Artificial Intelligence: History, Foundations, and Applications

Vocabulary flashcards covering the key milestones, foundational figures, paradigms, models, applications, and ethical frameworks of artificial intelligence from the 1950s to the present day.

Turing Test (Imitation Game)

A benchmark proposed by Alan Turing in his $1950$ paper 'Computing Machinery and Intelligence' stating that if a machine's responses are indistinguishable from those of a human, the machine passes as thinking.

Dartmouth Conference (1956)

A summer workshop convened by John McCarthy at Dartmouth College that formalized artificial intelligence as a distinct field of research and established its founding community.

Logic Theorist

The first symbolic AI program, created in $1955$ by Allen Newell and Herbert Simon, which demonstrated that machines could carry out logical reasoning.

SNARC

The first neural network machine, built in $1951$ by Marvin Minsky and Dean Edmonds.

Machine Learning (Samuel's Definition)

A term coined in $1959$ by Arthur Samuel to describe self-improving computer programs, demonstrated through his checkers-playing systems.

ELIZA

The world's first chatbot, developed at MIT in $1966$ by Joseph Weizenbaum.

First AI Winter

A period of reduced funding and interest from $1974$ to $1980$, triggered by the fundamental limits of symbolic AI and the UK's Lighthill Report.

Second AI Winter

A downturn from $1987$ to $1993$ precipitated by the collapse of the commercial expert systems market and the obsolescence of specialized AI hardware.

Expert Systems

Rule-based AI programs from the $1970\text{s}$ and $1980\text{s}$ that encoded human expertise into logical rules, which delivered real business value but suffered from brittleness when encountering novel scenarios.

MYCIN

A rule-based expert system developed at Stanford in $1972$ to diagnose bacterial infections, outperforming many human physicians in accuracy.

XCON

An automated configuration expert system deployed in $1980$ that saved Digital Equipment Corporation approximately $\text{\\textdollar}40\text{ million}$ per year.

Backpropagation

An algorithm popularized in $1986$ by Rumelhart, Hinton, and Williams that enabled multi-layer neural networks to learn effectively from errors.

Convolutional Neural Networks (CNNs)

A deep neural network architecture demonstrated by Yann LeCun in $1989$ for handwritten digit recognition, designed for processing grid-structured data like images.

Support Vector Machines (SVMs)

A class of statistical machine learning algorithms introduced in $1995$ by Vladimir Vapnik known for strong classification performance.

AlexNet

A convolutional neural network that won the ImageNet competition in $2012$ with a $10$-point error-rate lead, triggering the modern deep learning revolution.

Transformer Architecture

A neural network design introduced by Google Brain in the $2017$ paper 'Attention Is All You Need' that replaced recurrence with self-attention mechanisms, revolutionizing NLP and foundation models.

Self-Attention Mechanism

The core component of the Transformer architecture where a model computes relevance scores between all tokens in a sequence simultaneously to capture context.

Reinforcement Learning from Human Feedback (RLHF)

A training process used to refine and align foundation models using human ratings to ensure responses are safe, helpful, and accurate.

AlphaFold

An AI system that predicted the $3\text{D}$ structure of virtually every known protein, revolutionizing structural biology and therapeutic drug discovery.

Narrow AI

AI systems designed and optimized for a single, specific task that cannot transfer their learning to other domains, constituting $100\%$ of AI deployed today.

Artificial General Intelligence (AGI)

A hypothetical system matching or exceeding human cognitive abilities across all domains, characterized by the ability to generalize from few examples.

Agentic AI

An emerging paradigm of AI systems ($2025\text{--}2027$) capable of planning multi-step tasks, browsing the web, utilizing software tools, and collaborating autonomously with other agents.

Quantum AI

The application of quantum computing to AI training ($2027\text{--}2035$) aimed at shattering matrix computation bottlenecks in fields like drug discovery and materials science.

EU AI Act (2024)

The world's first comprehensive legal AI regulation, utilizing a risk-tiered framework that bans unacceptable-risk uses and enforces strict compliance on high-risk systems.