Define Overfitting and Underfitting by yourself

“If you are just got started in Machine Learning then you may have heard of these two terms often. And if not then you must.

Both the terms have the suffix ‘fitting’ and the interpretation of this word is same as in real life. Like suppose the dress which is not of fitting size can be expressed in two ways i.e. tight or loose.

In Maths we solve problems by implementing formulas/Algorithms and measures how good it is by accuracy, no. of steps,etc. the same way in Machine Learning we implement models(in fact you will find out later that they are also formulas) and we measures how good it is by accuracy, training time and many other things. And if our model is not doing good then we say it is either Underfitting or Overfitting.”



It’s the Learning That Matters

“Deep learning empowers better forecasting and better integration of relevant data. Production and demand forecasting improve dramatically by neural networks when added to more traditional analytical methods such as regression and classification techniques. This was true for 69 percent of all AI use cases we identified in our study. In fact, in only 16 percent of use cases did we find a “greenfield” AI solution that applied where other analytical methods would not be sufficient or relevant.”

AI technology is not just an experiment

“Most said the reason wasn’t that they wanted to keep their AI activities secret, but that they weren’t actually very far along and hence their projects were not worth discussing yet. They were doing lots of pilots, proofs of concept, and prototypes, but they had few production deployments. When they did have AI systems in production, most were machine learning-based systems that had been in place for many years. This is particularly true in financial services, where large-scale “scoring” has been used to evaluate customers for credit and potential fraud for well over a decade. Some said to us that they didn’t really consider these projects to be examples of AI — consistent with the common view of AI that it describes technology that is never really here yet. Others say that they have robotic process automation (RPA) implementations in place, but most are relatively small, and there is also debate about whether RPA is really AI or not.”

Como esta nova tecnologia baseada em Inteligência Artificial pode ser um benefício para a privacidade

“O novo sistema utiliza uma técnica de deep learning chamada de Generative Adversarial Networks (GANs), que coloca dois algoritmos de IA um contra o outro. A equipe projetou um conjunto de duas redes neurais: a primeira trabalhando para identificar rostos e a segunda trabalhando para interromper a tarefa de reconhecimento facial da primeira. Os dois estão constantemente lutando e aprendendo uns com os outros, estabelecendo uma competição contínua. O resultado é um filtro semelhante ao Instagram, que pode ser aplicado a fotos para proteger a privacidade.”

Veja mais em:

Uma introdução ao Deep Learning para dados tabulares

“Há uma técnica poderosa que está ganhando competições de Kaggle e é amplamente utilizada no Google (de acordo com Jeff Dean), Pinterest e Instacart, mas que muitas pessoas nem percebem que é possível: o uso de deep learning para dados tabulares, e em particular, a criação de incorporações para variáveis categóricas. Apesar do que você pode ter ouvido, você pode usar o aprendizado profundo para o tipo de dados que você pode manter em um banco de dados SQL, um Pandas DataFrame ou uma planilha do Excel (incluindo dados de série temporal). Refiro a isto como dados tabulares, embora também possa ser conhecido como dados relacionais, dados estruturados ou outros termos.”

Veja mais em: