ai Unit 4 Most important 40 VVi MCQs Practise || INT - 428 || Introduction to deep neural networks

本站添加: 2026-06-02

[00:00:02]Start unit number four introduction to deep neural networks and modern NLP.

[00:00:10]First question, an artificial neural network is inspired by human brain, database, operating system, compiler. So the right answer is human brain. Second, the basic processing unit of a neural network is token, neuron, parser layer.

[00:00:27]So the right answer is neuron. Third a perceptron is single layer neural network database model search algorithm for NLP model. So the right answer is single layer neural network. Four. Who proposed the perceptron? Alan turning coffee hinton. Frank Rosen Blat if John McKatti. So the right answer is Frank Rosen belt.

[00:00:55]MLP stands for multilayer perceptron machine learning processor multiple logic program a multi-linear process. So the right answer is multilayer perceptron.

[00:01:08]Six MLP contains input layer only output layer only input hidden and output layers if last option hidden layer only.

[00:01:16]So the right answer is input hidden and output layers. [clears throat] Seven.

[00:01:21]Which function introduced nonlinearity in neural networks? Activation function, sorting function, a search function, compiler function. So the right answer is activation function. Eight. Which of the following is a commonly used activation function? ReLU, SQL, BFS, FAD, DFS. So the right answer is RLU.

[00:01:44]Nine. CNN stands for central neural node, computer neural network, convolution node network, a convolutional neural network. So the right answer is convolutional neural network.

[00:02:00]10. CNN is mainly used for image processing, database design, networking, searching. So the right answer is image processing. 11. Which layer is the key component of CNN? Convolution layer, Q layer, heap layer, parser layer. So the right answer is convolution layer. 12. Pooling in CNN is used to increase dimension, reduce dimension, increase parameters, therefore increase complexity. So the right answer is reduce dimensions.

[00:02:31]13. RNN stands for recursive neural network, recurrent neural network, random neural network. Yeah. For last option, repeated neural network. So the right answer is recurrent neural network.

[00:02:46]14. RNN is mainly suitable for sequential data, images only, databases, F networks. So the right answer is sequential data. 15. Which application commonly used uses RNN? Language modeling, routing, file compression if a data storage. So the right answer is language modeling.

[00:03:08]16. The major limitation of basic RNN is large screen size, data compression, routing error. You have a vanishing gradient problem. So the right answer is vanishing gradient problem. 17. Which architecture largely replaced RNN in modern NLP? Transformers BFS CNN F decision tree. So the right answer is transformers.

[00:03:33]18. Transformer architecture was introduced in deep blue alpho. Attention is all you need if Eliza. So the right answer is attention is all you need. 19.

[00:03:44]The key idea behind transformer is attention mechanism sorting clustering a searching. So the right answer is attention mechanism.

[00:03:54]20. GPT is based on CNN decision tree perceptron a transformer architecture.

[00:04:00]So the correct answer is transformer architecture.

[00:04:05]21. But stands for birectional encoder representations from transformers.

[00:04:11]Binary encoder representation tool for birectional embedding retrieval technique. Last option basic encoder representation transformer. So the right answer is birectional encoder representation from transformers.

[00:04:26]22 but is primarily based on encoder part of transformer decoder part of transformer CNN for RNN. So the right answer is encoder part of transformer.

[00:04:39]23 GPT stands for general processing transformer, generative pre-trained transformer, graph processing technique.

[00:04:47]If your last option, general pre-trained tokenizer so the right answer is generative pre-trained transformers.

[00:04:54]24. GPT is mainly designed for image comprehension, network routing, database management, text generation. So the right answer is text generation.

[00:05:05]25. NLP stands for neural language processing, neural language programming, network language processing, if in natural logic programming. So the correct answer is natural language processing.

[00:05:19]26. Which of the following is the first step in NLP? Tokenization, translation, summarization, classification. So the correct answer is tokenization.

[00:05:30]27. Tokenization means breaking text into smaller units tokens, encrypting text, compressing text, y translating text. So the right answer is breaking text into smaller units tokens.

[00:05:45]28. In the sentence, I love AI. How many tokens are present after basic word tokenization?

[00:05:54]Option number third. Question number 29.

[00:05:57]Word embedding are used to uh represent words as numerical vector, encrypt words, remove words, sort words. So the right answer is represent words as numerical vectors. Uh question number 30. Which concept allows transformers to focus on important word in a sentence?

[00:06:16]Pooling, attention, regression, clustering. So the right answer is attention.

[00:06:23]Uh question number 31. Self attention helps a model focus on relevant word within the same sequence. Increase images resolution. Store data permanently. If last option, reduce storage. So the correct answer is focus on relevant words within the same sequence.

[00:06:42]32. Which AI application determine whether a review is positive or negative? Sentiment analysis, translation, summarization, tokenization. So the right answer is sentiment analysis. 33. This movie is amazing. Would most likely have negative sentiment, neutral sentiment, positive sentiment, mixed sentiment. So the right answer is positive sentiment. 34.

[00:07:08]Machine translation is used to convert text from one language to another, compress text, store text, a encrypt text. So the right answer is convert text from one language to another. 35.

[00:07:21]Converting English text into Hindi is an example of summarization, translation, classification if clustering. So the right answer is translation.

[00:07:31]36. Text summarization aims to increase document length. Generate a shorter meaningful version of text. Translate text encrypt text. So the right answer is generate a shorter meaningful version of text. 37. Siri, Alexa and Google Assistant are example of search engines, digital assistant, databases, in neural networks. So the correct answer is digital digital assistant.

[00:07:58]38. A chatbot is mainly designed to simulate human conversation, compress file, design hardware, you create networks. So the right answer is simulate human conversation.

[00:08:11]39. Which of the following is an application of NLP? Sentiment analysis, machine translation, text summarization, you all of the above. So the right answer is all of the above for teach is an example of generative EI model, database system, operating system, compiler. So the right answer is generative EI model. Thank you. Those unit five or six unit number five. Bye-bye.