\n<\/p>\n
Artificial intelligence is a deep and convoluted world. The scientists who work in this field often rely on jargon and lingo to explain what they\u2019re working on. As a result, we frequently have to use those technical terms in our coverage of the artificial intelligence industry. That\u2019s why we thought it would be helpful to put together a glossary with definitions of some of the most important words and phrases that we use in our articles.<\/p>\n
We will regularly update this glossary to add new entries as researchers continually uncover novel methods to push the frontier of artificial intelligence while identifying emerging safety risks.<\/p>\n
Artificial general intelligence, or AGI, is a nebulous term. But it generally refers to AI that\u2019s more capable than the average human at many, if not most, tasks. OpenAI CEO Sam Altman\u00a0recently<\/a>\u00a0described AGI as the \u201cequivalent of a median human that you could hire as a co-worker.\u201d Meanwhile,\u00a0OpenAI\u2019s charter<\/a>\u00a0defines AGI as \u201chighly autonomous systems that outperform humans at most economically valuable work.\u201d Google DeepMind\u2019s understanding differs slightly from these two definitions; the lab views AGI as \u201cAI that\u2019s at least as capable as humans at most cognitive tasks.\u201d Confused? Not to worry \u2014\u00a0so are experts at the forefront of AI research<\/a>.<\/p>\n An AI agent refers to a tool that uses AI technologies to perform a series of tasks on your behalf \u2014 beyond what a more basic AI chatbot could do \u2014 such as filing expenses, booking tickets or a table at a restaurant, or even writing and maintaining code. However, as we\u2019ve explained before<\/a>, there are lots of moving pieces in this emergent space, so \u201cAI agent\u201d might mean different things to different people. Infrastructure is also still being built out to deliver on its envisaged capabilities. But the basic concept implies an autonomous system that may draw on multiple AI systems to carry out multistep tasks.<\/p>\n Given a simple question, a human brain can answer without even thinking too much about it \u2014 things like \u201cwhich animal is taller, a giraffe or a cat?\u201d But in many cases, you often need a pen and paper to come up with the right answer because there are intermediary steps. For instance, if a farmer has chickens and cows, and together they have 40 heads and 120 legs, you might need to write down a simple equation to come up with the answer (20 chickens and 20 cows).<\/p>\n In an AI context, chain-of-thought reasoning for large language models means breaking down a problem into smaller, intermediate steps to improve the quality of the end result. It usually takes longer to get an answer, but the answer is more likely to be correct, especially in a logic or coding context. Reasoning models are developed from traditional large language models and optimized for chain-of-thought thinking thanks to reinforcement learning.<\/p>\n (See: Large language model<\/a>)<\/p>\n A subset of self-improving machine learning in which AI algorithms are designed with a multi-layered, artificial neural network (ANN) structure. This allows them to make more complex correlations compared to simpler machine learning-based systems, such as linear models or decision trees. The structure of deep learning algorithms draws inspiration from the interconnected pathways of neurons in the human brain.<\/p>\n Deep learning AI models are able to identify important characteristics in data themselves, rather than requiring human engineers to define these features. The structure also supports algorithms that can learn from errors and, through a process of repetition and adjustment, improve their own outputs. However, deep learning systems require a lot of data points to yield good results (millions or more). They also typically take longer to train compared to simpler machine learning algorithms \u2014 so development costs tend to be higher.<\/p>\n (See: Neural network<\/a>)<\/p>\n Diffusion is the tech at the heart of many art-, music-, and text-generating AI models. Inspired by physics, diffusion systems slowly \u201cdestroy\u201d the structure of data<\/a> \u2014 e.g. photos, songs, and so on \u2014 by adding noise until there\u2019s nothing left. In physics, diffusion is spontaneous and irreversible \u2014 sugar diffused in coffee can\u2019t be restored to cube form. But diffusion systems in AI aim to learn a sort of \u201creverse diffusion\u201d process to restore the destroyed data, gaining the ability to recover the data from noise.<\/p>\n Distillation is a technique used to extract knowledge from a large AI model with a \u2018teacher-student\u2019 model. Developers send requests to a teacher model and record the outputs. Answers are sometimes compared with a dataset to see how accurate they are. These outputs are then used to train the student model, which is trained to approximate the teacher\u2019s behavior.<\/p>\n Distillation can be used to create a smaller, more efficient model based on a larger model with a minimal distillation loss. This is likely how OpenAI developed GPT-4 Turbo, a faster version of GPT-4.<\/p>\n While all AI companies use distillation internally, it may have also been used by some AI companies to catch up with frontier models. Distillation from a competitor usually violates<\/a> the terms of service of AI API and chat assistants.<\/p>\n This refers to the further training of an AI model to optimize performance for a more specific task or area than was previously a focal point of its training \u2014 typically by feeding in new, specialized (i.e., task-oriented) data.\u00a0<\/p>\n Many AI startups are taking large language models as a starting point to build a commercial product but are vying to amp up utility for a target sector or task by supplementing earlier training cycles with fine-tuning based on their own domain-specific knowledge and expertise.<\/p>\n (See: Large language model [LLM]<\/a>)<\/p>\n A GAN, or Generative Adversarial Network, is a type of machine learning framework that underpins some important developments in generative AI when it comes to producing realistic data \u2013 including (but not only) deepfake tools. GANs involve the use of a pair of neural networks, one of which draws on its training data to generate an output that is passed to the other model to evaluate. This second, discriminator model thus plays the role of a classifier on the generator\u2019s output \u2013 enabling it to improve over time.\u00a0<\/p>\n The GAN structure is set up as a competition (hence \u201cadversarial\u201d) \u2013 with the two models essentially programmed to try to outdo each other: the generator is trying to get its output past the discriminator, while the discriminator is working to spot artificially generated data. This structured contest can optimize AI outputs to be more realistic without the need for additional human intervention. Though GANs work best for narrower applications (such as producing realistic photos or videos), rather than general purpose AI.<\/p>\n Hallucination is the AI industry\u2019s preferred term for AI models making stuff up \u2013 literally generating information that is incorrect. Obviously, it\u2019s a huge problem for AI quality.\u00a0<\/p>\n Hallucinations produce GenAI outputs that can be misleading and could even lead to real-life risks \u2014 with potentially dangerous consequences (think of a health query that returns harmful medical advice). This is why most GenAI tools\u2019 small print now warns users to verify AI-generated answers, even though such disclaimers are usually far less prominent than the information the tools dispense at the touch of a button.<\/p>\n The problem of AIs fabricating information is thought to arise as a consequence of gaps in training data. For general purpose GenAI especially \u2014 also sometimes known as foundation models \u2014 this looks difficult to resolve. There is simply not enough data in existence to train AI models to comprehensively resolve all the questions we could possibly ask. TL;DR: we haven\u2019t invented God (yet).\u00a0<\/p>\n Hallucinations are contributing to a push towards increasingly specialized and\/or vertical AI models \u2014 i.e. domain-specific AIs that require narrower expertise \u2013 as a way to reduce the likelihood of knowledge gaps and shrink disinformation risks.<\/p>\n Inference is the process of running an AI model. It\u2019s setting a model loose to make predictions or draw conclusions from previously-seen data. To be clear, inference can\u2019t happen without training; a model must learn patterns in a set of data before it can effectively extrapolate from this training data.<\/p>\n Many types of hardware can perform inference, ranging from smartphone processors to beefy GPUs to custom-designed AI accelerators. But not all of them can run models equally well. Very large models would take ages to make predictions on, say, a laptop versus a cloud server with high-end AI chips.<\/p>\n [See: Training<\/a>]<\/p>\n Large language models, or LLMs, are the AI models used by popular AI assistants, such as ChatGPT<\/a>, Claude<\/a>, Google\u2019s Gemini<\/a>, Meta\u2019s AI Llama<\/a>, Microsoft Copilot<\/a>, or Mistral\u2019s Le Chat<\/a>. When you chat with an AI assistant, you interact with a large language model that processes your request directly or with the help of different available tools, such as web browsing or code interpreters.<\/p>\n AI assistants and LLMs can have different names. For instance, GPT is OpenAI\u2019s large language model and ChatGPT is the AI assistant product.<\/p>\n LLMs are deep neural networks made of billions of numerical parameters (or weights, see below<\/a>) that learn the relationships between words and phrases and create a representation of language, a sort of multidimensional map of words.<\/p>\n These models are created from encoding the patterns they find in billions of books, articles, and transcripts. When you prompt an LLM, the model generates the most likely pattern that fits the prompt. It then evaluates the most probable next word after the last one based on what was said before. Repeat, repeat, and repeat.<\/p>\n (See: Neural network<\/a>)<\/p>\n A neural network refers to the multi-layered algorithmic structure that underpins deep learning \u2014 and, more broadly, the whole boom in generative AI tools following the emergence of large language models.\u00a0<\/p>\n Although the idea of taking inspiration from the densely interconnected pathways of the human brain as a design structure for data processing algorithms dates all the way back to the 1940s, it was the much more recent rise of graphical processing hardware (GPUs) \u2014 via the video game industry \u2014 that really unlocked the power of this theory. These chips proved well suited to training algorithms with many more layers than was possible in earlier epochs \u2014 enabling neural network-based AI systems to achieve far better performance across many domains, including voice recognition, autonomous navigation, and drug discovery.<\/p>\n (See: Large language model [LLM]<\/a>)<\/p>\n Developing machine learning AIs involves a process known as training. In simple terms, this refers to data being fed in in order that the model can learn from patterns and generate useful outputs.<\/p>\n Things can get a bit philosophical at this point in the AI stack \u2014 since, pre-training, the mathematical structure that\u2019s used as the starting point for developing a learning system is just a bunch of layers and random numbers. It\u2019s only through training that the AI model really takes shape. Essentially, it\u2019s the process of the system responding to characteristics in the data that enables it to adapt outputs towards a sought-for goal \u2014 whether that\u2019s identifying images of cats or producing a haiku on demand.<\/p>\n It\u2019s important to note that not all AI requires training. Rules-based AIs that are programmed to follow manually predefined instructions \u2014 for example, such as linear chatbots \u2014 don\u2019t need to undergo training. However, such AI systems are likely to be more constrained than (well-trained) self-learning systems.<\/p>\n Still, training can be expensive because it requires lots of inputs \u2014 and, typically, the volumes of inputs required for such models have been trending upwards.<\/p>\n Hybrid approaches can sometimes be used to shortcut model development and help manage costs. Such as doing data-driven fine-tuning of a rules-based AI \u2014 meaning development requires less data, compute, energy, and algorithmic complexity than if the developer had started building from scratch.<\/p>\n [See: Inference<\/a>]<\/p>\n A technique where a previously trained AI model is used as the starting point for developing a new model for a different but typically related task \u2013 allowing knowledge gained in previous training cycles to be reapplied.\u00a0<\/p>\n Transfer learning can drive efficiency savings by shortcutting model development. It can also be useful when data for the task that the model is being developed for is somewhat limited. But it\u2019s important to note that the approach has limitations. Models that rely on transfer learning to gain generalized capabilities will likely require training on additional data in order to perform well in their domain of focus<\/p>\n