The purpose of discourse integration in natural language processing is:

to analyze the meaning of sentences in context

According to the passage under 'Architecture', what is the central component of a chatbot's architecture?

The natural language processing engine

Which of the following are types of deep learning models commonly used in natural language processing? Select all that apply.

Recurrent neural networks (RNNs)

Transformer neural networks (transformer NNs)

Convolutional neural networks (CNNs)

What is the function of the hidden layers in an RNN, according to the passage?

The hidden layers are responsible for maintaining the memory of the network.

The hidden layers are used only for input normalization.

The hidden layers are responsible for generating the output directly.

The hidden layers are used to store the weights of the network.

Refer to Figure 3: The input layer, hidden layers and output layer of a recurrent neural network (RNN). What is the primary function of an RNN?

To process variable-length sequential data by maintaining a memory of previous inputs in the form of hidden states

To process fixed-length data only

An RNN adjusts input, output and hidden layer weights during training using _________

backpropagation through time (BPTT)

What is the key innovation in Transformer NNs that differentiates them from LSTM?

The application of a self-attention mechanism.

The use of convolutional layers for sequence modeling.

The reliance on recurrent connections for memory retention.

The implementation of max pooling for feature extraction.

Which large language model uses the self-attention mechanism?

Generative Pre-trained Transformer 3 (GPT-3) created by Open AI.

Long Short-Term Memory (LSTM) network.

What is confirmation bias in the context of datasets?

Confirmation bias occurs when the dataset is biased towards a particular viewpoint, such as training data that only include customer queries related to certain types of policies.

Confirmation bias is when datasets are always balanced and contain equal representation of all viewpoints.

Confirmation bias refers to the process of randomly selecting data points to avoid any bias.

Confirmation bias is when datasets are only used for testing and not for training purposes.

What is linguistic bias in the context of datasets?

Linguistic bias occurs when the dataset is biased towards certain linguistic features, such as dialect or vocabulary, making it difficult for the model to interpret informal language.

Linguistic bias refers to the tendency of datasets to favor numerical data over textual data.

Linguistic bias is when a dataset contains only formal language and excludes all technical terms.

Linguistic bias is the process of translating datasets into multiple languages to improve accuracy.

What is sampling bias in the context of datasets?

Sampling bias occurs when the training dataset is not representative of the entire population, such as training data that only include customer queries from one demographic.

Sampling bias is when data is collected from all possible sources equally.

Sampling bias refers to the process of increasing the size of a dataset by duplicating samples.

Sampling bias is the error caused by measuring the entire population instead of a sample.

What is selection bias in the context of datasets?

Selection bias occurs when the training data are not randomly selected but are instead chosen based on some criteria, leading to biased predictions for certain demographics.

Selection bias occurs when all data points are given equal importance regardless of their source.

Selection bias is when the dataset contains only numerical values and no categorical data.

Selection bias refers to the process of increasing the size of the dataset to improve model accuracy.

List the steps for preparing a machine learning model.

Pre-processing the input data 2. Training the model 3. Deployment of the model

Collecting data 2. Ignoring outliers 3. Skipping validation

Training the model 2. Pre-processing the input data 3. Deployment of the model

Deployment of the model 2. Pre-processing the input data 3. Training the model

What does pre-processing involve?

Cleaning, selection, transformation, and reduction of data to improve its quality and accuracy.

Only collecting raw data from various sources.

Visualizing data using charts and graphs.

Storing data in databases without any modification

What specialized hardware is recommended for achieving the best performance when deploying a large language model?

CPUs (central processing units)

FPGAs (field-programmable gate arrays)

ASICs (application-specific integrated circuits)

List any two ethical challenges identified by the RAKT team.

Any two of: Data privacy and security, Bias and fairness, Accountability and responsibility, Transparency, Misinformation and manipulation.

Any two of: Cost reduction, Increased speed, Automation, Scalability, Efficiency.

Any two of: Marketing strategies, Customer engagement, Brand awareness, Product placement, Sales growth.

Any two of: Environmental sustainability, Renewable energy, Carbon footprint, Recycling, Green technology.

Which of the following is a challenge faced in natural language processing?

Evaluating methods for reducing latency

Understanding the five stages of natural language processing

Analysing the processing power requirements for a natural language processing model

Bag-of-words is a model used in natural language processing where a text is represented as an unordered collection of words, disregarding grammar and word order but keeping multiplicity.

Bag-of-words is a model that focuses on the grammatical structure of sentences and the order of words.

Bag-of-words is a technique for compressing images by representing them as a collection of pixels.

Bag-of-words is a method for sorting words in a dictionary based on their frequency of use.

Define: Biases (Confirmation, Historical, Labelling, Linguistic, Sampling, Selection)

Biases are systematic errors in data or algorithms. Types include: Confirmation (favoring information that confirms beliefs), Historical (bias from past data), Labelling (errors in data labeling), Linguistic (language-related bias), Sampling (non-representative data), Selection (biased selection of data).

Biases are always intentional manipulations in data. Types include: Confirmation (ignoring all evidence), Historical (using only future data), Labelling (never labeling data), Linguistic (using only one language), Sampling (using all data), Selection (randomly selecting data).

Biases are random errors in data collection. Types include: Confirmation (disregarding all beliefs), Historical (bias from future data), Labelling (perfect labeling), Linguistic (no language involved), Sampling (using only large samples), Selection (choosing data without criteria).

Biases are errors that occur only in algorithms. Types include: Confirmation (accepting all information), Historical (bias from current data), Labelling (labeling all data correctly), Linguistic (using technical language only), Sampling (using only small samples), Selection (selecting data alphabetically).

A dataset is a collection of data, often used for training or evaluating machine learning models.

A dataset is a type of algorithm used to process data.

A dataset is a programming language for data analysis.

A dataset is a hardware device for storing information.

Define: Deep learning

Deep learning is a subset of machine learning involving neural networks with many layers that can learn complex patterns in data.

Deep learning is a type of database management system used for storing large datasets.

Deep learning refers to shallow learning algorithms that do not use neural networks.

Deep learning is a programming language used for artificial intelligence applications.

Define: Graphical processing unit (GPU)

A graphical processing unit (GPU) is a specialized processor designed to accelerate graphics rendering and parallel computations, often used in machine learning.

A graphical processing unit (GPU) is a type of storage device used to save large files and documents.

A graphical processing unit (GPU) is a network device that manages internet connections for computers.

A graphical processing unit (GPU) is a software application used for editing text documents.

Latency is the delay between a user's action and the response from a system, often measured in milliseconds.

Latency is the total amount of data a system can process in a second.

Latency is the process of encrypting data before transmission.

Latency is the number of users a system can support simultaneously.

Define: Long short-term memory (LSTM)

Long short-term memory (LSTM) is a type of recurrent neural network architecture designed to remember information for long periods and avoid the vanishing gradient problem.

Long short-term memory (LSTM) is a type of feedforward neural network used for image classification tasks.

Long short-term memory (LSTM) is a data compression algorithm for reducing file sizes.

Long short-term memory (LSTM) is a hardware component used in computer memory management.

Define: Loss function

A loss function is a mathematical function used to measure the difference between predicted and actual values in machine learning.

A loss function is a type of neural network used for image recognition.

A loss function is a data structure for storing large datasets efficiently.

A loss function is a programming language used for statistical analysis.

Define: Memory cell state

Memory cell state refers to the internal state of a memory cell in neural networks like LSTM, which helps retain information over time.

Memory cell state is the process of deleting all stored data in a neural network.

Memory cell state refers to the external input provided to a neural network at each time step.

Memory cell state is the output layer of a neural network responsible for classification.

Define: Lexical analysis (in NLP)

Lexical analysis is the process of converting a sequence of characters into a sequence of tokens in NLP.

Lexical analysis is the process of translating tokens into syntax trees in NLP.

Lexical analysis is the process of generating speech from text in NLP.

Lexical analysis is the process of understanding the meaning of sentences in NLP.

Define: Syntactical analysis (parsing) (in NLP)

Syntactical analysis, or parsing, is the process in NLP of analyzing the grammatical structure of a sentence.

Syntactical analysis, or parsing, is the process in NLP of translating text into another language.

Syntactical analysis, or parsing, is the process in NLP of identifying the sentiment of a sentence.

Syntactical analysis, or parsing, is the process in NLP of recognizing named entities in a sentence.

Define: Recurrent neural network (RNN)

A recurrent neural network (RNN) is a type of neural network designed for processing sequential data by maintaining a memory of previous inputs.

A recurrent neural network (RNN) is a type of neural network used only for image classification tasks.

A recurrent neural network (RNN) is a neural network that does not use any memory of previous inputs.

A recurrent neural network (RNN) is a type of neural network designed exclusively for static data analysis.

IB CS Case study

12th Grade

•

53 Qs

Similar activities

ÔN TẬP GIỮA HỌC KỲ II - LỚP 11

12th Grade

•

51 Qs

KUIZ VLE - PENGETAHUAN AM

10th - 12th Grade

•

50 Qs

year3 revision

2nd Grade - University

•

50 Qs

Latihan Soal AIJ 12

12th Grade

•

57 Qs

11th COMPUTER SCIENCE CH-1 to CH-6 TN SB I-MID

11th Grade - University

•

53 Qs

software dev

9th - 12th Grade

•

50 Qs

HTML

9th - 12th Grade

•

50 Qs

pangkalan data lanjutan tingkatan 5

1st Grade - University

•

50 Qs

IB CS Case study

Quiz

•

Computers

•

12th Grade

•

Practice Problem

•

Hard

Alasdair Lightbody

FREE Resource

MULTIPLE CHOICE QUESTION

30 sec • 1 pt

What is a potential risk mentioned regarding the chatbot's ethical challenges?

Slow response time

Inability to handle complex language

Revealing personal information from its training dataset

Poor accuracy in customer queries

MULTIPLE CHOICE QUESTION

30 sec • 1 pt

What is one way to overcome dependencies in machine learning models for chatbots?

To transform unstructured text into machine-actionable information using natural language understanding (NLU).

To ignore user input and provide random responses.

To use only rule-based systems for all conversations.

To avoid using any training data for the chatbot.

Access all questions and much more by creating a free account

Create resources

Host any resource

Get auto-graded reports

Continue with Google

Continue with Email

Continue with Classlink

Continue with Clever

or continue with

Microsoft

Apple

Others

Already have an account?

Similar Resources on Wayground

50 questions

EXAM REVIEW: Computing Principles

Quiz

•

9th - 12th Grade

49 questions

PENGENDALI SISTEM ROBOTIK

Quiz

•

12th Grade

50 questions

ATS Informatika Genap

Quiz

•

10th Grade - University

55 questions

Computer Science AQA A-Level Paper 2 Practice

Quiz

•

11th - 12th Grade

50 questions

Databases

Quiz

•

KG - 12th Grade

57 questions

MTA Mobility and Devices

Quiz

•

9th - 12th Grade

55 questions

SKT5 : Bab 1 : c) Senibina Komputer

Quiz

•

10th - 12th Grade

57 questions

Ulangkaji ASK T1 Bab 2.2 Ukuran Data

Quiz

•

1st - 12th Grade

Popular Resources on Wayground

15 questions

Fractions on a Number Line

Quiz

•

3rd Grade

20 questions

Equivalent Fractions

Quiz

•

3rd Grade

25 questions

Multiplication Facts

Quiz

•

5th Grade

$fractions$

22 questions

fractions

Quiz

•

3rd Grade

20 questions

Main Idea and Details

Quiz

•

5th Grade

20 questions

Context Clues

Quiz

•

6th Grade

15 questions

Equivalent Fractions

Quiz

•

4th Grade

20 questions

Figurative Language Review

Quiz

•

6th Grade

Discover more resources for Computers

18 questions

[AP CSP] JavaScript Programming Lesson 2025-2026

Lesson

•

9th - 12th Grade

IB CS Case study

53 questions

Which of the following is NOT listed as a problem to be addressed for the chatbot in the scenario?

What does the term 'latency' refer to in the context of the chatbot's performance?

Fill in the blank: The chatbot's training dataset is not ______ enough, leading to poor accuracy in understanding and responding to customer queries.

Which problem area is described as the chatbot's language model struggling to respond appropriately to ambiguous statements?

The chatbot's architecture is too simplistic and unable to handle complex language.

What is a potential risk mentioned regarding the chatbot's ethical challenges?

What is one way to overcome dependencies in machine learning models for chatbots?

According to the passage, what is the first stage of natural language processing?

Which stage of natural language processing involves interpreting the grammar and structure of the sentence?

Fill in the blank: ________ analysis focuses on the meaning of the sentence and its individual words, such as understanding that the sentence is about the customer’s desire to make a claim related to a car accident.

Access all questions and much more by creating a free account

Similar Resources on Wayground

Popular Resources on Wayground

Discover more resources for Computers