FUNAI - 1st Prelim

Artificial Intelligence (AI)

The ability of a machine to perform tasks that normally require human intelligence: learning, reasoning, understanding language, and solving problems

Turing Test

Test proposed by Alan Turing in 1950 to evaluate machine intelligence

Dartmouth Conference

The 1956 event where the term "Artificial Intelligence" was formally coined

AI Winters

Periods (1970s–80s) when AI funding and interest declined as early promises outpaced actual results

Deep Blue

IBM's system that defeated chess champion Garry Kasparov in 1997

Watson

IBM's system that won Jeopardy! against human champions in 2011

ImageNet

The 2012 deep learning breakthrough that sparked the modern AI boom

ChatGPT

Generative AI system launched in 2022, bringing generative AI into everyday use

ANI (Narrow AI)

AI designed to perform a single task or narrow set of tasks well; exists today (e.g. Siri, spam filters, Google Maps, ChatGPT)

AGI (General AI)

AI that would match human-level reasoning and learning across any task; hypothetical, doesn't exist yet

ASI (Superintelligence)

AI that would surpass human intelligence in every domain; purely theoretical

Reactive Machines

Responds only to the current input, with no memory of the past (e.g. IBM Deep Blue)

Limited Memory

Uses recent, stored data to inform decisions (e.g. self-driving cars, chatbots)

Theory of Mind

Would understand emotions, beliefs, and intentions; still a research goal

Self-Aware

Would have its own consciousness and self-understanding; purely hypothetical

Machine Learning

Systems that learn patterns from data instead of following fixed rules

Natural Language Processing (NLP)

Lets machines understand, interpret, and generate human language

Computer Vision

Enables machines to interpret and make sense of images and video

Robotics

Pairs AI with physical machines that sense and act in the real world

Algorithm

A set of rules

LLM

Per the notes, described as a scoring system for words and a database containing large amounts of information

Supervised Learning

Learns from labeled examples, like flashcards with the answer on the back

Unsupervised Learning

Finds patterns in data with no labels, like sorting a mixed bag of candy by feel

Reinforcement Learning

Learns by trial and error, guided by rewards, like training a dog with treats

Tokenization ("Break it into pieces")

Splitting a sentence into individual tokens (e.g. "I love pizza" → "I", "love", "pizza")

Translation

Converts text between languages

Sentiment Analysis

Detects whether text sounds positive, negative, or neutral

Chatbots

Hold a conversation and answer questions

Autocomplete

Predicts the next word as you type

Summarization

Condenses a long text into key points