UGC NET - Planning, natural language processing, multi-agent systems, fuzzy sets, genetic algorithms and artificial neural networks (supervised, unsupervised and reinforcement learning). Online Test (MCQ Quiz)

1. In AI planning, what is the primary goal of a planner?

A) To generate a sequence of actions to achieve a goal state from an initial state. B) To learn patterns from data without explicit guidance. C) To enable multiple agents to cooperate or compete to solve a problem. D) To represent and reason with uncertain or imprecise information.

2. Which AI planning approach involves exploring a state space using a search algorithm?

A) Constraint Satisfaction B) Heuristic Search C) Logic Programming D) Knowledge Representation

3. What does STRIPS stand for in the context of AI planning?

A) Stanford Research Institute Problem Solver B) Structured Representation for Intelligent Planning Systems C) Sequential Task Realization and Integration Protocol D) Systematic Tactical Resource Integration Planning

4. In PDDL (Planning Domain Definition Language), what represents the preconditions for an action?

A) Effects B) Actions C) Types D) Predicates

5. What is the main challenge in dealing with complex, real-world AI planning problems?

A) The combinatorial explosion of the state space. B) The lack of available data for learning. C) The difficulty in defining fuzzy rules. D) The inability of neural networks to handle sequential data.

6. Which of the following is a core component of Natural Language Processing (NLP)?

A) Image recognition B) Speech synthesis C) Robotic control D) Database management

7. What is the process of breaking down a sentence into its constituent words or tokens called in NLP?

A) Parsing B) Stemming C) Tokenization D) Lemmatization

8. Which NLP technique aims to reduce words to their root form?

A) Lemmatization B) Stop word removal C) Stemming D) Part-of-speech tagging

9. What is syntactic ambiguity in NLP?

A) A word having multiple meanings. B) A sentence having multiple possible grammatical structures. C) A sentence containing words not present in the lexicon. D) A sentence whose meaning is unclear due to context.

10. Which NLP task involves assigning a grammatical category (like noun, verb, adjective) to each word in a sentence?

A) Named Entity Recognition (NER) B) Sentiment Analysis C) Part-of-Speech (POS) Tagging D) Machine Translation

11. What is a 'multi-agent system'?

A) A system composed of a single intelligent agent. B) A system where multiple autonomous agents interact to solve problems. C) A system that processes information from a single source. D) A system designed for a single, specific task.

12. In multi-agent systems, what is the concept of 'cooperation'?

A) Agents working against each other to achieve individual goals. B) Agents coordinating their actions to achieve a common goal. C) Agents acting independently without any interaction. D) Agents attempting to deceive each other.

13. What is 'negotiation' in the context of multi-agent systems?

A) A process where agents communicate to reach an agreement. B) A process where agents compete for limited resources. C) A process where agents learn from observing others. D) A process where agents divide tasks among themselves.

14. Which type of multi-agent system involves agents with conflicting goals?

A) Cooperative systems B) Competitive systems C) Mixed-motive systems D) Hierarchical systems

15. What is a common challenge in designing multi-agent systems?

A) Ensuring individual agent rationality leads to global optimality. B) Preventing agents from learning. C) Simplifying agent communication protocols. D) Limiting the number of agents in the system.

16. Fuzzy sets are used to represent and reason with:

A) Precise and exact values. B) Uncertainty and vagueness. C) Binary logic (true/false). D) Deterministic relationships.

17. What is a 'membership function' in fuzzy set theory?

A) A function that defines the degree to which an element belongs to a fuzzy set. B) A function that maps crisp inputs to crisp outputs. C) A function that determines the probability of an event. D) A function that describes the structure of a neural network.

18. In fuzzy logic, what is a 'linguistic variable'?

A) A variable that takes precise numerical values. B) A variable whose values are words or sentences in natural language (e.g., 'hot', 'cold'). C) A variable representing a boolean condition. D) A variable used in traditional programming.

19. Which fuzzy logic operator corresponds to the logical AND operation?

A) Union B) Intersection (minimum or product) C) Complement D) Negation

20. What is the process of converting a fuzzy output set back into a single crisp value called?

A) Fuzzification B) Inference C) Defuzzification D) Normalization

21. Genetic algorithms are inspired by which biological process?

A) Cell division B) Photosynthesis C) Natural selection and evolution D) DNA replication

22. In a genetic algorithm, what is a 'chromosome' typically used to represent?

A) A single data point. B) A potential solution to the problem. C) The fitness function. D) The initial population.

23. What is the purpose of the 'fitness function' in a genetic algorithm?

A) To randomly initialize the population. B) To measure how well a solution (chromosome) solves the problem. C) To select parents for reproduction. D) To introduce mutations into the population.

24. Which genetic operator allows for the exchange of genetic material between two chromosomes?

A) Mutation B) Selection C) Crossover D) Elitism

25. What is the primary role of 'mutation' in a genetic algorithm?

A) To ensure the best solution survives to the next generation. B) To introduce new genetic material and prevent premature convergence. C) To combine genetic material from two parent chromosomes. D) To evaluate the quality of a solution.

26. Which type of learning involves training a model on labeled data?

A) Unsupervised learning B) Reinforcement learning C) Supervised learning D) Semi-supervised learning

27. In supervised learning, what is the goal of the learning process?

A) To find hidden structures in unlabeled data. B) To learn a mapping from input features to output labels. C) To learn through trial and error based on rewards and penalties. D) To reduce the dimensionality of the data.

28. Which of the following is a common algorithm for supervised learning?

A) K-Means clustering B) Principal Component Analysis (PCA) C) Linear Regression D) Apriori algorithm

29. What is 'unsupervised learning' primarily used for?

A) Predicting future outcomes based on past labeled data. B) Discovering patterns and structures in unlabeled data. C) Learning optimal actions in an environment. D) Classifying data into predefined categories.

30. Which unsupervised learning algorithm groups data points into a specified number of clusters?

A) Decision Trees B) Support Vector Machines (SVM) C) K-Means Clustering D) Naive Bayes

31. What is the fundamental concept behind 'reinforcement learning'?

A) Learning from a dataset of input-output pairs. B) Learning by exploring an environment and receiving feedback (rewards/penalties). C) Finding hidden patterns in unlabeled data. D) Optimizing a model's parameters based on a predefined objective.

32. In reinforcement learning, what does an 'agent' do?

A) Receives rewards or penalties from the environment. B) Takes actions within an environment. C) Defines the state space of the problem. D) Represents the optimal policy.

33. What is the 'state' in a reinforcement learning problem?

A) The action taken by the agent. B) The reward received by the agent. C) A description of the current situation of the environment. D) The goal to be achieved.

34. What is the 'policy' in reinforcement learning?

A) The reward function defined by the environment. B) A strategy that the agent uses to decide which action to take in a given state. C) The set of all possible states in the environment. D) The final outcome of an agent's actions.

35. Which type of artificial neural network is primarily used for image recognition tasks?

A) Recurrent Neural Network (RNN) B) Convolutional Neural Network (CNN) C) Multilayer Perceptron (MLP) D) Autoencoder

36. What is the main characteristic of a Recurrent Neural Network (RNN)?

A) It uses convolutional layers to process spatial hierarchies. B) It has connections that form directed cycles, allowing it to exhibit temporal dynamic behavior. C) It is primarily used for unsupervised feature learning. D) It processes data in parallel without memory of past inputs.

37. In an artificial neural network, what is a 'neuron' or 'node'?

A) A connection between layers. B) A processing unit that receives inputs, performs a computation, and produces an output. C) The entire network structure. D) The training data.

38. What is the role of an 'activation function' in an artificial neuron?

A) To define the learning rate of the network. B) To introduce non-linearity into the output of the neuron. C) To determine the number of layers in the network. D) To store the weights of the connections.

39. Which learning paradigm involves training a neural network using input data with corresponding correct outputs?

A) Unsupervised learning B) Reinforcement learning C) Supervised learning D) Meta-learning

40. What is the 'backpropagation' algorithm primarily used for in neural networks?

A) Initializing the network weights. B) Calculating the gradients of the loss function with respect to the network weights. C) Performing feature extraction. D) Generating new data samples.

41. An Autoencoder is a type of artificial neural network used for:

A) Generating sequences of data. B) Dimensionality reduction and feature learning (unsupervised). C) Classifying data with labeled examples. D) Controlling robotic systems.

42. What is the 'learning rate' in the context of training neural networks?

A) The speed at which the network processes data. B) The size of the step taken during gradient descent to update weights. C) The number of epochs for training. D) The complexity of the activation function.

43. Which of the following is a common issue in training deep neural networks that can be mitigated by techniques like dropout?

A) Underfitting B) Overfitting C) Slow convergence D) High bias

44. What is the 'hidden layer' in a neural network?

A) The input layer that receives raw data. B) The output layer that produces the final prediction. C) An intermediate layer between the input and output layers that performs computations. D) The layer responsible for error calculation.

45. In planning, what is a 'heuristic' function used for?

A) To guarantee finding the optimal plan. B) To estimate the cost from a given state to the goal state, guiding the search. C) To represent the effects of actions. D) To define the initial state of the problem.

46. Which NLP technique is used to identify and categorize named entities in text, such as names of people, organizations, and locations?

A) Topic Modeling B) Named Entity Recognition (NER) C) Text Summarization D) Question Answering

47. What is the 'credit assignment problem' in reinforcement learning?

A) Determining which actions led to a positive reward. B) Determining which actions led to a specific state. C) Determining which past actions are responsible for a current reward or penalty. D) Assigning credit to different agents in a multi-agent system.

48. Which type of fuzzy inference system uses rules of the form 'IF X is A AND Y is B THEN Z is C'?

A) Sugeno-type B) Mamdani-type C) Takagi-Sugeno-Kang (TSK) D) Babu-type

49. What is the primary advantage of using Genetic Algorithms for optimization problems?

A) They guarantee finding the global optimum. B) They can efficiently search large and complex solution spaces, avoiding local optima. C) They are computationally inexpensive to run. D) They require precise mathematical models of the problem.

Planning, natural language processing, multi-agent systems, fuzzy sets, genetic algorithms and artificial neural networks (supervised, unsupervised and reinforcement learning). - Online Test

Test Results