Kips Class 8 AI Unit 1: Excite

Unit 1: Excite

Question and Answer
Question 1: Define the term AI.
Answer: Artificial Intelligence (AI) refers to the computer systems (including hardware and software) that are capable of doing things that normally require human intelligence. Example: recognizing images, videos, speech, and decision making etc.

Question 2: List the various behaviours associated with human intelligence.
Answer: The various behaviours associated with human intelligence are
• Planning
• Learning
• Reasoning
• Problem solving
• Perception
• Motion
• Knowledge representation
• Manipulation
• Creativity
• Social Intelligence

Question 3: Difference between common tasks and expert tasks for humans.
Answer: Common tasks
• Human beings routinely perform these tasks.
• Example
→ Identifying living and non-living things.
→ Communicating using languages, sign, symbol.
→ Moving around while avoiding or navigating obstacles.

Expert tasks
• Only human beings with special skills can do these tasks.
• Example
→ Medical diagnosis of illnesses
→ Playing complex games such as chess, checkers.
→ Mathematical modelling on problem solving

Question 4: Differentiate between common tasks and expert tasks for computers.
Answer: Common tasks
• Computers find it difficult to perform tasks that are considered routine for humans. Human beings can process and interpret complex sensory information.
• Example
→ Recognising objects, navigation etc.

Expert tasks
• Computers are able to perform some tasks quite easily.
• Example
→ Medical diagnosing; in certain cases they perform better than human beings.
→ Designing mathematical problems
→ Solving mathematical problems

Question 5: Differentiate between human intelligence and artificial intelligence.
Answer: Differences between human and artificial intelligence are
• The word 'Artificial Intelligence' itself states that it is artificial in nature. Human intelligence on the other hand is natural.
• Human decision-making is based on cognition. Artificial intelligence, on other hand, uses models for making decisions. These models try to mimic human behaviour.
• Human decision-making can be biased but the decision-making of AI will be free from bias.
• Human intelligence is general purpose intelligence i.e. capable of performing different tasks whereas artificial intelligence systems can only perform the tasks for which they are developed.

Question 6: Explain the types of AI.
Answer: There are 2 types of AI
• Artificial Narrow Intelligence (ANI) - Narrow intelligent systems are the systems that are in use today. These systems have been developed to perform simple tasks like Alexa, Google Now etc. AI systems are only designed to recognize voice and act upon it.
• Artificial General Intelligence (AGI) - Artificial General Intelligence is still at research stage. These systems will be as intelligent as humans, capable of learning and developing. Unlike ANI, AGI has the capability to learn, adapt and improve itself so that it can perform various tasks

Question 7: List any four capabilities of Artificial Narrow Intelligence.
Answer: Artificial Narrow Intelligence can perform a number of tasks such as
• Interpreting videos to understand what is happening, like a video of the roof of house for identifying repairs needed.
• Acting as a personal secretary for organising personal and business calendars, making appointments, scheduling meetings etc.
• Acting a as a customer service representative for answering simple customer questions.
• Booking hotels, ordering groceries etc.

Question 8: Differentiate between strong AI and weak AI.
Answer:
• Strong AI systems are intelligent. But weak AI systems act intelligently.
• Strong AI will become self-conscious on capable of becoming self-conscious. Weak AI can't achieve self-consciousness.
• All current AI systems are weak AI systems, i.e. they pretend to be intelligent.

Question 9: Differentiate between General AI and Narrow AI.
Answer:
• Narrow AI can handle only one task, the task for which it has been programmed. AGI can handle any intellectual task thrown at it.
• All current AI systems are narrow. The AGI systems currently exist in science fiction only.
• Many scientists have abandoned the work on AGI due to lack of process. Narrow AI field, on the other hand is blooming.

Question 10: List the contribution of AI in various fields.
Answer: The contribution of AI in various fields are
• Maths
• Biology and Neuroscience
• Sociology
• Psychology
• Computer Science
• Philosophy

Question 11: Draw a layout of dream smart house with minimum five features of AI.
Answer:

Question 12: Explain various fields related to AI.
Answer: The various fields related to AI are
• Machine Learning - Machine learning are the systems that improve their performances with more experience or information. This is sometimes regarded as a sub-field of AI and at a other times separate subfield of computer science. These systems are adaptive systems.
• Deep Learning - Deep Learning or Deep Neutral learning are systems that are capable of learning from unstructured on unlabelled data.
• Data Science - Data science is a term that is used for referring to different related sub-disciplines. These sub-disciplines include machine learning statistics, data storage web development etc.
• Robotics - The field of robotics deals with construction and programming of robots. Robots are computing machines that have both hardware and software. The artificial intelligence development is used for the software or programming part of the computer.

Question 13: Explain the applications of AI.
Answer: Some of the applications of AI are listed below
• Gaming - Game developers are using AI technology to improve the strategic aspects of games. In addition to it game developers also use AI to improve the user experience in a wide variety of games. These games learn and adapt themselves, providing better gaming experience.
• Natural Language Processing - This helps computer-based system to understand the natural language that humans use. The game Mystery Animal is based on this technology
• Expert systems - These systems are used to provide advice, explanations, solutions etc to the users. These are based on the integration of a number of different programs and sources.
• Vision system - These systems are use AI technology to understand the images. These include identification of faces, diagnosing of patients, identifying damages to building and fields etc.
• Handwriting recognition - It is related to the technology used in vision systems. In these systems, the computer converts text on the paper, or inputs given by the users from stylus into the computer readable format.
• Intelligent robots - Artificial Intelligence is widely being used to improve the working of robots so that they can perform better, able to perform tasks assigned by humans

Question 14: Explain the 3 domains of AI.
Answer: There are 3 domains of AI
• Data - Data in AI systems refers to the data that is right for the AI system in question. Data is at core of artificial intelligence. Current AI systems cannot be developed or made functional without adequate data. There are 3 types of data - Sound data, text data, image and video data.
• Computer vision - Computer Vision is a technology that is used to make machines see and percieve the human world as humans do. This requires image or visual data. The development of computer vision is progressing fast pace. The uses of computer vision are face recognition, image retrieval, gaming and controls, surveillances, and smart cars.
• Natural Language Processing - The third domain of AI is Natural Language Processing. It works on enabling the computers to understand naturally written an spoken languages like Hindi, English. This will make computer capable of natural communication. There are 2 types of NLP - Natural Language Understanding and Natural Language Generation. For AI systems, NLU is much difficult than NLG.

Question 15: Write the concept of AI behind the following games
a) Rock, paper and scissors
b) Mystery animal
c) Emoji Scavenger Hunt
d) Tic-Tac-Toe
e) Teachable Snake
f) Thing Translator
Answer:
a) Rock, paper and scissors - This game shows the power of AI and machine learning to learn adapt and respond. In this game, 2 players make one of the 3 hand shapes (paper, rock and scissors) on the count of 3. For a human player, the choices made by him will look random but they aren't. The computer searches for patterns in his data and uses them to defeat him.
b) Mystery Animal - It is a simple 20 questions game. In this game, one player pretends to be an animal. The other player can ask 20 Yes/No questions to him. This game shows the power of NLP. Google, in this game uses machine learning and knowledge graph to understand the question and reply. Mystery animal showcase the endless possibilities when open-ended natural conversation are brought into applications.
c) Emoji Scavenger Hunt - This game uses the power of a phone camera and AI to identify the real life versions of the daily used emoji. This game uses neural network to recognize and identify the images clicked by the user.
d) Tic-Tac-Toe - It is a simple and old game that is played on 3x3 grid. In this game, players take turn by placing X and O markings on the spaces in the grid. It can be won by any 8 ways. The computer calculates all the possible outcomes and uses MINIMAX function to select best of the possible outcomes.
e) Teachable Snake - Teachable Snake uses pre-trained neural networks to control classic snake game with the help of webcam images. The neural network has been trained to recognize different types of arrows. The computer system identifies the direction of the arrow on the paper and moves the snake on the screen accordingly.
f) Thing translator - This game is based on neural networks. This game speaks aloud the name of the object in the clicked image. The neutral network in the game is trained to recognize different images. The computer system identifies the image and translates it into one of the nine languages. The system works on probability of what the AI thought the object was against what it translated.

Question 16: Explain the limitations of AI.
Answer: Current AI systems are not perfect, and they suffer from a number of limitations. Open ended world times require unconstrained thinking and solutions. AI systems are constrained by a number of variables.
Following are the limitations of AI
• understanding natural language
• surfing the internet
• interpreting visuals (images and videos)
• learning natural language.