Image preprocessing

Images is nothing more than a two-dimensional array of numbers(or pixels) : it is a matrices of pixel values. Black and…

Artificial Intelligence Techniques

Approaches to artificial intelligence and its history

The Drivetrain Approach to designing great data products

A systematic approach

As data products get more and more complex we sometimes lose sight of the real problem we are trying to solve. The Drivetrain Approach aims to better couple the data…

Crash course on how Numpy Arrays, PyTorch Tensors, PIL, Colab & Computer Images work!

Dataloaders in AI

The Dataloaders task particularly important task in AI system and is typically one amongst the most important roadblocks for the new AI professionals. In this paper, we are going to discuss the dataloder shortly together with the example of Dataloders for the AI model.

Data Ethics Issues in Recourse & feedback loops

Machine Learning Techniques

Within artificial intelligence, machine learning there is two different ways to analyse big amount of data: supervised learning and unsupervised learning. The main difference between those 2 technique is one uses labeled data to help predict outcome, the other one not. But There…

模型构建和实施中的偏见

随着机器学习对人类和社会的影响越来越大，重要的是我们要努力对有害后果的潜在来源有一个全面和统一的理解。例如，下游损害组织经常被归咎于“有偏见的数据”，但是这个概念包含了太多的问题，在开发解决方案中是有用的。

张量和数组

1. 张量是什么？

数组是numpy的一部分，numpy是python里最流行的数组编程库。张量是PyTorch版本的numpy数组。它和数组的代码是完全一样的。在绝大多数情况下，张量和数组的行为几乎相同。但是有一点需要注意的是，张量不仅可以用CPU…