As much as if you use the AI of Sam Altman’s company, as if you have used other chatbots such as Google Gemini or the interesting Grok of Elon Musk’s company, You are likely to find some words that you have not heard before And it is that the world of cognitive computing is broad.
As in all technology, There are terms that can become commonespecially if you frequently use this type of tools. For example, machine learning, DeepSearchmodel and several more things that probably leave you with doubts.
If you have come up with any and do not understand it, you don’t have to worry, because sooner or later you must learn them. However, making part of your knowledge from the beginning can be the key to preparing for the evolution of artificial intelligence.
Important terms related to artificial intelligence that you should know
Unplash
At AF as chatgpt you can ask you many things, how to know your opinion about whether Elon Musk is good or bad and even Create Ghibli style images with the GPT 4.o model. Every time a new update arrives at these projects, it is innovated in various areas, so it is necessary to adapt to this revolution.
There are many types of artificial intelligences, since some are dedicated to certain activities, such as Operator who is an AI or Sora agent that specifically serves to create videos. Over the years, the development of these characteristics has been expanded, making it more concepts that define practices, elements or actions.
Leaving aside the synonyms of artificial intelligence, such as AI, Chatbot and Cognitive Computing, as well as well -known OpenAi figures such as Sam Altman, Ilya Sutskever and Greg Brockman, in this list you have some terms that is undoubtedly worth knowing:
- IA debil o limited: It is a chatbot that is only responsible for making a specific action, such as one that generates text or one that generates only images. Most of the current platforms are of this type, although it also depends on the power of the implemented model.
- General Artificial Intelligence (AGI): It is a theoretical objective that engineers want to arrive, which would be the most advanced AI. If it is achieved, it would be autonomous, intellectual, with reasoning just like human beings and automatic learning.
- Singularity: It is another concept that represents the evolution of AI to the point where it is able to overcome by much human intelligence, practically as it is a new life with conscience equal to that of an organic being.
- Autonomous agents: They are artificial intelligence systems created to function autonomously, with repetitive tasks or certain activities. For example, as the technology that is implemented in an autonomous car such as the Tesla Robotaxi.
- Ethics: It is a field in which moral regulations, values and all the characteristics related to ethics are determined so that it is not used inappropriately and adapts to the rules of society.
- Algorithm: They are the instructions that the model follows depending on its programming. With this you can analyze parameters or prompts to continue learning and improve, as well as offer the appropriate results.
- Multimodal: A chatbot that has several entry functions, such as attaching text, images, videos or audio to give you results. In the same way, that these formats can generate with applications.
- Red neuronal: It is the composition of AI, a process behind the AI that simulates the human brain with the characteristics of patterns and data necessary to work.
- Parameters: They are elements, values or processes of a model.
- Alignment: It refers to adjusting the operation of the model until you can follow the desired instructions in the scope of programming.
- Hallucination: When the chatbot responds incorrectly despite the good use of the prompt and/or deviates from the subject completely. You could talk about Elon Musk’s Optimus robots, but I would give you answers that have nothing to do, as something related to a movie like My neighbor Totoro.
- Bias: They are errors that occur when training information is applied. The AI grants results that are consistent with stereotypes of races, genres, cultures or human groups, which can be very good or very bad.
- Emerging behavior: suge when developing an unwanted ability, response or action.
- Fast Onset of Overwhelming Mastery (FOOM): In Spanish, fast takeoff or forced takeoff. It is a concept that defines a situation in which AI becomes powerful so fast that it would later be very difficult to stop its evolution, such as it, for example.
- Tokens: It refers to a word, symbol, character or other text elements that are processed to formulate the phrases or indications. They are essential to be able to interact.
- Overjustment: problem that happens when the model adapts too well to automatic learning parameters, but preventing good things from learning. Therefore, it only recognizes specific data, instead of the generalized ones.
- Large Language Model (LLM): A model that uses advanced neural networks that allow it to better understand and generate natural conversations similar to that of a human.
- Automatic learning (ML): Technique attributed to artificial intelligences so that they can learn new things without requiring programmed or added data manually.
- Deep Learning): It is also part of automatic learning and focuses on recognizing patterns in various formats through the data included in the neuronal network.
- Zero shooting learning (ZSL): It is the characteristic that allows the model to identify and organize elements or definitions without having previous direct knowledge with them.
- End -to -end learning (E2E): This is another automatic learning process that consists of performing a complete task without intervention.
- Antagonistic generative networks (GAN): It is a type of AI model that works with two neuronal networks, a so -called «generator» and another «discriminatory.» Both focus on a competition technique to constantly improve.
- Natural language processing (PLN or NLP): It is a field that focuses on making machines can understand, generate and use human language through various processes.
- Stochastic parrot: It is a concept that refers to the fact that the functioning of AI is based on patterns or existing information, being unable to actually understand what they say. «Loro» because he only repeats things without understanding them.
- Turing test: Test created by one of the AI parents, Alan Turing. It is done to determine if a machine can replicate the way humans act.
These are just some of the relevant things that revolve around artificial intelligence. At first glance they are too many and complicated to remember, but over time you get used to it.
It doesn’t matter if you study to develop models such as Google AI (Gemini), Anthropologic (Claude), Openai (Chatgpt), Microsoft (Copilot) and IBM (Watson Assistant) or if you are only a curious user, With these terms you will expand your knowledge.
In addition, it is good that you continue to learn new things that are related to this technology, since little by little more concepts will emerge that are probably equally important.
Know How we work in NoticiasVE.
Tags: Artificial Intelligence, Curiosities