents and their approach to solving problems. Problem-solving agents (1) formulate a goal based on the current situation, (2) formulate the problem by defining relevant states and actions, and (3) search for a solution by exploring sequences of actions that lead to the goal state. Several examples of problems are provided, including the 8-puzzle, robotic assembly, the 8 queens problem, and the missionaries and cannibals problem. For each problem, the relevant states, actions, goal tests, and path costs are defined.
The simplex method is a linear programming algorithm that can solve problems with more than two decision variables. It works by generating a series of solutions, called tableaus, where each tableau corresponds to a corner point of the feasible solution space. The algorithm starts at the initial tableau, which corresponds to the origin. It then shifts to adjacent corner points, moving in the direction that optimizes the objective function. This process of generating new tableaus continues until an optimal solution is found.
The document discusses functions and the pigeonhole principle. It defines what a function is, how functions can be represented graphically and with tables and ordered pairs. It covers one-to-one, onto, and bijective functions. It also discusses function composition, inverse functions, and the identity function. The pigeonhole principle states that if n objects are put into m containers where n > m, then at least one container must hold more than one object. Examples are given to illustrate how to apply the principle to problems involving months, socks, and selecting numbers.
The document discusses relations and their representations. It defines a binary relation as a subset of A×B where A and B are nonempty sets. Relations can be represented using arrow diagrams, directed graphs, and zero-one matrices. A directed graph represents the elements of A as vertices and draws an edge from vertex a to b if aRb. The zero-one matrix representation assigns 1 to the entry in row a and column b if (a,b) is in the relation, and 0 otherwise. The document also discusses indegrees, outdegrees, composite relations, and properties of relations like reflexivity.
This document discusses logic and propositional logic. It covers the following topics:
- The history and applications of logic.
- Different types of statements and their grammar.
- Propositional logic including symbols, connectives, truth tables, and semantics.
- Quantifiers, universal and existential quantification, and properties of quantifiers.
- Normal forms such as disjunctive normal form and conjunctive normal form.
- Inference rules and the principle of mathematical induction, illustrated with examples.
1. Set theory is an important mathematical concept and tool that is used in many areas including programming, real-world applications, and computer science problems.
2. The document introduces some basic concepts of set theory including sets, members, operations on sets like union and intersection, and relationships between sets like subsets and complements.
3. Infinite sets are discussed as well as different types of infinite sets including countably infinite and uncountably infinite sets. Special sets like the empty set and power sets are also covered.
The document discusses uncertainty and probabilistic reasoning. It describes sources of uncertainty like partial information, unreliable information, and conflicting information from multiple sources. It then discusses representing and reasoning with uncertainty using techniques like default logic, rules with probabilities, and probability theory. The key approaches covered are conditional probability, independence, conditional independence, and using Bayes' rule to update probabilities based on new evidence.
Planning involves representing an initial state, possible actions, and a goal state. A planning agent uses a knowledge base to select action sequences that transform the initial state into a goal state. STRIPS is a common planning representation that uses predicates to describe states and logical operators to represent actions and their effects. A STRIPS planning problem specifies the initial state, goal conditions, and set of operators. A solution is a sequence of ground operator instances that produces the goal state from the initial state.
6th International Conference on Machine Learning & Applications (CMLA 2024)ClaraZara1
6th International Conference on Machine Learning & Applications (CMLA 2024) will provide an excellent international forum for sharing knowledge and results in theory, methodology and applications of on Machine Learning & Applications.
A SYSTEMATIC RISK ASSESSMENT APPROACH FOR SECURING THE SMART IRRIGATION SYSTEMSIJNSA Journal
The smart irrigation system represents an innovative approach to optimize water usage in agricultural and landscaping practices. The integration of cutting-edge technologies, including sensors, actuators, and data analysis, empowers this system to provide accurate monitoring and control of irrigation processes by leveraging real-time environmental conditions. The main objective of a smart irrigation system is to optimize water efficiency, minimize expenses, and foster the adoption of sustainable water management methods. This paper conducts a systematic risk assessment by exploring the key components/assets and their functionalities in the smart irrigation system. The crucial role of sensors in gathering data on soil moisture, weather patterns, and plant well-being is emphasized in this system. These sensors enable intelligent decision-making in irrigation scheduling and water distribution, leading to enhanced water efficiency and sustainable water management practices. Actuators enable automated control of irrigation devices, ensuring precise and targeted water delivery to plants. Additionally, the paper addresses the potential threat and vulnerabilities associated with smart irrigation systems. It discusses limitations of the system, such as power constraints and computational capabilities, and calculates the potential security risks. The paper suggests possible risk treatment methods for effective secure system operation. In conclusion, the paper emphasizes the significant benefits of implementing smart irrigation systems, including improved water conservation, increased crop yield, and reduced environmental impact. Additionally, based on the security analysis conducted, the paper recommends the implementation of countermeasures and security approaches to address vulnerabilities and ensure the integrity and reliability of the system. By incorporating these measures, smart irrigation technology can revolutionize water management practices in agriculture, promoting sustainability, resource efficiency, and safeguarding against potential security threats.
Presentation of IEEE Slovenia CIS (Computational Intelligence Society) Chapte...University of Maribor
Slides from talk presenting:
Aleš Zamuda: Presentation of IEEE Slovenia CIS (Computational Intelligence Society) Chapter and Networking.
Presentation at IcETRAN 2024 session:
"Inter-Society Networking Panel GRSS/MTT-S/CIS
Panel Session: Promoting Connection and Cooperation"
IEEE Slovenia GRSS
IEEE Serbia and Montenegro MTT-S
IEEE Slovenia CIS
11TH INTERNATIONAL CONFERENCE ON ELECTRICAL, ELECTRONIC AND COMPUTING ENGINEERING
3-6 June 2024, Niš, Serbia
Harnessing WebAssembly for Real-time Stateless Streaming PipelinesChristina Lin
Traditionally, dealing with real-time data pipelines has involved significant overhead, even for straightforward tasks like data transformation or masking. However, in this talk, we’ll venture into the dynamic realm of WebAssembly (WASM) and discover how it can revolutionize the creation of stateless streaming pipelines within a Kafka (Redpanda) broker. These pipelines are adept at managing low-latency, high-data-volume scenarios.
Advanced control scheme of doubly fed induction generator for wind turbine us...IJECEIAES
This paper describes a speed control device for generating electrical energy on an electricity network based on the doubly fed induction generator (DFIG) used for wind power conversion systems. At first, a double-fed induction generator model was constructed. A control law is formulated to govern the flow of energy between the stator of a DFIG and the energy network using three types of controllers: proportional integral (PI), sliding mode controller (SMC) and second order sliding mode controller (SOSMC). Their different results in terms of power reference tracking, reaction to unexpected speed fluctuations, sensitivity to perturbations, and resilience against machine parameter alterations are compared. MATLAB/Simulink was used to conduct the simulations for the preceding study. Multiple simulations have shown very satisfying results, and the investigations demonstrate the efficacy and power-enhancing capabilities of the suggested control system.
Using recycled concrete aggregates (RCA) for pavements is crucial to achieving sustainability. Implementing RCA for new pavement can minimize carbon footprint, conserve natural resources, reduce harmful emissions, and lower life cycle costs. Compared to natural aggregate (NA), RCA pavement has fewer comprehensive studies and sustainability assessments.
2. Objectives
1. To conceptualize the basic ideas and
techniques underlying the design of intelligent
systems.
2. To make students understand and explore the
mechanism of mind that enable intelligent
thought and action.
3. To make students understand advanced
representation formalism and search
techniques.
4. To make students understand how to deal with
uncertain and incomplete information.shiwani gupta 2
3. Outcomes
Learner will be able to
1. develop a basic understanding of AI building blocks
presented in intelligent agents.
2. choose an appropriate problem solving method and
knowledge representation technique.
3. analyze the strength and weaknesses of AI
approaches to knowledge – intensive problem
solving.
4. design models for reasoning with uncertainty as well
as the use of unreliable information.
5. design and develop the AI applications in real world
scenario.
shiwani gupta 3
4. Learning Outcomes
shiwani gupta 4
At the end of this course, learner should be able to:
• Knowledge and understanding
know and understand the basic concepts of Artificial
Intelligence including Search, Game Playing, KBS (including
Uncertainty), Planning and Machine Learning.
• Intellectual skills
use this knowledge and understanding of appropriate principles
and guidelines to synthesize solutions to tasks in AI and to
critically evaluate alternatives.
• Practical skills
use a well known declarative language (Prolog) and to
construct simple AI systems.
• Transferable Skills
solve problems and evaluate outcomes and alternatives.
5. List of AI Practical / Experiments
All the programs should be implemented in C/C++/Java/Prolog
under Windows or Linux environment. Experiments can also be
conducted using available open source tools.
1. One case study on NLP/Expert system based papers
published in IEEE/ACM/Springer or any prominent journal.
2. Program on uninformed and informed search methods.
3. Program on Local Search Algorithm.
4. Program on Optimization problem.
5. Program on adversarial search.
6. Program on Wumpus world.
7. Program on unification.
8. Program on Decision Tree.
Any other practical covering the syllabus topics and subtopics can
be conducted.shiwani gupta 5
6. REFERENCE BOOKS (Practicals)
1. Ivan Bratko "PROLOG Programming for Artificial Intelligence",
Pearson Education, Third Edition.
2. Elaine Rich and Kevin Knight "Artificial Intelligence "Third
Edition
3. Davis E.Goldberg, "Genetic Algorithms: Search, Optimization
and Machine Learning", Addison Wesley, N.Y., 1989.
4. Han Kamber, “Data Mining Concepts and Techniques”,
Morgann Kaufmann Publishers. Text Books:
shiwani gupta 6
7. TEXT BOOKS
1. Stuart J. Russell and Peter Norvig, "Artificial Intelligence A
Modern Approach “Second Edition" Pearson Education.
2. Saroj Kaushik “Artificial Intelligence” , Cengage Learning.
3. George F Luger “Artificial Intelligence” Low Price Edition ,
Pearson Education., Fourth edition.
shiwani gupta 7
8. REFERENCE BOOKS
1. Ivan Bratko “PROLOG Programming for Artificial Intelligence”,
Pearson Education, Third Edition.
2. Elaine Rich and Kevin Knight “Artificial Intelligence” Third
Edition
3. Davis E.Goldberg, “Genetic Algorithms: Search, Optimization
and Machine Learning”, Addison Wesley, N.Y., 1989.
4. Hagan, Demuth, Beale, “Neural Network Design” CENGAGE
Learning, India Edition.
5. Patrick Henry Winston , “Artificial Intelligence”, Addison-
Wesley, Third Edition.
6. Han Kamber, “Data Mining Concepts and Techniques”,
Morgann Kaufmann Publishers.
7. N.P.Padhy, “Artificial Intelligence and Intelligent Systems”,
Oxford University Press.
shiwani gupta 8
10. • Introduction
• History of Artificial Intelligence
• Intelligent Systems: Categorization of Intelligent
System
• Components of AI Program
• Foundations of AI
• Sub-areas of AI
• Applications of AI
• Current trends in AI
shiwani gupta 10
12. shiwani gupta 12
Can Computers beat Humans at Chess?
Chess Playing is a classic AI problem
– well-defined problem
– very complex: difficult for humans to play well
Human World Champion
Deep Blue
Deep Thought
PointsRatings
1200
1400
1600
1800
2000
2200
2400
2600
2800
3000
1966 1971 1976 1981 1986 1991 1997
Ratings
13. shiwani gupta 13
Can Computers Talk?
• This is known as “speech synthesis”
– translate text to phonetic form
• e.g., “fictitious” -> fik-tish-es
– use pronunciation rules to map phonemes to actual sound
• e.g., “tish” -> sequence of basic audio sounds
• Difficulties
– sounds made by this “lookup” approach sound unnatural
– sounds are not independent
• e.g., “act” and “action”
• modern systems can handle this pretty well
– a harder problem is emphasis, emotion, etc
• humans understand what they are saying
• machines don’t; so they sound unnatural
• Conclusion:
– NO, for complete sentences
– YES, for individual words
14. shiwani gupta 14
Can Computers Recognize Speech?
• Speech Recognition:
– mapping sounds from a microphone into a list of words
eg. A deaf human
• Recognizing single words from a small vocabulary
• systems can do this with high accuracy (order of 99%)
• e.g., directory inquiries
– limited vocabulary (area codes, city names)
– computer tries to recognize you first, if unsuccessful hands
you over to a human operator
– saves millions of dollars a year for the telephone companies
15. shiwani gupta 15
• Recognizing normal speech is much more difficult
– speech is continuous: where are the boundaries between words?
• e.g., “John’s car has a flat tire”
– large vocabularies
• can be many thousands of possible words
• we can use context to help figure out what someone said
e.g., hypothesize and test
– try telling a waiter in a restaurant:
“I would like some sugar in my coffee”
– background noise, other speakers, accents, colds, etc
– on normal speech, modern systems are only about 60-70%
accurate
• Conclusion:
– NO, normal speech is too complex to accurately recognize
– YES, for restricted problems (small vocabulary, single speaker)
16. shiwani gupta 16
Can Computers Understand speech?
• Understanding is different to recognition:
– “Where is the water?”
• assume the computer can recognize all the words
• how many different interpretations are there?
– 1. in chemistry lab, it must be pure
– 2. when you are thirsty, it must be potable
– 3. dealing with a leaky roof, it can be filthy
but how could a computer figure this out?
– clearly humans use a lot of implicit commonsense
knowledge in communication
• Conclusion: NO, much of what we say is beyond the capabilities of a
computer to understand at present
17. shiwani gupta 17
Can Computers Learn and Adapt ?
• Learning and Adaptation
– consider a computer learning to drive on the freeway
– we could teach it lots of rules about what to do and what not to do
– or we could let it drive and steer it back on course when it heads for the
embankment
• systems like this are under development (e.g., Daimler Benz)
• e.g., RALPH at CMU
– in mid 90’s it drove 98% of the way from Pittsburgh to San Diego
without any human assistance
– machine learning allows computers to learn to do things without explicit
programming
– many successful applications require some “set-up”: does not mean your
PC can learn to forecast the stock market or become a brain surgeon
• Conclusion: YES, computers can learn and adapt, when presented with
information in the appropriate way
18. shiwani gupta 18
• Recognition vs Understanding (like Speech)
– Recognition and Understanding of Objects in a scene
• look around this room
• you can effortlessly recognize objects
• human brain can map 2d visual image to 3d “map”
• Why is visual recognition a hard problem?
Conclusion:
– mostly NO: computers can only “see” certain types of objects
under limited circumstances
– YES for certain constrained problems (e.g. face recognition)
Can Computers “see”?
19. shiwani gupta 19
Can computers plan and make optimal
decisions?• Intelligence
– involves solving problems and making decisions and plans
– e.g. you want to take a holiday in Brazil
• you need to decide on dates, flights
• you need to get to the airport, etc
• involves a sequence of decisions, plans, and actions
• What makes planning hard?
– the world is not predictable:
• your flight is canceled or there’s a backup
– there are a potentially huge number of details
• do you consider all flights? all dates?
– no: commonsense constrains your solutions
– AI systems are only successful in constrained planning problems
• Conclusion: NO, real-world planning and decision-making is still beyond the
capabilities of modern computers
exception: very well-defined, constrained problems
20. shiwani gupta 20
Summary of State of AI Systems in Practice
• Speech synthesis, recognition and understanding
– very useful for limited vocabulary applications
– unconstrained speech understanding is still too hard
• Computer vision
– works for constrained problems (hand-written zip-codes)
– understanding real-world, natural scenes is still too hard
• Learning
– adaptive systems are used in many applications: have their limits
• Planning and Reasoning
– only works for constrained problems: e.g., chess
– real-world is too complex for general systems
• Overall
– many components of intelligent systems are “doable”
– there are many interesting research problems remaining
22. shiwani gupta 22
Evolution / History of AI
• The gestation of artificial intelligence (1943-1956)
• Early enthusiasm, great expectations (1952-1969)
• A dose of reality (1966-1974)
• Knowledge-based systems: The key to power? (1969-
1979)
• AI becomes an industry (1980-1988)
• The return of neural networks (1986-present)
• Recent events (1987-present)
23. shiwani gupta 23
• 1943 Warren McCulloch & Walter Pitts: Boolean circuit model of
brain
• 1949 Donald Hebb discovered how to set connection strengths b/w
neurons
• 1950 Turing's "Computing Machinery and Intelligence“- Turing Test
(the imitation game)
• 1950 Marvin Minsky and Dean Edmonds built first neural net
computer (Snarc)
• 1950s Early AI programs, including Samuel's checkers program,
Newell & Simon's Logic Theorist, Gelernter's Geometry
Engine
• 1952 Arthur Samuel-checkers programs; learned how to improve,
quickly eclipsing Samuel himself
• 1952-69 Look, Ma, no hands! (first toy eg. with lots of enthusiasm)
• 1956 Dartmouth workshop: "Artificial Intelligence" adopted
• 1958 LISP from MIT by John McCarthy
• 1959 Hebert Gelernter (Geometry Theorem Prover)
• 1963 James Slagle-Saint solved basic integration problems.
Evolution / History of AI
24. shiwani gupta 24
• 1963 McCarthy founds AI lab at Stanford
• 1965 Robinson's complete algorithm for logical reasoning
• 1966-74 AI discovers computational complexity
1966-74 Neural network research almost disappears after Minsky and
Papert’s book in 1969
• 1967 Daniel Bobrow-Student solved algebra story problems
• 1969 DENDRAL by Buchanan et al..
• 1976 MYCIN by Edward Shortliffle in early 1970s.
• 1979 PROSPECTOR by Duda et al..
• 1980-88 Expert systems are a major industry
• 1981 Japan’s 10 year 5th generation project
• Mid 1980s Backpropogation learning algorithm reinvented
• 1985-95 Neural networks resurface connectionism turn to popularity
• 1988- Probability enters into general use
• 1988 Novel AI (ALife, Gas, soft computing,…)
• 1995- The emergence of intelligent agents as part of internet boom
• 2003- Human level AI back as topic of study
Evolution / History of AI
25. Defining AI
The branch of computer science concerned with
making computers behave like humans.
Study of agents that exist in an environment and
perceive and act.
AI strives to build intelligent entities and understand
them.
The term was coined in 1956 by John McCarthy at
the Massachusetts Institute of Technology.
26. shiwani gupta 26
"The automation of activities that we
associate with human thinking, activities such
as decision-making, problem solving, learning
..."(Bellman, 1978)
"The study of mental faculties through the
use of computational models“ (Charniak and
McDermott, 1985)
"The exciting new effort to make computers
think . . . machines with minds, in the full and
literal sense" (Haugeland, 1985)
"A field of study that seeks to explain and
emulate intelligent behavior in terms of
computational processes" (Schalkoff, 1990)
"The art of creating machines that perform
functions that require intelligence when
performed by people" (Kurzweil, 1990)
"The study of how to make computers do
things at which, at the moment, people are
better" (Rich and Knight, 1991)
"a collection of algorithms that are
computationally tractable, adequate
approximations of intractably specified
problems" (Partridge, 1991)
"the field of computer science that studies
how machines can be made to act
intelligently“ (Jackson, 1986)
"the getting of computers to do things that
seem to be intelligent" (Rowe, 1988).
"a very general investigation of the nature of
intelligence and the principles and
mechanisms required for understanding or
replicating it" (Sharpies et ai, 1989)
"a field of study that encompasses
computational techniques for performing
tasks that apparently require intelligence
when performed by humans" (Tanimoto,
1990)
"The study of the computations that make it
possible to perceive, reason, and act“
(Winston, 1992)
"The branch of computer science that is
concerned with the automation of intelligent
behavior" (Luger and Stubblefield, 1993)
"the enterprise of constructing a physical
symbol system that can reliably pass the
Turing Test" (Ginsberg, 1993)
27. shiwani gupta 27
Overview of AI
• Artificial intelligence: Computers with the ability to mimic or duplicate
the functions of the human brain
• Artificial intelligence systems: The people, procedures, hardware,
software, data, and knowledge needed to develop computer systems
and machines that demonstrate the characteristics of intelligence
• Intelligent behavior
– Learn from experience
– Apply knowledge acquired from experience
– Handle complex situations
– Solve problems when important information is missing
– Determine what is important
– React quickly and correctly to a new situation
– Understand visual images
– Process and manipulate symbols
– Be creative and imaginative
– Use heuristics
28. shiwani gupta 28
Major Branches of AI
Perceptive system
• A system that approximates the way a human sees, hears, and feels objects
Vision system
• Capture, store, and manipulate visual images and pictures
Robotics
• Mechanical and computer devices that perform tedious tasks with high
precision
Expert system
• Stores knowledge and makes inferences
Learning system
• Computer changes how it functions or reacts to situations based on feedback
Games playing
• Programming computers to play games such as chess and checkers
Natural language processing
• Computers understand and react to statements and commands made in a
“natural” language, such as English
Neural network
• Computer system that can act like or simulate the functioning of the human
brain
30. shiwani gupta 30
HAL: from the movie 2001
‘2001: A Space Odyssey” epic science fiction movie in
1968
part of the story centers around an intelligent computer
called HAL 9000
HAL is the “brains” of an intelligent spaceship
in the movie, HAL can
• speak easily with the crew
• see and understand the emotions of the crew
• navigate the ship automatically
• diagnose on-board problems
• make life-and-death decisions
• display emotions
In 1968 this was science fiction: is it still science fiction?
31. AI in Movies
• The Avengers: Age of Ultron (2015)
• Chappie (2015)
• Ex Machina (2015)
• Paul Blart: Mall Cop 2 (2015)
• Ash in Alien (1979)
• Bishop in Aliens (1986)
• Roy Batty in 'Blade Runner' (1982)
• WOPR in 'WarGames' (1983)
• Skynet in 'The Terminator' Series (1984 - 2015)
• Data in 'Star Trek: First Contact' (1996)
• Agent Smith in 'The Matrix' (1999)
• David in 'A.I.: Artificial Intelligence' (2001)
• Gerty in 'Moon' (2009)
• Samantha in 'Her' (2013)
shiwani gupta 31
32. shiwani gupta 32
Most common languages for AI
• LISP- HLL. Features of the language that are good for AI programming include:
garbage collection, dynamic typing, functions as data, uniform syntax, interactive
environment, and extensibility.
• PROLOG- This language wins 'cool idea' competition. It wasn't until the 70s that
people began to realize that a set of logical statements plus a general theorem
prover could make up a program. Prolog combines the high-level and traditional
advantages of Lisp with a built-in unifier, which is particularly useful in AI. Prolog
seems to be good for problems in which logic is intimately involved, or whose
solutions have a succinct logical characterization. Its major drawback is that it's hard
to learn.
• C/C++- The speed demon of the bunch, C/C++ is mostly used when the program is
simple, and execution speed is the most important. Statistical AI techniques such as
neural networks are common examples of this. Back propagation is only a couple of
pages of C/C++ code, and needs every ounce of speed that the programmer can
master.
• Java- The newcomer, Java uses several ideas from Lisp, most notably garbage
collection. Its portability makes it desirable for just about any application, and it has a
decent set of built in types. Java is still not as high-level as Lisp or Prolog, and not
as fast as C, making it best when portability is paramount.
• Python- This language does not have widespread acceptance yet, but several
people have suggested to me that it might end up passing Java soon. According to
Peter Norvig, "Python can be seen as either a practical (better libraries) version of
Scheme, or as a cleaned-up (no $@&%) version of Perl."
35. shiwani gupta 35
Views of AI
(understand and build intelligent entities)
The area of CS that focuses on creating machines that can
engage on behaviors that human consider intelligent
Thinking humanly Thinking rationally
Acting humanly Acting rationally
Modern AI focuses on acting rationally
36. Systems that think like humans:
cognitive modeling
• Humans as observed from ‘inside’
• How do we know how humans think?
– Introspection vs. psychological
experiments
• Cognitive Science
• “The exciting new effort to make computers
think … machines with minds in the full and
literal sense” (Haugeland)
• “[The automation of] activities that we
associate with human thinking, activities such
as decision-making, problem solving, learning
…” (Bellman)
37. shiwani gupta 37
Acting (doing) humanly:
“The Turing Test Approach”
• “The study of how to make computers do things at which, at the
moment, people are better.” (Rich and Knight)
• Alan Mathison Turing (1912-1954)
• A.M. Turing Award…..ACM's most prestigious technical award
is accompanied by a prize of $250,000. It is given to an
individual selected for contributions of a technical nature made
to the computing community. Financial support of the Turing
Award is provided by the Intel Corporation and Google Inc.
• Suggested major components of AI: Natural language
processing; Knowledge Representation; Automated reasoning;
Machine Learning
• Predicted that by 2000, a machine might have a 30% chance of
fooling a lay person for 5 minutes
38. shiwani gupta 38
The imitation game: “Computing machinery and intelligence”
devised by Alan Turing in 1950 defines intelligent behavior as the
ability to achieve human level performance in all cognitive tasks,
sufficient to fool an interrogator.
Test passes if human interrogator cannot distinguish AI system
from human when interrogated via a teletype (a computer
keyboard and screen) 70% of the time
Original Turing Test abstracts out physical interaction
Total Turing Test adds Computer Vision (perception) and Robotics
(manipulation)
39. shiwani gupta 39
Systems that act like humans
• natural language processing to enable it to
communicate successfully in English (or some
other human language)
• knowledge representation to store information
provided before or during the interrogation
• automated reasoning to use the stored
information to answer questions and to draw
new conclusions
• machine learning to adapt to new
circumstances and to detect and extrapolate
patterns.
40. shiwani gupta 40
ELIZA
(one of the first chatterbots in existence)
ELIZA was a computer program and an early example of primitive
natural language processing.
ELIZA operated by processing users' responses to scripts, the
most famous of which was DOCTOR.
In this mode, ELIZA mostly rephrased the user's statements as
questions and posed those to the 'patient.'
ELIZA was written by Joseph Weizenbaum between 1964 to
1966.
In DOCTOR mode, ELIZA might respond to "My head hurts" with
"Why do you say your head hurts?"
The response to "My mother hates me" would be "Who else in
your family hates you?"
ELIZA was implemented using simple pattern matching
techniques, but was taken seriously by several of its users,
even after Weizenbaum explained to them how it worked.
41. shiwani gupta 41
Reverse Turing Test
• STANDARD TURING TEST: judge is human
• REVERSE TURING TEST: judge is computer
The challenge would be for the computer to be able to determine
if it were interacting with a human or another computer.
• CAPTCHA is a form of reverse Turing test. Before being
allowed to perform some action on a website, the user is
presented with alphanumerical characters in a distorted graphic
image and asked to type them out. This is intended to prevent
automated systems from being used to abuse the site.
The rationale is that software sufficiently sophisticated to read and
reproduce the distorted image accurately does not exist (or is
not available to the average user), so any system able to do so
is likely to be a human.
42. shiwani gupta 42
Thinking rationally (ideally):
“The Laws of Thought Approach"
• Humans are not always ‘rational’
• Logic can’t express everything (e.g. uncertainty)
• Aristotle was one of the first to attempt to codify “right
thinking”, i.e., irrefutable reasoning processes.
– Given correct premises; his syllogisms gave correct
conclusions
– eg. Socrates is a man; all men are mortal. → Socrates is
mortal.
• Formal logic provides a precise notation and rules for
representing and reasoning with all kinds of things in the
world.
• What is the purpose of thinking? What thought should I have
and what thought could I have?
43. shiwani gupta 43
Laws of thought Approach emphasizes on correct inferences
Obstacles:
• it is not easy to take informal knowledge and state it in formal
terms, particularly when the knowledge is less than 100%
certain.
• there is a big difference between being able to solve a problem
"in principle" and doing so in practice. Even problems with just
a few dozen facts can exhaust the computational resources of
any computer unless it has some guidance as to which
reasoning steps to try first.
44. Systems that act rationally:
“Rational agent”
• Rational behavior: doing the right thing
• The right thing: that which is expected to
maximize goal achievement, given the available
information
• Giving answers to questions is ‘acting’.
45. Rational agents
An agent is an entity that perceives and acts
This course is about designing rational agents
For any given class of environments and
tasks, we seek the agent (or class of agents)
with the best performance
computational limitations make perfect
rationality unachievable
→ design best program for given machine
resources
46. From the above two definitions, we can see
that AI has two major roles:
– Study the intelligent part concerned with
humans.
– Represent those actions using computers.
47. shiwani gupta 47
Acting rationally (ideally):
The Rational Agent Approach
Acting so as to achieve one’s goals, given one’s beliefs (assuming
them to be correct). Does not necessarily involve thinking.
• Advantages
– More general than the “laws of thought” approach.
– More amenable to scientific development than human
behavior or human thought based approaches.
• Problems
– Always doing the right thing is not possible in complicated
environments
– Computational demands are just too high
A basic agent is just something that perceives and acts
48. shiwani gupta 48
• Cognitive skills required:
– Ability to represent knowledge and reason with it
– Ability to generate comprehensible sentences in natural
language
– Ability to perceive to get better idea of what an action might
achieve
• Rational agent: doing the right thing ( that which is expected to
achieve best expected outcome, given the available
information)
Doesn't necessarily involve thinking – e.g., blinking reflex –
but thinking should be in the service of rational action
Requires same skills as for Turing test and act even when
no provably correct way to act
Are broader in scope than previous ones.
Aristotle Every act and every enquiry, and similarly every
action and pursuit, is thought to aim at some good.
50. Prof Saroj Kaushik 50
Components of AI Program
• AI techniques must be independent of
the problem domain as far as possible.
• AI program should have
– knowledge base
– navigational capability
– inferencing
51. 51
Knowledge Base
• AI programs should be learning in
nature and update its knowledge
accordingly.
• Knowledge base consists of facts and
rules.
• Characteristics of Knowledge:
– It is voluminous in nature and requires
proper structuring
– It may be incomplete and imprecise
– It may keep on changing (dynamic)
52. 52
Navigational Capability
• Navigational capability contains
various control strategies
• Control Strategy
– determines the rule to be applied
– some heuristics (thump rule) may be
applied
55. shiwani gupta 55
AI Foundation / Pre-History
o Philosophy(428 B.C.- present)
• Logic (no formal expression)
• Aristotle first to formulate a precise set of rules of rational
derivation
• methods of reasoning… A dog is an animal, all animals
have four legs → all dogs have four legs
• The emergence of intelligence in a physical brain
• Foundations of learning language and rationality
o Mathematics(c.800- present)
• Formal representation and proof
• Main areas: logic, computation and probability
• Logic: Mathematical Formulation
• Algorithms: First Euclid’s algorithm… calculate GCD
(analyze (un)decidability and (in)tractability)
• Probability theory: uncertainty in AI
56. shiwani gupta 56
AI Foundation / Pre-History
o Economics(1776- present)
• Formal theory for rational decision making
• The concept of utility
• Decision theory (expected utility)
• Game theory (distributed models)
• Markov Models (OR)
o Neuroscience(1861- present)
• Broca study of aphasia ca → functional areas in brain
• Models for memory
• Basic model for action generation
o Psychology(1879- present)
• Behaviorism- study only objective measures of percepts
• Cognitive psychology- cognitive science- adaptation
• Reasoning- action generation and derivation
57. shiwani gupta 57
AI Foundation / Pre-History
o Computer Engg.(1940- present)
• construction of efficient computers
• Languages for efficient implementation-
FORTRAN, LISP, PROLOG, BASIC, PASCAL,
C/C++, JAVA…
o Control theory and cybernetics(1948- present)
• Computer control of physical systems
• Basis for development of robotics, vision, language
processing
o Linguistics(1957- present)
• For understanding natural languages
• Formal languages
• Syntactic and semantic analysis
• Knowledge representation
59. Subareas of AI
shiwani gupta 59
Search
Vision
Planning
Machine
Learning
Knowledge
RepresentationLogic
Expert
SystemsRoboticsNLP
60. Prof Saroj Kaushik 60
Sub-areas of AI
– Knowledge representation
– Theorem proving
– Game playing
– Reasoning dealing with uncertainty and decision making
– Learning models, inference techniques, pattern recognition,
search and matching etc.
– Logic (fuzzy, temporal, modal) in AI
– Planning and scheduling
– Natural language understanding
– Computer vision
– Understanding spoken utterances
– Intelligent tutoring systems
– Robotics
– Machine translation systems
62. shiwani gupta 62
Applications of AI
• IBM supercomputer Deep Blue defeated the reigning world chess champion
Garry Kasparov (at grandmaster level) in 1997
• Proved a mathematical conjecture (Robbins conjecture) unsolved for
decades
• No hands across America (driving autonomously 98% of the time from
Pittsburgh to San Diego)
• ALVINN, grand challenge; cars can by now drive 200 km autonomously
• During the 1991 Gulf War, US forces deployed an AI logistics planning and
scheduling program that involved up to 50,000 vehicles, cargo, and people
• NASA's on-board autonomous planning program controlled the scheduling
of operations for a spacecraft
• Proverb solves crossword puzzles better than most humans
• MEDICAL EXPERT SYSTEM in 1980 is the first expert level performance
diagnosis of blood infections, diabetes, muscle diseases
• CHESS examines 5 billion positions per second
• ROBOTIC races in desert and urban environments by fully autonomous
vehicles; succeeded
• 2006: face recognition software available in consumer cameras
63. AI Applications
• Autonomous Planning & Scheduling:
– Telescope scheduling
– Analysis of data
– Autonomous rovers
• Medicine:
– Image analysis and enhancement
– Image guided surgery
• Robotic toys:
• Games:
• Transportation:
– Autonomous vehicle control
69. shiwani gupta 69
AI APPLICATIONS: Machine Translation
• Language problems in international business
– e.g., at a meeting of Japanese, Korean, Vietnamese and Swedish investors, no
common language
– or: you are shipping your software manuals to 127 countries
– solution; hire translators to translate
– would be much cheaper if a machine could do this
• How hard is automated translation
– very difficult! e.g., English to Russian
– “The spirit is willing but the flesh is weak” (English)
– “the vodka is good but the meat is rotten” (Russian)
– not only must the words be translated, but their meanings also!
– is this problem “AI-complete”?
• Nonetheless....
– commercial systems can do a lot of the work very well (e.g.,restricted
vocabularies in software documentation)
– algorithms which combine dictionaries, grammar models, etc.
– Recent progress using “black-box” machine learning techniques
71. Prof Saroj Kaushik 71
Latest Perception of AI
• Three typical components of AI Systems
THE WORLD
Perception Action
Reasoning
72. Prof Saroj Kaushik 72
Recent AI
• Heavy use of
– probability theory
– decision theory
– statistics
– logic (fuzzy, modal, temporal)
73. shiwani gupta 73
Intelligent Systems in Your Everyday Life
• Post Office
– automatic address recognition and sorting of mail
• Banks
– automatic check readers, signature verification systems
– automated loan application classification
• Customer Service
– automatic voice recognition
• The Web
– Identifying your age, gender, location, from your Web surfing
– Automated fraud detection
• Digital Cameras
– Automated face detection and focusing
• Computer Games
– Intelligent characters/agents
75. – more powerful and more useful computers
– new and improved interfaces
– solving new problems
– better handling of information
– relieves information overload
– conversion of information into knowledge
Some Advantages of Artificial
Intelligence
76. The Disadvantages
– increased costs
– difficulty with software development - slow
and expensive
– few experienced programmers
– few practical products have reached the
market as yet.
77. shiwani gupta 77
Question Bank
• Explain information, knowledge and intelligence
• What is AI? Explain components of AI with suitable eg. Or block
diagram.
• What do you mean by Intelligent agent? Explain various types.
State limitation of each and how is it overcome in other.
• Explain structure of intelligent agents that keep track of the
world.
• Describe environment simulator programs with performance
measure that can be used as test beds for agent programs.
• Consider vacuum cleaner problem and explain
➢ How is it rational? Which behavior will be irrational.
➢ Give success function. Explain performance measure.