IDSC Topics 11 and 12
artificial intelligence (AI)
Computer software that can mimic or improve upon functions that would otherwise require human intelligence.
Hallucination
In the context of AI, hallucination refers to answers that generate results that are incorrect and aren't based on facts. Some refer to AI hallucination as “made up” answers.
Prompt
A request made to generative artificial intelligence systems.
the AI winter
generative AI
A type of artificial intelligence whose output is some type of media: e.g., text, images, audio, video. Results are usually generated based on a prompt, such as typed or spoken text. This technology “learns” the patterns and structure of data used during training, and can then generate new output based on the characteristics of this input.
constitutional AI
A method for providing alignment and safety in an AI by incorporating a set of specific rules or guidelines that an AI must follow as machine learning takes place.
genetic algorithms
AI technologies that seek an optimal model by transforming or “mutating” an algorithm (versus neural networks, which add weights and mappings to a combination of inputs)—iteratively testing the result and choosing the best outcome.
neural networks
Statistical techniques used in AI and particularly in machine learning. Neural networks hunt down and expose patterns, building multilayered relationships that humans can’t detect on their own.
Parameters
Values that are created by the foundation model and that are used by LLMs to determine text elements, relationships between these elements, and that are further refined during training.
machine learning (ML)
A type of artificial intelligence that leverages massive amounts of data so that computers can improve the accuracy of actions and predictions on their own without additional programming.
deep learning
A type of machine learning that uses multiple layers of interconnections among data to identify patterns and improve predicted results.
Turing test
A test of a computer’s ability to exhibit behavior so that it is indistinguishable from a human.
Corpus
The data used to train an AI.
foundation model
The model or base technology used to train a large language model. In the example of ChatGPT-3, OpenAI's GPT-3 is the base model that is used to train the LLM that users eventually use to ask and answer questions.
prompt engineering
The practice of designing inputs for generative AI tools that will produce optimal outputs.
large language model (LLM)
A type of AI that is used for general-purpose language understanding and generation. LLMs are trained by putting a corpus of training data through a foundation model.
reinforcement learning from human feedback
A machine learning training technique that uses a reward model and human evaluators that will provide feedback to continually tune results.
artificial general intelligence (AGI)
Refers to software that’s capable of learning any task or subject, including developing reasoning about topics not presented through a training corpus.
supervised learning
A type of machine learning where algorithms are trained by providing explicit examples of results sought, like defective versus error-free, or stock price.
expert systems
AI systems that leverage a set of programmed decision rules or example outcomes to perform a task in a way that mimics applied human expertise. Expert systems are used in tasks ranging from medical diagnoses to product configuration.
self-supervised learning
Sometimes called unsupervised learning, where systems build pattern-recognizing algorithms using data that has not been pre-classified.
pair programming
A software development technique where two programmers work side-by-side, looking at the same screen, sharing their collective insights to write, test, and debug code.
OCR
Optical Character Recognition. Software that can scan images and identify text within them.
legacy systems
Older information systems that are often incompatible with other systems, technologies, and ways of conducting business. Incompatible legacy systems can be a major roadblock to turning data into information, and they can inhibit firm agility, holding back operational and strategic initiatives.