Artificial Intelligence Glossary: ​​AI Terms Everyone Should Learn

We’ve compiled a list of useful phrases and concepts for understanding artificial intelligence, especially the new generation of AI-enabled chatbots like ChatGPT, Bing, and Bard.

If you don’t understand these explanations or want to know more, you can consider asking the chatbots themselves. Answering such questions is one of their most useful skills, and one of the best ways to understand AI is to use it. But keep in mind that they are sometimes wrong.

Bing and Bard chatbots are slowly rolling out, and you may need to join their waitlists to gain access. ChatGPT currently does not have a waitlist, but it does require creating a free account.

To learn more about AI, check out the New York Times five-part series on how to become a chatbot expert.

Anthropomorphism: The tendency of people to attribute human qualities or characteristics to an AI chatbot. For example, you can assume she’s nice or cruel based on her responses, even though she’s not capable of having emotions, or you can believe the AI ​​is sentient because she’s very good at it. to imitate human language.

Bias: A type of error that can occur in a large language model if its output is skewed by model training data. For example, a model may associate specific traits or occupations with a certain race or gender, resulting in inaccurate predictions and offensive responses.

Emergent behavior: Unexpected or unintended capabilities in a large language model, enabled by models and the model’s learning rules from its training data. For example, models trained on programming and coding sites can write new code. Other examples include creative abilities such as composing poetry, music, and fictional stories.

Generative AI: Technology that creates content, including text, images, video, and computer code, by identifying patterns in large amounts of training data and then creating original material with similar characteristics. Examples include ChatGPT for text and DALL-E and Midjourney for images.

Hallucination: A well-known phenomenon in large language models, in which the system provides an answer that is factually incorrect, irrelevant, or nonsensical, due to the limitations of its training data and architecture.

Large tongue model: A type of neural network that learns skills – including generating prose, conducting conversations and writing computer code – by analyzing large amounts of text on the Internet. The basic function is to predict the next word in a sequence, but these models have surprised experts by learning new abilities.

Natural language processing: Techniques used by large language models to understand and generate human language, including text classification and sentiment analysis. These methods often use a combination of machine learning algorithms, statistical models, and linguistic rules.

Neural network: A mathematical system, modeled after the human brain, that learns skills by finding statistical patterns in data. It consists of layers of artificial neurons: the first layer receives the input data and the last layer produces the results. Even experts who create neural networks don’t always understand what’s going on in between.

Settings: Numerical values ​​that define the structure and behavior of a large language model, like clues that help it guess which words come next. Systems like GPT-4 are believed to have hundreds of billions of parameters.

Reinforcement learning: A technique that teaches an AI model to find the best result through trial and error, receiving rewards or punishments from an algorithm based on its results. This system can be improved by humans who give their opinion on its performance, in the form of evaluations, corrections and suggestions.

Transformer model: A useful neural network architecture for understanding language that doesn’t have to parse words one by one but can look at an entire sentence at a time. This was a breakthrough in AI, as it allowed models to understand context and long-term dependencies in language. Transformers use a technique called self-attention, which allows the model to focus on particular words that are important to understanding the meaning of a sentence.