INTRODUCTION – AI Concepts, Terminology, and Application Areas
This week’s agenda focuses on foundational AI concepts, providing an in-depth look at how AI learns and its wide range of applications. By examining the processes behind AI learning and exploring its practical uses, you will develop a strong grasp of the core principles powering this groundbreaking technology.
Learning Objectives
- Understand fundamental AI concepts.
- Describe Machine Learning, Deep Learning, and Neural Networks.
- Identify and explain key application areas of AI.
GRADED: AI CONCEPTS, TERMINOLOGY, AND APPLICATION AREAS
1. Which of these statements is true?
- Cognitive systems can derive mathematically precise answers following a rigid decision tree approach
- Cognitive systems can learn from their successes and failures (CORRECT)
- Cognitive systems can only process neatly organized structured data
- Cognitive systems can only translate small volumes of audio data into their literal text translations at massive speeds
Cognitive systems continuously learn, adapt, and improve by analyzing their interactions with humans and reflecting on their successes and failures, much like the way humans learn.
2. Which of these statements is true?
- Deep Learning is a specialized subset of Machine Learning that uses layered neural networks to simulate human decision-making (CORRECT)
- Artificial Intelligence and Machine Learning refer to the same thing since both the terms are often used interchangeably
- Data Science is a subset of AI that uses machine learning algorithms to extract meaning and draw inferences from data
- AI is the subset of Data Science that uses Deep Learning algorithms on structured big data
Data Science is an independent interdisciplinary field that encompasses the entire process of collecting, processing, analyzing, and interpreting data. While it may utilize AI techniques to extract insights, it is not a subset of AI but a broader domain that integrates various methods and tools from statistics, computer science, and domain expertise.
3. Which of the following is NOT an attribute of Machine Learning?
- Takes data and rules as input and uses these inputs to develop an algorithm that will give us an answer (CORRECT)
- Machine Learning defines behavioral rules by comparing large data sets to find common patterns
- Machine Learning models can be continuously trained
- Takes data and answers as input and uses these inputs to create a set of rules that determine what the Machine Learning model will be
Machine Learning leverages computer algorithms to analyze data and make intelligent decisions by establishing behavioral patterns based on learned information, rather than relying on explicit programming. These algorithms continuously evolve and improve through ongoing learning.
4. Which of the following is NOT an attribute of Unsupervised Learning?
- It is useful for clustering data, where data is grouped according to how similar it is to its neighbors and dissimilar to everything else
- It is useful for finding hidden patterns and or groupings in data and can be used to differentiate normal behavior with outliers such as fraudulent activity
- Takes data and rules as input and uses these inputs to develop an algorithm that will give us an answer (CORRECT)
- The algorithm ingests unlabeled data, draws inferences, and finds patterns from unstructured data
This statement does not accurately describe Machine Learning or Unsupervised Learning. In Machine Learning, particularly in unsupervised learning, algorithms are not provided with predefined rules. Instead, they analyze data to identify patterns and determine rules autonomously.
5. Which of the following is an attribute of Supervised Learning?
- Relies on providing the machine learning algorithm with a set of rules and constraints and letting it learn how to achieve its goals
- Relies on providing the machine learning algorithm human-labeled data – the more samples you provide, the more precise the algorithm becomes in classifying new data (CORRECT)
- Tries its best to maximize its rewards by trying different combinations of allowed actions within the provided constraints
- Relies on providing the machine learning algorithm unlabeled data and letting the machine infer qualities
Supervised learning involves training an algorithm using human-labeled data, where each input is paired with the correct output. The accuracy of the algorithm in classifying or predicting new data improves as it is trained on a larger and more diverse set of labeled examples.
6. Which of the following statements about datasets used in Machine Learning is NOT true?
- Validation data subset is used to validate results and fine-tune the algorithm’s parameters
- Training subset is the data used to train the algorithm
- Testing data is data the model has never seen before and is used to evaluate how good the model is
- Training data is used to fine-tune algorithm’s parameters and evaluate how good the model is (CORRECT)
Training data is used to train the algorithm. It is the Validation data that is used to fine-tune algorithm’s parameters and evaluate how good the model is.
7. When creating deep learning algorithms, developers configure the number of layers and the type of functions that connect the outputs of each layer to the inputs of the next.
- True (CORRECT)
- False
Deep Learning algorithms utilize multiple layers of interconnected processing units, or neurons, where each layer processes input data and passes its output to the next layer. Developers configure the number of layers and the functions that define how the outputs of one layer are connected to the inputs of the next, enabling the system to learn complex patterns and representations.
8. Which of the following fields of application for AI can be used at the airport to flag weapons within luggage passing through the X-ray scanner?
- Computer Vision (CORRECT)
- Speech
- Chatbots
- Natural Language
Computer Vision allows machines to analyze and interpret digital images and video sequences, enabling them to perform tasks such as object detection, identification, and recognition.
9. Which of these activities is NOT required in order for a neural network to synthesize human voice?
- Continue to correct the sample and run it through the classifier, repetitively, till an accurate voice sample is created
- Deconstruct sentences to decipher the context of use (CORRECT)
- Generate audio data and run it through the network to see if it validates it as belonging to the subject
- Ingest numerous samples of a person’s voice until it can tell whether a new voice sample belongs to the same person
The process of generating natural voice begins with a neural network analyzing samples of a person’s voice to learn its unique characteristics and determine whether a new voice sample matches the same individual. A second neural network then generates audio data and iteratively tests it through the first network. This process continues until the generated voice sample is validated as accurate and consistent with the subject’s voice.
10. Which one of these ways is NOT how AI learns?
- Unsupervised Learning
- Reinforcement Learning
- Supervised Learning
- Proactive Learning (CORRECT)
AI learns in three different ways – Supervised, Unsupervised, and Reinforcement Learning.
11. Cognitive Systems can interpret data to generate hypotheses about what it means
- True (CORRECT)
- False
Cognitive systems mimic human decision-making processes by interpreting data and generating hypotheses based on the information they analyze. This allows them to make informed conclusions and predictions, much like how humans reason and draw inferences from available data.
12. Is the following an application of Machine Learning and AI:
A machine that beats human in a game in which all rules and moves have been pre-programmed into the machine – Yes or No?
- Yes
- No (CORRECT)
Programming all the rules and moves of a game is not a true application of AI. Instead, Machine Learning involves training a system to learn from data, allowing it to figure out optimal moves and strategies for winning through experience and pattern recognition, rather than relying on predefined rules.
13. Data Science is a subset of AI that uses machine learning algorithms to extract meaning and draw inferences from data.
- True
- False (CORRECT)
Data Science is an interdisciplinary field that covers the entire data processing methodology, from data collection to analysis and interpretation. While it incorporates AI techniques to derive insights from data, it is not a subset of AI but a distinct field that integrates various methods, including statistics, machine learning, and domain expertise.
14. Which of the following are attributes of Machine Learning?
- Defines behavioral rules by comparing large data sets to find common patterns (CORRECT)
- Machine learning algorithms can be continuously trained and used in the future to predict values (CORRECT)
- Takes data and answers as input and use these inputs to create a set of rules that determine what the Machine Learning model will be (CORRECT)
- In Machine Learning models, when we submit inputs, we get answers based on predefined rules
Machine Learning employs computer algorithms to analyze large datasets, identifying common patterns and defining behavioral rules based on what it has learned.
Unlike traditional algorithms, which provide answers based on a predefined set of rules, machine learning algorithms analyze input data and desired outcomes to generate their own set of rules for achieving those outcomes. The model can be continuously improved and refined by training it on new datasets.
In essence, machine learning analyzes input data and responses to determine the optimal rules and develop a learning algorithm that adapts over time.
15. Which of the following are attributes of Classification?
- Forms of classification include decision trees, support vector machines, logistic regression and random forests (CORRECT)
- Classification is the process of predicting the class of given data points (CORRECT)
- Using classification models we extract features from data and classify results into multiple categories (CORRECT)
- Classification models are built by looking at the relationships between features and results, where results are a continuous variable
Classification is the process of extracting features from data and classifying the results into one or more categories.
16. Neural networks are the reason deep learning algorithms become more efficient as the datasets increase in volume.
- True (CORRECT)
- False
Neural networks are the foundation of deep learning algorithms, enabling them to continuously learn and improve as they process more data. As datasets grow in volume over time, neural networks can adapt and enhance the quality and accuracy of their results by refining their internal models through repeated exposure to new data.
17. Which of the following are attributes of Perceptrons?
- Input layers forward the input values to the next layer by means of multiplying by a weight and summing the results (CORRECT)
- Perceptrons are single-layered neural networks consisting of input nodes connected directly to an output node (CORRECT)
- Each layer of neurons conducts a mathematical operation on the output of the previous layer
- An activation function determines how a node responds to its inputs (CORRECT)
Perceptrons are single-layer neural networks where input nodes are connected directly to an output node. The input values are forwarded to the next layer, where they are multiplied by weights and summed to produce an output. An activation function then determines how the node responds to its inputs, playing a crucial role in the performance and success of the neural network by introducing non-linearity and enabling the network to model complex patterns.
18. Which of these is the most complex data for machine learning to work with?
- Structured data
- Big data
- Natural Language (CORRECT)
- Training data
Natural language is one of the most complex forms of data for machine learning due to its highly contextual nature. Unlike other types of data, such as auditory or visual data, which often have more discernible patterns, language is used by humans conceptually rather than literally. This makes it challenging for machine learning models to understand, as meaning can change depending on context, tone, and cultural nuances.
19. Which of the following is an attribute of Natural Language Processing (NLP)?
- NLP systems can identify the emotion in which a word is spoken, for example, frustration, confusion, irritation, or fun etc. (CORRECT)
- NLP systems are provided recorded voice samples with corresponding text to help them discern common patterns
- NLP systems can understand intent (CORRECT)
- NLP systems use a broad array of linguistic models and algorithms to draw inferences from language (CORRECT)
Natural Language Processing (NLP) uses machine learning and deep learning algorithms to analyze sentences grammatically, relationally, and structurally, allowing it to understand both the semantic meaning of words and their context. NLP systems can also interpret intent and emotion by drawing inferences from a wide range of linguistic models and algorithms, enabling them to process and respond to human language more effectively.
20. In order for a self-driving vehicle to navigate accurately, it needs to piece together a complete view of its driving environment, which it does with the help of:
- Laser data (CORRECT)
- Vision data (CORRECT)
- Radar data (CORRECT)
- Data Science
Self-driving vehicles combine laser data, vision data, and radar data to generate a comprehensive three-dimensional view of their surroundings. This fusion of data allows the vehicle to accurately perceive its environment, enabling it to make informed decisions and navigate the road safely and efficiently.
21. Which is the biggest limitation of human vision that computer vision can help make up for?
- Gauging distance between objects
- Visual Attention
- Being able to predict in which direction objects are headed (CORRECT)
Visual attention is a limitation of human vision, as humans cannot focus on every element in their visual field simultaneously. Computer vision, on the other hand, can overcome this limitation by processing and analyzing large amounts of visual information at once, allowing it to detect and focus on multiple objects or details in real-time, enhancing its ability to interpret complex scenes.
CONCLUSION – AI Concepts, Terminology, and Application Areas
In conclusion, this week’s exploration of fundamental AI concepts has laid the groundwork for understanding this rapidly advancing field. By examining how AI learns and exploring its real-world applications, you’ve gained key insights into its potential and importance across various domains.
As you continue your journey into the world of artificial intelligence, keep in mind that this is just the beginning—an exciting springboard for further learning and discovery in this constantly evolving field.