MINI GUÍA
Una guía visual de la A a la Z para la palabra de moda de la década "IA".
Things are changing again. AI is bringing in a new era, and yes, it’s got its own set of tricky words. We’re not just about choosing the right colours anymore; it’s time to get smart with tech words! We can’t just be the 'can we have it green and on the right?' squad. We need to level up!
But we've been here before.. we learned to design while keeping business needs in mind, thinking about the KPIs, the outcomes, increasing retention and much more. We learned these terms. Now it's just another dictionary to help us share our ideas in terms that the techies understand. It’s like learning a new language just to order your favourite food without any unwanted surprises!
Tokens
Unidades básicas que pueden ser codificadas. Las palabras se convierten en lenguaje que las máquinas entienden.
Modelo de Lenguaje Grande o LLMs
Okay, everyone is using this abbreviation left and right! It stands for Large Language Models. Huh?
What that means is that the language we humans speak, AI can now train on that vast amount of data and even generate. It's essentially limited to everything we do through conversations, through language (which is basically everything!)
In short, AI hacked the Operating System of a human = LLM.
Autoatención
Para entender la relación entre las palabras. Cada palabra se valora en función de otras palabras.
GPT 4 vs Hugging face vs Llama 2?
Aprendizaje supervisado
Start with labelled data set.
Eg: Analyzing a picture & having a human expert label it as a cat or a dog. Then separating it into training and validation data set.
Train the machine on the pattern which you know is correct, then,
Machine predicts if it’s a cat or a dog.
Aprendizaje no supervisado
Unlike the previous approach, in unsupervised leaning, the algorithm finds “natural” groupings, without known outcomes or labels.
It finds observations and patterns that may not be obvious to human being.
Aprendizaje por refuerzo (RL)
You don’t start with a data set or try to recognize patterns, like the previous two, you just have a starting point and a performance metric.
Start somewhere, probe around (explore), chose one (exploit), check if the performance improved or worsened, then continue or go back to try again.
Coming soon.
Subscribe above to get notified.
aiverse.design