Origin of Causality
The Birth of Causal Inference
Last updated
The Birth of Causal Inference
Last updated
“Be a free thinker and don’t accept everything you hear as truth. Be critical and evaluate what you believe in....” ~Aristotle
The critical intellectual distinction between humans and machines is their ability to define goals and reason towards achieving them. Understanding cause and effect is a big aspect of what we call common sense. It is the ability of humans to interpret and process information that makes humans dominant on this planet.
The process of experimentations, accepting, and rejecting hypotheses to reach conclusions needs to be accepted and implemented by the AI models. Over time researchers have agreed that modeling and reasoning about the interventions can help the current Machine Learning techniques to understand and resolve the current problems and take the field of Artificial Intelligence to the Next Level!!!
AI research began in the 1950s. Today, Artificial Intelligence is considered to be the New Revolution. Today the most important resource is not oil, but Data. For the efficient and smart use of this data or information, we need AI. Starting from the Industrial Revolution to the AI Revolution there has been a concern about the ethical use of these new coming technologies
“The last 10 years have been about building a world that is mobile-first. In the next 10 years, we will shift to a world that is AI-first.” - Sundar Pichai, CEO of Google
Statisticians have been considering and researching on causality for a very long time. Statisticians started working on it in 1920. Causality was recognized as another area of research since the 1970s. Causal Inference has done immense progress in the past 30 years!!
Mainly there are 2 Frameworks :
Rubin Causal Model (RCM) by Rubin-Neyman - Potential Outcomes Framework **Control Group Experiments
Structural Causal Model (SCM) by Judea Perl - Causal Graphical Models, Causal Diagrams and Do-Calculus
Both of these frameworks have different usage but overall focus on the same concept!!
Potential Outcome Framework has been really popular amongst statisticians. Donald Rubin has done a lot of work in developing the Potential Outcomes Framework. This started with more of a statistical approach earlier. Later, Judea Perl started working on formalizing a new world of Causal Inference. He did a lot Jof work earlier in Bayesian Networks and worked on the graphical approach to formalize causality. Then he came up with the Structural Causal Model(SCM) which provides a clear picture of mathematical tools to be used for causal conclusions. Judea Perl has contributed a lot in developing this field of Causal Inference and stating its importance in the field of Artificial Intelligence. His recent work "The Book of Why", states how understanding causality has revolutionized science and will revolutionize Artificial Intelligence. His approach has ignited a discussion about the future of AI, and whether deep learning and other breakthroughs are enough for approaching the human-level intelligence.
There has been a Causal Revolution - from Statistics to Counterfactual !!
The practice of causal Inference is well established in the medical field where it is termed as Medical Trials. The testing of most new medications is done through this process of medical trials, which is proven to be a very effective method of testing a causal relationship between the treatment and outcome. Randomized controlled trials used here are considered high in the pyramid of evidence.
In recent years the field of causal inference has grown in scope and impact. Today, it's not only Judea Perl, but also researchers of MIT, and various other universities that agree on the need of causal models for human-like intelligence in machines. - link
Causal Inference is now making its way into Machine Learning and Artificial Intelligence as people are increasingly talking about this significant research area. Also, after all these years there are authentic connections between Machine Learning and Causality, which are considered very crucial for growth in the field of Artificial Intelligence.