Category: Artificial intelligence

Deep learning and artificial intelligence have invaded our daily lives. Evidence that these techniques are so widespread is before our eyes through success stories such as digital assistants and self-driving cars.

Deep learning techniques have their roots in previous centuries. In recent decades there have been some really important evolutions that have led to the algorithms we use today. Let’s learn about the evolutions that led to today’s technology.

Several methodologies and algorithms that are part of deep learning have been developed in recent years. But are these techniques actually that modern? Let’s look at a bit of history, starting from the Middle Ages and arriving at the present day, to understand the roots of deep learning and artificial intelligence.

Supervised methods require that our example data also provide us with the labels we should predict. However, these learning methods are limited. In some cases we either do not know what we are looking for or the environment itself can give us hints as to how our model should evolve. Unsupervised and reinforcement learning methods address these types of problems.

In addition to classical classification and regression methods, there are other approaches and fields in which machine learning is used. Indeed, we can assign different labels to the same sample, create models to improve search in our applications, or suggest highly relevant content based on an individual user’s profile. Finally, we can build models that use data sequences to produce new sequences, such as translations or conversion from audio to text and vice versa.

When we need to analyze data, we have several techniques at our disposal. In deep learning, but more generally in data science, we can make use of some techniques called superivision learning. These consist of looking at some example data to predict values and/or labels on new data.

Data scientists who want to use machine learning and/or deep learning techniques need to be clear about a few key concepts: data, models, objective functions, and optimization algorithms. Let’s analyze them in detail to understand how to use them in our projects.

Deep learning is a tool that has invaded everyday life. Many of the applications we use daily are based on models built with specific techniques that accumulate experience by looking at the data available to us. Let’s find out what it is and how it affects the use and experience of our applications.

What are the NLP models that have made history? In this guide we will discover some of these as well as an overview of the most widely used libraries for developing models. Finally, we will cover some controversies that have arisen in this field.

Natural language processing (NLP) allows us to create systems that can interpret what we write. But how are the data underlying these models processed? And what techniques are most commonly used? In this guide we will look at these issues.