Artificial Intelligence Applied Skills + Build Your Own AI

Glossary:

A

Adversarial [Network, Model]: Models that fool other models.

AGI: Artificial General Intelligence, versatile AI.

AI Assistant: Digital helper powered by AI.

AI Content: Material generated or curated by AI.

AI Detector: Tool identifying AI-generated content.

Anthropic: AI research and safety company.

API, API Call: Interface for software interaction.

Arc Search: Advanced search algorithm.

ATS Resume Review: Filters candidates automatically.

Autonomous Agent: Self-governing AI entity.

B

Back End Developer: Manages server, database, application.

Bias: Prejudice in data or algorithms.

Big Data: Extremely large data sets.

BI, PowerBI: Business intelligence tools.

C

Chatbot: Automated conversation system.

Chat GPT 3.5, 4, 5, V (Vision): Iterations of OpenAI's language models.

Classifier: Categorizes data into classes.

Claude: AI language model by Anthropic.

Cloud: Internet-based computing resources.

Compute (Time, Cost): Resources for AI processing.

Copilot: AI-based coding assistant.

CRM: Customer Relationship Management software.

CPU: Central Processing Unit, computer's brain.

Custom GPT: Tailored Generative Pre-trained Transformer.

D

Data Analysis, File Analysis: Examining data to extract insights.

Data Center: Facility for computer systems.

Data Cutoff: Latest point of information inclusion.

Data Leak: Unauthorized data exposure.

Data Scientist, Data Analyst: Professionals analyzing data.

Decentralized: Distributed control or authority.

Deep Learning: Learning through multiple processing layers.

Deepfake: Realistic AI-generated video or audio.

DeepMind: Leading AI research company.

Einstein: Salesforce's AI platform.

E

Emergent Capabilities: Unanticipated AI abilities.

ERP: Enterprise Resource Planning software.

F

Front End Developer: Develops user interface and experience.

G

Gemini / Gemma: AI models or projects (hypothetical).

GenAI, Generative AI: AI creating original content.

GPU: Graphics Processing Unit, accelerates processing.

Gradient Descent: Optimization algorithm for learning.

GUI: Graphical User Interface.

H

Hallucinate: AI generating false or imagined data.

Huggingface: AI community and model hub.

I

J

Jailbreak: Bypassing device or software restrictions.

K

K Means Cluster: Partitioning data into k groups.

L

LaMDA: Language Model for Dialogue Applications.

Labeling: Assigning labels to data for training.

Llama 2: AI model by Meta.

LLM: Large Language Model.

Local Model: AI model running on local hardware.

M

Machine Learning: Algorithms improving through experience.

Malicious Behavior: Intentionally harmful actions by AI.

Meta Data: Data about data.

Microsoft Graph: API for Microsoft 365 services.

Mistral: AI project or model (hypothetical).

Model: Simplified representation for predictions.

N

Nemotron: AI project or model (hypothetical).

Neural Network: Network mimicking the human brain.

Nightshade: AI project or model (hypothetical).

NLP: Natural Language Processing.

No Code: Development without traditional coding.

Nvidia: Leading GPU manufacturer.

O

One Shot Learning: Learning from a single example.

Open Source: Software with publicly available code.

P

Palm: AI project or model (hypothetical).

Parameters: Variables defining a model's function.

Persistent Memory: Retains data without power.

Power BI: Microsoft's business analytics service.

Prompt / Prompt Engineering: Crafting inputs for AI models.

Python: Popular programming language for AI.

Python Library: Reusable code for Python.

Q

Q* (Q Star): Optimal action-value function.

Q

R

Reasoning: Logical thinking by AI.

Red Teaming: Security testing by simulating attacks.

Regression: Predicting numerical values.

Reinforcement RLHF: Learning from feedback to improve.

Relational Database: Structured data storage system.

Recommender System: Suggests items to users.

S

Sci-Kit: Python library for machine learning.

Semantic: Relating to meaning in language.

SoRa: AI project or model (hypothetical).

Speech Recognition: Translating spoken words into text.

SQL: Language for managing databases.

Structured Data, Unstructured Data: Organized vs. unorganized data.

Supervised Learning: Learning with labeled data.

T

Testing Set: Data for evaluating a model.

Train Training: Process of teaching a model.

Training Set / Training Data: Data for training a model.

Transformer: Architecture behind modern AI models.

U

Unsupervised: Learning without labeled data.

V

Voice Clone: AI-generated replica of a voice.

W

Watson AI: IBM's suite of AI services.

Weights and Biases: Parameters in neural networks.

Workflow: Sequence of processes or tasks.

X

Y

Z

Zero Shot Learning: Learning without prior examples.