AI Glossary Infographic

🧠 AI Fundamentals

Artificial Intelligence (AI)

Development of systems that can perform tasks typically requiring human intelligence, such as reasoning, learning, and problem-solving.

Fundamental in advancing scientific research and addressing complex challenges across domains.

General AI (AGI)

AI systems with human-like versatility, capable of performing a wide range of tasks across various domains with adaptability and reasoning.

Aims to generalize and replicate human intelligence across diverse situations.

Narrow AI

AI systems designed to perform a specific task or limited range of tasks with high efficiency and precision.

Most prevalent form of AI today, driving advancements in specialized applications.

Agentic AI

AI systems designed to act autonomously, perceiving their environment, making decisions, and taking actions to achieve specific goals.

Crucial for applications requiring independence and proactive behavior.

Adaptable AI

Systems capable of dynamically adjusting their behavior, strategies, or models in response to changing environments without explicit reprogramming.

Critical for developing robust systems that operate effectively in unpredictable scenarios.

📊 Machine Learning

Machine Learning (ML)

A subset of AI that uses statistical methods to enable systems to learn from data and improve performance over time.

Drives advances in AI by automating learning processes and analyzing vast datasets.

Classic Machine Learning

Traditional algorithms that rely on structured, tabular data without neural networks. Includes Decision Trees, SVMs, k-NN, and regression.

Valued for simplicity, efficiency, and interpretability on smaller datasets.

Supervised Learning

Machine learning approach where models are trained on labeled datasets, learning to map inputs to outputs.

Essential for classification and regression tasks requiring high accuracy.

Unsupervised Learning

Machine learning approach using unlabeled data to identify patterns and structures through clustering and dimensionality reduction.

Crucial for anomaly detection and discovering hidden relationships in data.

Reinforcement Learning

Learning method where an agent learns by interacting with its environment, maximizing cumulative rewards over time.

Valuable for developing adaptive systems and autonomous agents.

Transfer Learning

Technique where a model developed for one task is adapted to a different but related task by leveraging learned knowledge.

Accelerates AI development by reducing data and computational requirements.

Fine-Tuning

Adapting a pre-trained model to a specific task by training it further on new data with adjusted parameters.

Enables efficient reuse of pre-trained models with limited data.

Zero-shot Learning

Enables AI models to perform tasks without prior specific training, leveraging general knowledge and contextual understanding.

Reduces need for extensive labeled datasets, making AI more adaptable.

🧮 Neural Networks

Neural Networks

Machine learning technique modeled after the human brain's structure, consisting of interconnected nodes that process data using weights and activation functions.

Fundamental to modern AI, powering advancements in recognition and processing tasks.

Deep Neural Networks (DNNs)

Neural networks with multiple layers enabling hierarchical learning of features from raw data.

Revolutionized AI by achieving state-of-the-art performance in complex tasks.

Convolutional Neural Networks (CNNs)

Deep learning models designed for processing grid-like data such as images, using convolutional layers to extract hierarchical features.

Foundational to modern computer vision tasks and image recognition.

Recurrent Neural Networks (RNNs)

Neural networks designed for sequential data processing, where previous step outputs are used as current step inputs.

Essential for tasks involving sequential or time-series data.

LSTM (Long Short-Term Memory)

Specialized RNN designed to overcome limitations by effectively managing long-term dependencies in sequential data.

Addresses vanishing gradient problem, highly effective for sequence-based tasks.

Transformers

Neural network architecture for sequence-to-sequence tasks, leveraging self-attention mechanisms to process input data efficiently.

Revolutionized AI by enabling scalable training and state-of-the-art results.

Autoencoders

Neural networks designed to learn efficient, low-dimensional representations by encoding input into compressed form and reconstructing it.

Widely used for dimensionality reduction, anomaly detection, and data denoising.

GANs (Generative Adversarial Networks)

Neural network architecture with two competing networks: a generator creating synthetic data and a discriminator evaluating authenticity.

Revolutionized content creation, enabling generation of highly realistic synthetic data.

🚀 Applied AI

Computer Vision

Field of AI focused on enabling machines to interpret and understand visual information from images and videos.

Powers applications like autonomous vehicles, medical imaging, and security systems.

Natural Language Processing (NLP)

Field of AI that enables machines to understand, interpret, generate, and interact using human languages.

Bridges gap between human communication and machine processing.

Large Language Models (LLMs)

Subset of large models specifically designed for natural language processing tasks, trained on vast amounts of text data.

Revolutionized natural language processing with state-of-the-art performance.

Multimodal AI

AI systems that process and integrate multiple types of data such as text, images, audio, and video.

Enhances AI versatility, aligning capabilities closer to human-like perception.

Generative AI

AI systems that produce new content such as text, images, music, or videos by learning patterns from existing data.

Revolutionizes creative fields and enables synthetic data generation.

Robotics

Interdisciplinary field combining AI, engineering, and computer science to develop robots capable of autonomous or semi-autonomous tasks.

Transforms industries by automating tasks and handling complex real-world challenges.

Predictive AI

Focuses on analyzing historical and real-time data to forecast future trends, behaviors, or events.

Fundamental for proactive decision-making and strategic planning.

⚖️ Ethics & Governance

Ethical AI

Development and use of AI systems aligned with moral principles, societal norms, and regulatory frameworks for responsible outcomes.

Essential for safeguarding human rights and fostering trust in AI technologies.

Responsible AI

Practice of developing, deploying, and managing AI systems upholding ethical principles including fairness, transparency, and accountability.

Ensures AI technologies align with societal values and human rights.

Fairness

Principle ensuring AI systems provide equitable and unbiased outcomes for all individuals regardless of personal attributes.

Fundamental to ethical AI deployment, preventing perpetuation of societal inequalities.

Explainable AI (XAI)

Methods and systems designed to make AI model inner workings transparent, providing clear explanations for their decisions.

Critical for fostering accountability and trust in high-stakes applications.

Auditable AI

Systems designed with mechanisms allowing their processes, decisions, and outcomes to be reviewed and verified.

Ensures accountability and compliance with regulatory standards.

Hallucination

Phenomenon where language models generate incorrect or misleading information that appears plausible due to lack of factual grounding.

Crucial to identify and mitigate for ensuring trust in AI-generated outputs.

🔬 Advanced Concepts

Knowledge Representation and Reasoning (KR&R)

Field focused on encoding world information in structured forms and enabling systems to use this knowledge for reasoning and decision-making.

Fundamental to AI, bridging human understanding and machine processing.

Symbolic AI

Paradigm focusing on representing and manipulating knowledge using symbols, rules, and formal logic for reasoning and decision-making.

Fosters transparency and explainability in AI systems.

Neuro-Symbolic AI

Hybrid AI integrating neural networks with symbolic reasoning to create systems capable of both learning and logic-based reasoning.

Enhances AI by addressing explainability, generalization, and reasoning challenges.

Hybrid AI

Approaches integrating symbolic reasoning with data-driven methods to combine advantages of both paradigms.

Bridges gap between interpretability and adaptability.

Continual Learning

Machine learning paradigm where AI systems incrementally learn from new data while preserving previously acquired knowledge.

Mimics human cognitive processes, enabling AI systems to evolve and adapt continuously.

RAG (Retrieval-Augmented Generation)

AI technique integrating external knowledge sources with language models to generate more accurate and informed responses.

Enhances AI reliability by retrieving relevant information from structured sources.

Prompt Engineering

Practice of designing and refining prompts to achieve precise and desired outputs from AI models.

Essential for maximizing AI effectiveness and improving reliability.

Tokens

Fundamental units of processing in LLMs, representing chunks of text such as words, subwords, or characters.

Managing tokens efficiently is crucial for optimizing AI responses and controlling costs.

Scalable AI

AI systems designed to maintain efficiency and performance as data volume, task complexity, or user demands increase.

Crucial for handling large-scale applications and ensuring long-term reliability.