A
AI Ethics
AI ethics is the field that examines how artificial intelligence should be designed and used responsibly. It focuses on issues like fairness, accountability, transparency, privacy, and preventing harm. Ethical frameworks help ensure AI systems benefit society while minimizing risks such as bias, surveillance, or misuse. Companies and governments are increasingly adopting AI ethics guidelines to build trust and protect users.
AI Hallucination
An AI hallucination occurs when an artificial intelligence system generates information that is incorrect, nonsensical, or entirely fabricated but presents it as fact. This is common in large language models, which may produce confident but false answers. Addressing hallucinations is a major focus in making AI systems more reliable and trustworthy.
Algorithm
An algorithm is a set of step-by-step instructions that a computer follows to perform a task or solve a problem. In the context of artificial intelligence, algorithms are the mathematical rules that guide how a system processes data, learns patterns, and makes predictions.
For example, a simple algorithm might sort numbers from smallest to largest, while a complex machine learning algorithm can analyse millions of images to recognize faces. Algorithms form the backbone of AI systems, telling them how to transform input data into useful output.
They can be designed for specific tasks, such as recommending products, detecting fraud, or optimizing routes. The effectiveness of an AI model depends largely on the quality and design of the algorithms behind it.
Anomaly Detection
Anomaly detection is an AI technique used to identify unusual patterns or outliers in data that don’t match expected behavior. It is widely used in fraud detection, cybersecurity, predictive maintenance, and healthcare monitoring. For example, banks use anomaly detection to spot suspicious credit card transactions that differ from a customer’s normal spending habits.
Artificial Intelligence (AI)
Artificial Intelligence, often shortened to AI, refers to computer systems designed to perform tasks that normally require human intelligence. These tasks include learning from data, understanding language, recognizing images, making decisions, and even generating new content. At its core, AI works by processing large amounts of information, finding patterns, and using those patterns to make predictions or create outputs.
There are two main types of AI. Narrow AI is specialized for specific tasks, like recommending a movie on Netflix or powering a voice assistant such as Alexa. General AI, which is still theoretical, would have the ability to think and reason like a human across many different areas.
AI is now embedded in everyday life, from navigation apps and spam filters to advanced tools that write text, compose music, and design art. Businesses use it for automation and analytics, while researchers explore its potential in healthcare, education, and climate science.
B
Bias in AI
Bias in AI refers to unfair or skewed outcomes produced by artificial intelligence systems due to the data they were trained on or the way their algorithms were designed. Since AI learns from historical data, if that data contains human biases — such as gender, racial, or socioeconomic imbalances — the AI can reproduce and even amplify those biases.
For example, if a hiring algorithm is trained on past resumes that favor one gender, it may unfairly recommend candidates of that gender more often. Similarly, medical AI trained primarily on data from one demographic may produce less accurate results for others.
Addressing bias in AI is crucial for building fair, ethical, and trustworthy systems. This involves using diverse datasets, transparent design, and continuous monitoring to reduce discriminatory outcomes.
Big Data
Big Data describes extremely large and complex datasets that traditional computing methods cannot easily process. It is characterized by high volume (amount of data), velocity (speed of generation), and variety (different types of data, such as text, images, and video).
AI relies on Big Data to learn and improve. For example, training a language model like ChatGPT requires analyzing billions of sentences. Similarly, image recognition AI is trained on massive collections of labeled photos. The more data an AI system has, the better it can detect patterns and make predictions.
Big Data is not just about size — it’s about extracting meaningful insights. Industries like healthcare, finance, and retail use Big Data analytics to forecast trends, optimize operations, and personalize services.
C
Chatbot
A chatbot is an AI-powered program designed to simulate conversation with humans, typically through text or voice. Chatbots use natural language processing (NLP) to understand questions, provide information, and perform tasks in a way that feels conversational.
Common examples include customer service bots on websites, virtual assistants like Siri and Alexa, and messaging apps that help users book tickets or check account balances. Some chatbots follow scripted flows, while advanced ones use machine learning to handle complex, open-ended questions.
Chatbots are widely used in industries like retail, banking, healthcare, and travel because they reduce wait times, automate routine inquiries, and improve customer service around the clock.
Clustering
Clustering is a technique in unsupervised learning where AI groups data points into clusters based on similarities, without needing predefined labels. The goal is to find hidden patterns or natural groupings within a dataset.
For example, a retailer might use clustering to segment customers into groups with similar shopping habits. In healthcare, clustering can group patients by similar symptoms or responses to treatments. Social media platforms use clustering to recommend content or identify communities with shared interests.
Clustering is widely used because it helps reveal structure in complex data, making it easier to analyze and act upon. Popular algorithms for clustering include k-means, hierarchical clustering, and DBSCAN.
Computer Vision
Computer vision is a field of AI that enables machines to interpret and understand visual information from the world, such as images and video. By analyzing pixels, patterns, and shapes, AI systems can identify objects, detect motion, and even interpret facial expressions.
Applications of computer vision are everywhere: self-driving cars use it to detect pedestrians and traffic signals, social media platforms use it for facial recognition and photo tagging, and healthcare systems use it to analyze medical images like X-rays and MRIs.
Computer vision relies heavily on deep learning, where neural networks process large amounts of labelled image data to learn visual patterns. This allows AI systems to achieve human-like, and sometimes superhuman, accuracy in visual tasks.
Conversational AI
Conversational AI refers to technologies like chatbots and voice assistants that can engage in human-like dialogue using natural language processing (NLP). It powers tools such as Siri, Alexa, and customer service bots, enabling users to interact with systems more naturally. Conversational AI is improving rapidly with large language models that understand context and generate fluent, human-like responses.
D
Data Augmentation
Data augmentation is the process of artificially expanding a dataset by creating modified versions of existing data. In computer vision, for example, images may be rotated, cropped, or color-adjusted to train more robust AI models. This technique helps reduce overfitting, improves generalization, and boosts accuracy when real-world data is limited.
Data Labelling
Data labeling is the process of tagging raw data — such as images, text, or audio — with meaningful labels so that machine learning models can learn from it. For example, labeling photos with “cat” or “dog” helps an AI system learn to recognize animals. High-quality labeled data is crucial for training accurate AI systems.
Data Mining
Data mining is the process of discovering patterns, trends, and useful information from large sets of data. In AI, it refers to extracting insights from datasets that can then be used to train models or make predictions.
For example, data mining might reveal that certain shopping habits lead to repeat purchases, which can help businesses target loyal customers. In healthcare, it can uncover patterns in patient records to improve diagnosis and treatment plans.
The process typically involves cleaning data, analysing it with algorithms, and identifying correlations or anomalies. Data mining is essential in building AI systems because it provides the knowledge that algorithms use to learn and improve.
Deepfake
A deepfake is synthetic media, usually a video or image, created using AI techniques like generative adversarial networks (GANs) to replace one person’s likeness with another’s. Deepfakes can be used for entertainment, satire, or creative projects, but they also raise serious ethical concerns when used for misinformation, fraud, or non-consensual content. Detecting and regulating deepfakes has become a key challenge in AI ethics.
Deep Learning
Deep learning is a subset of machine learning that uses large neural networks with many layers (hence “deep”) to process data and recognize complex patterns. Inspired by how the human brain works, deep learning enables AI systems to handle tasks like image recognition, natural language processing, and speech translation.
What makes deep learning powerful is its ability to automatically learn features from raw data without manual programming. For example, in image recognition, early layers of the network detect edges and shapes, while deeper layers recognize objects like faces or animals.
Deep learning powers many of today’s most advanced AI applications, including self-driving cars, virtual assistants, generative AI, and medical image analysis. It has become the foundation of modern AI breakthroughs.
Deep Reinforcement Learning
Deep reinforcement learning (DRL) combines deep learning with reinforcement learning, allowing AI to learn complex behaviors through trial and error. By using neural networks, DRL can handle high-dimensional inputs like images and video. It has powered breakthroughs such as AlphaGo beating world champions, robotics learning complex movements, and advanced game-playing AIs.
Digital Twin
A digital twin is a virtual replica of a physical object, system, or process that uses real-time data and AI to simulate performance, predict outcomes, and optimize operations. For example, digital twins are used in manufacturing to monitor machinery, in healthcare to model patient health, and in smart cities to simulate traffic flows.
Dimensionality Reduction
Dimensionality reduction is a technique in machine learning that simplifies large datasets by reducing the number of input variables while preserving important patterns. This makes models faster and easier to interpret. Methods like Principal Component Analysis (PCA) and t-SNE are commonly used. For instance, reducing thousands of genetic markers into a smaller set can still reveal useful insights in medical research.
E
Edge AI
Edge AI refers to running artificial intelligence models directly on devices (the “edge”) rather than relying solely on cloud servers. Examples include AI in smartphones, smart cameras, or autonomous vehicles. By processing data locally, edge AI reduces latency, enhances privacy, and allows real-time decision-making. It’s especially important for applications like self-driving cars, drones, and IoT devices.
Explainability
Explainability in AI refers to how easily humans can understand the reasoning behind an AI system’s outputs. Black-box models like deep neural networks often make accurate predictions but are difficult to interpret. Explainability techniques help highlight which features influenced a decision, making AI more transparent and accountable.
Explainable AI (XAI)
Explainable AI (XAI) refers to methods and tools that make the decision-making process of AI systems more transparent and understandable to humans. Since many advanced AI models, especially deep learning networks, operate like “black boxes,” it can be difficult to know why they produced a certain output.
For example, if an AI system denies a loan application, XAI techniques can highlight which factors — such as income, credit history, or employment status — influenced that decision. This builds trust and allows users to challenge or verify results.
Explainable AI is especially important in sensitive areas like healthcare, finance, and law, where understanding the “why” behind an AI recommendation is as critical as the result itself.
F
Feature Engineering
Feature engineering is the process of selecting, creating, or transforming variables (features) in a dataset to improve the performance of a machine learning model. Features are the inputs an algorithm uses to make predictions, and choosing the right ones can make a significant difference.
For example, in predicting house prices, raw data might include square footage, number of bedrooms, and location. Feature engineering could involve creating new variables like “price per square foot” or categorizing neighborhoods into zones. These engineered features often help the model identify more accurate patterns.
Effective feature engineering combines domain knowledge, creativity, and technical skills. While deep learning reduces the need for manual feature engineering by learning directly from raw data, it remains a crucial step in many AI projects.
Federated Learning
Federated learning is a machine learning technique where models are trained across multiple devices without sharing raw data. Instead, only the model updates are sent to a central server. This approach improves privacy while still enabling powerful AI training. It’s commonly used in mobile devices for predictive text, personalized recommendations, and health applications.
G
Generative AI
Generative AI is a type of artificial intelligence that can create new content rather than just analyze existing data. By learning from large datasets, these models can produce original text, images, music, video, and even computer code.
What sets generative AI apart from traditional AI is its ability to generate something new. For example, ChatGPT can write essays, articles, or poems, while DALL·E can create images from written prompts. In music, tools like AIVA compose original scores, and in design, AI can suggest new product concepts.
Generative AI relies heavily on advanced neural networks, particularly large language models (LLMs) and diffusion models. These systems learn patterns, context, and style from training data, then use that knowledge to craft outputs that appear human-made.
This technology is transforming industries, from marketing and entertainment to healthcare and education, while also raising ethical questions about originality, copyright, and trust.
Generative Adversarial Network (GAN)
A Generative Adversarial Network, or GAN, is a type of machine learning model made up of two neural networks that compete with each other: a generator and a discriminator. The generator creates new data, such as images or music, while the discriminator evaluates whether the data looks real or fake. Through this competition, the generator becomes increasingly skilled at producing realistic outputs.
GANs are behind many AI breakthroughs, including the creation of lifelike deepfake videos, realistic artwork, and even AI-generated fashion designs. While powerful, they also raise ethical concerns around authenticity and misuse.
H
Heuristic
A heuristic is a problem-solving approach that uses practical shortcuts or rules of thumb rather than exhaustive calculations. In AI, heuristics guide algorithms to find good-enough solutions more efficiently, especially in complex search problems like chess or route optimization. While not always perfect, heuristics make AI systems faster and more scalable.
Human-in-the-Loop (HITL)
Human-in-the-loop is an AI training and deployment approach where human judgment is integrated into the system. Instead of leaving decisions entirely to algorithms, humans provide feedback, corrections, or approvals. HITL improves accuracy, reduces bias, and ensures that AI systems align with human values. For instance, content moderation often uses HITL, where AI flags harmful posts but humans make the final decision.
Hyperparameter
A hyperparameter is a setting in a machine learning model that is chosen before training begins and controls how the model learns. Examples include the learning rate, number of layers in a neural network, or the number of clusters in a clustering algorithm. Choosing the right hyperparameters can greatly improve model performance, and the process of finding them is called hyperparameter tuning.
Hybrid AI
Hybrid AI combines different approaches to artificial intelligence, such as machine learning, rule-based systems, and symbolic reasoning, into a single model. This makes AI systems more flexible and powerful because they can use both data-driven learning and human-like logic. For example, a healthcare AI might use machine learning to analyse medical images while relying on symbolic reasoning to explain its recommendations.
I
Image Recognition
Image recognition is a branch of computer vision where AI identifies objects, people, text, or scenes within digital images. It is used in applications like facial recognition on smartphones, quality control in manufacturing, and diagnosing diseases from medical scans. By training on large datasets of labelled images, AI systems learn to detect patterns and achieve high accuracy in visual tasks.
J
Joint Probability Distribution
A joint probability distribution is a mathematical concept used in AI and statistics to describe the likelihood of two or more variables occurring together. In machine learning, it helps models understand relationships between features. For example, in a shopping recommendation system, the joint probability of a user liking both shoes and jackets can guide more accurate suggestions.
K
K-Means Clustering
K-Means clustering is a popular unsupervised learning algorithm used to group data points into clusters based on similarity. The algorithm assigns each point to the nearest cluster center, then adjusts the centers until the groups stabilize. It is commonly used in customer segmentation, image compression, and anomaly detection.
Knowledge Graph
A knowledge graph is a way of organizing information so that machines can understand the relationships between different concepts. Instead of just storing data in tables, a knowledge graph connects facts like a web, showing how people, places, and things are linked.
Google uses knowledge graphs to power its search results. When you look up a famous person, for example, the panel that shows their biography, works, and related figures comes from a knowledge graph. This structure helps AI provide more accurate, context-rich answers instead of isolated facts.
L
Large Language Model (LLM)
A Large Language Model, or LLM, is a type of AI trained on massive amounts of text to understand and generate human-like language. These models, such as GPT, can answer questions, write essays, summarize information, and even hold conversations.
The strength of LLMs lies in their scale — the more data they process and the larger their parameters, the more nuanced their understanding becomes. LLMs are now used in chatbots, translation services, content creation tools, and research assistance.
Logistic Regression
Logistic regression is a statistical method widely used in machine learning for classification tasks. Unlike linear regression, which predicts continuous values, logistic regression predicts probabilities of categories, such as “spam” or “not spam.” It is simple, interpretable, and effective for binary classification problems in fields like healthcare, finance, and marketing.
M
Machine Learning (ML)
Machine Learning, or ML, is a branch of artificial intelligence that allows computers to improve their performance without being explicitly programmed. Instead of following step-by-step rules, a machine learning system is trained on data and learns to identify patterns and make predictions on its own.
For example, when a spam filter learns which emails to block, it does so by studying thousands of examples of spam and non-spam messages. Similarly, a recommendation engine like the one on YouTube learns your preferences by analyzing your viewing history and comparing it to patterns from millions of other users.
Machine learning can be supervised, where the model learns from labeled examples, or unsupervised, where it finds patterns without guidance. A third type, reinforcement learning, trains models through trial and error with rewards and penalties.
ML powers many AI applications today, including voice recognition, fraud detection, and medical image analysis.
Multimodal AI
Multimodal AI refers to artificial intelligence systems that can process and combine different types of input — such as text, images, audio, and video — to generate more accurate and versatile outputs. For example, a multimodal AI might analyze an X-ray image alongside a doctor’s notes to provide a better diagnosis, or generate a caption for a photo by combining visual recognition with natural language processing. Tools like OpenAI’s GPT-4 and Google’s Gemini are leading examples of multimodal AI in action.
N
Natural Language Generation (NLG)
Natural language generation is a branch of AI that produces human-like text from structured data. It powers applications like automated news reports, business intelligence summaries, and chatbots. NLG is often combined with natural language understanding (NLU) to create systems that can both interpret and generate language naturally.
Natural Language Processing (NLP)
Natural Language Processing, or NLP, is a field of AI that focuses on enabling computers to understand, interpret, and generate human language. It bridges the gap between how humans communicate and how machines process information.
Common applications of NLP include chatbots, translation apps, and voice assistants like Siri and Google Assistant. When you type a query into a search engine or ask Alexa to play a song, NLP algorithms process your words, analyze context, and deliver a meaningful response.
NLP combines techniques from linguistics, computer science, and machine learning. It involves tasks like tokenization (breaking down text into words), sentiment analysis (detecting emotions in text), and entity recognition (identifying names, places, or dates). Modern NLP models, such as large language models, have advanced to the point where they can hold conversations, summarize documents, and even write creative text.
NLP plays a critical role in making AI accessible, since language is the most natural interface humans use.
Neural Network
A neural network is a type of machine learning model inspired by the human brain. Just like neurons in our brain pass signals to each other, artificial “neurons” in a neural network process information in layers. These layers work together to recognize patterns, make predictions, or generate outputs.
At a basic level, each neuron takes in inputs, processes them through a mathematical function, and sends the result to the next layer. When stacked together, these layers can learn very complex patterns. For instance, in image recognition, one layer may detect edges, the next shapes, and later layers entire objects.
Neural networks are the backbone of modern AI. Deep learning, a subset of machine learning, uses large and complex neural networks to achieve breakthroughs in areas like self-driving cars, natural language processing, and generative AI.
Because of their ability to mimic human pattern recognition, neural networks have made AI systems far more powerful and flexible than earlier approaches.
O
Overfitting
Overfitting is a common problem in machine learning where a model learns the training data too well, including its noise and small details, instead of focusing on general patterns. As a result, the model performs very accurately on the training dataset but struggles to make correct predictions on new, unseen data.
Think of it like a student who memorizes every question and answer from past exams but cannot handle slightly different questions on the real test. Overfitting makes AI models less reliable in practical applications.
To prevent overfitting, data scientists use techniques like cross-validation, regularization, and dropout methods. Ensuring a diverse and balanced training dataset also helps models generalize better to new situations.
Overfitting vs. Underfitting
Overfitting and underfitting are two common issues in machine learning. Overfitting happens when a model learns the training data too well, including its noise and details, making it less effective on new data. Underfitting is the opposite — the model is too simple and fails to capture important patterns in the data.
Imagine teaching a child math problems. If they memorize answers without understanding (overfitting), they struggle with new questions. If they only learn a vague idea without practice (underfitting), they also perform poorly. The goal in AI is to find the right balance so the model generalizes well to new situations.
P
Predictive Analytics
Predictive analytics is a branch of AI that uses data, statistical algorithms, and machine learning techniques to forecast future outcomes. Instead of just describing what happened in the past, predictive analytics provides insights into what is likely to happen next.
For example, airlines use predictive analytics to forecast flight delays, retailers predict customer buying habits, and healthcare providers anticipate patient needs. By identifying patterns in large datasets, AI systems can help organizations make proactive decisions.
Predictive analytics is widely used across industries because it improves efficiency, reduces risk, and supports smarter planning. It is one of the most practical and impactful applications of AI today.
Prompt Engineering
Prompt engineering is the practice of designing and refining inputs (prompts) to get the best possible outputs from generative AI systems, such as large language models. By carefully choosing wording, context, and instructions, users can guide AI to produce more accurate, creative, or useful results. It has become a key skill in working effectively with AI tools.
R
Reinforcement Learning
Reinforcement learning is a type of machine learning where an AI system learns by interacting with its environment and receiving feedback in the form of rewards or penalties. Instead of being told the correct answer, the system figures it out through trial and error.
This approach is inspired by how humans and animals learn — by experimenting, observing the results, and adjusting behavior. For example, an AI program trained to play chess learns by making moves, seeing whether it wins or loses, and refining its strategy over time.
Reinforcement learning is behind some of the most impressive AI breakthroughs, such as AlphaGo defeating world champions in board games and robots learning to navigate complex environments.
Reinforcement Learning with Human Feedback (RLHF)
Reinforcement Learning with Human Feedback is a training method where AI learns not only from trial and error but also from guidance provided by humans. This process helps align AI systems with human values and preferences.
For example, when training a language model, humans rank different responses, and the AI uses this feedback to improve future answers. RLHF has become a cornerstone for making conversational AI systems safer, more helpful, and more aligned with human expectations.
S
Sentiment Analysis
Sentiment analysis is a natural language processing (NLP) technique that identifies the emotional tone of text — whether positive, negative, or neutral. Businesses use it to analyze customer reviews, monitor social media, and track public opinion. For example, a brand can quickly see how people feel about a new product by analyzing thousands of tweets. Sentiment analysis helps organizations understand audiences at scale.
Supervised Learning
Supervised learning is a type of machine learning where a model is trained on labeled data. This means each example in the training dataset comes with both the input (such as an image) and the correct output (such as the label “cat” or “dog”). The model learns to map inputs to outputs so it can make predictions on new data.
For example, a supervised learning model trained on thousands of labeled photos of fruit can learn to distinguish between apples, oranges, and bananas. Once trained, it can then identify fruit in new, unseen images.
Supervised learning is widely used for applications like spam detection, credit scoring, and medical diagnosis, where accurate labelled data is available.
Swarm Intelligence
Swarm intelligence is a field of AI inspired by the collective behaviour of animals like ants, bees, or birds. In these systems, simple agents follow basic rules but produce complex group behaviour when combined. AI applies this concept to optimization problems, robotics coordination, and traffic management. For instance, swarm intelligence algorithms can efficiently route delivery trucks or coordinate drone fleets.
Synthetic Data
Synthetic data is artificially generated information that mimics real-world data without exposing sensitive details. It is used to train AI models when real data is scarce, expensive, or restricted due to privacy concerns. For example, synthetic medical records can be created to train healthcare AI systems without violating patient confidentiality.
T
Tokenization
Tokenization is the process of breaking down text into smaller units, called tokens, that AI models can process. Tokens can be words, characters, or subwords, depending on how the system is designed.
Large language models like GPT rely heavily on tokenization to understand input and generate output. For example, the sentence “AI is amazing” might be split into the tokens “AI,” “is,” and “amazing.” This allows the model to analyse patterns in language at scale.
Transfer Learning
Transfer learning is a technique in machine learning where a model trained on one task is reused as the starting point for another, related task. Instead of training a model from scratch, you leverage the knowledge it has already gained.
For example, a neural network trained on millions of general images can be adapted to recognize medical X-rays with far less data. This makes AI development faster, cheaper, and more accurate, especially in fields where labelled data is scarce.
Turing Test
The Turing Test, proposed by computer scientist Alan Turing in 1950, is a measure of a machine’s ability to exhibit intelligent behavior indistinguishable from that of a human. In the test, a human judge interacts with both a machine and another human without knowing which is which. If the judge cannot reliably tell them apart, the machine is said to have passed the test.
While modern AI systems can generate human-like responses, most experts agree that passing the Turing Test doesn’t necessarily mean true intelligence — only that the machine can mimic human conversation effectively.
U
Unsupervised Learning
Unsupervised learning is a type of machine learning where the model is trained on data without labels. Instead of being told the correct answers, the system looks for hidden patterns, groupings, or structures within the dataset.
For example, an unsupervised learning algorithm analyzing customer data might discover that shoppers naturally fall into different groups based on their spending habits, even though those groups weren’t predefined.
This approach is commonly used in clustering, market segmentation, anomaly detection, and recommendation systems. While it can be harder to evaluate than supervised learning, unsupervised learning is powerful for uncovering insights from raw, unstructured data.
V
Vectorization
Vectorization is the process of converting data — such as words, images, or sounds — into numerical vectors that AI systems can understand. These vectors capture meaning, relationships, or features in a mathematical form that algorithms can process.
For example, in natural language processing, the words “king” and “queen” can be represented as vectors in a way that captures their semantic relationship. In computer vision, an image is broken down into a vector of pixel values or features.
W
Weak AI
Weak AI, also known as narrow AI, refers to systems designed to perform specific tasks very well but without general intelligence or consciousness. Voice assistants like Alexa, recommendation engines like Netflix’s, and image recognition models are all examples of weak AI.
This is in contrast to the concept of general AI, which would be able to learn, reason, and adapt across many domains like a human. Weak AI dominates today’s real-world applications and powers most of the technology we interact with daily.
X
XAI (Explainable AI)
Explainable AI, often shortened to XAI, refers to techniques that make AI systems more transparent by showing how they reach decisions. Instead of working like a “black box,” XAI highlights the factors that influenced an output. This is especially important in sensitive areas like healthcare, finance, or hiring, where users need to trust and understand AI recommendations.
Y
YOLO (You Only Look Once)
YOLO is a popular deep learning algorithm used for real-time object detection in images and video. Unlike older methods that scan an image in parts, YOLO processes the entire image at once, making it extremely fast and efficient. It is widely used in applications such as self-driving cars, surveillance systems, and robotics.
Z
Zero-Shot Learning
Zero-shot learning is a machine learning approach where an AI system can correctly perform tasks or recognize objects it has never seen during training. It works by transferring knowledge from related tasks or using descriptive information. For example, a model trained to recognize cats and dogs might also identify a zebra if given the right descriptive labels, even without zebra images in its training data.
