Can Computers Think?

This is PART 1 of a multi-part series on Causal AI. Links to the previous article can be found below:

The immense power of computers to process large quantities of data is one of the critical factors that has enabled the rapid development of artificial intelligence and machine learning in recent years. Modern computers can process vast amounts of data at extremely high speeds, making it possible to analyze and extract insights from large datasets in a short time.

One of the key technologies that has contributed to the increased processing power of computers is the development of parallel processing architectures, which allow computers to perform multiple tasks simultaneously using multiple processors (Gharachorloo, Laudon, & Lenoski, 1990). Increased processing quantity made it possible to simultaneously increase data processing speed by distributing the workload across multiple processors.

In addition to the use of parallel processing, the use of specialized hardware such as graphics processing units (GPUs) has also contributed to the increased processing power of computers (Krizhevsky, Sutskever, & Hinton, 2012). GPUs are designed to perform many simple mathematical operations in parallel, making them particularly well-suited for tasks such as image and video processing.

Increased computing power has dramatically enhanced the capabilities of artificial intelligence (AI). As computing power has increased, the ability to process large amounts of data and perform complex calculations has also increased. Increased computing power has allowed AI algorithms to analyze and learn from data much faster, leading to significant improvements in the performance of AI systems.

For example, in a paper published in the journal Nature, researchers at Google reported that they could train a deep learning algorithm to recognize images with an error rate of only 3.6% using a network with over 1 billion parameters (Silver et al., 2016). This level of accuracy was only made possible by the vast amount of computing power and data available to the researchers.

In addition, the ability to perform complex calculations quickly has allowed AI systems to analyze and make decisions based on vast amounts of data in real time. Real-time analysis has led to the development of AI-powered systems that can assist with decision-making, data analysis, and prediction tasks.

Causal AI, also known as "causal machine learning" or "causal inference," is a subfield of artificial intelligence that focuses on using data to identify causal relationships between different variables (Pearl, 2018). Causal machine learning is in contrast to traditional machine learning techniques, which primarily focuses on predicting the value of a target variable based on a set of input variables without necessarily considering the underlying causal relationships (Hernán, Robins, & Brumback, 2020).