Role of explainability in machine learning

Introduction

Laymen explanation

As domains like healthcare look to deploy artificial intelligence and deep learning systems, where questions of accountability and transparency are particularly important, Explaining reason for decision is important aspect.

Current widely used ML algorithms doesn't explain reason. If you are interested to know about this important aspect, this document helps

Technical explanation

Explainability is the extent to which the internal mechanics of a machine or deep learning system can be explained in human terms.

It doesn’t take a data scientist to work out that the machine and deep learning algorithms built into automation and artificial intelligence systems lack transparency.

Why explanation is needed?

A typical example is diagnosing medical conditions. ML models can be produced using data such as X-rays and CT scans, and then be fed with new data and asked to identify whether a medical condition is present or not. In situations like these, however, getting an outcome is not enough: we need to know the explanation behind it, and this is where it gets tricky.

Explainable AI behaviour

Please notice the reason for risk section in below picture. It is produced by the ML.

Why widely used models are not explainable?

Through the introduction of probabilistic learning, applications became increasingly successful, but increasingly opaque.

Impact

The lack of understanding within Artificial Neural Networks has been a large barrier to the adoption of machine learning. This uncertainty comes from the “Black Box” of the neural net.

It is very important for the adoption of machine learning models and ANN’s to be incorporated into fields such as healthcare and finance.

What is the solution?

Explainable AI deals with the implementation of transparency and traceability of statistical black‐box machine learning methods, particularly deep learning (DL).

Current status

It is emerging field. This paper evaluates explainable AI models

Point to remember

Interpretability and explainability are different terms

Reference

https://www.kdnuggets.com/2018/12/machine-learning-explainability-interpretability-ai.html

https://images.app.goo.gl/b2r6Pm6SVK5insAy6

https://images.app.goo.gl/cgzcoQ3xqnLhSqg38

https://images.app.goo.gl/ag4wp1hxeSkFTPhj6

https://onlinelibrary.wiley.com/doi/full/10.1002/widm.1312

https://medium.com/@shagunm1210/the-explainable-neural-network-8f95256dcddb

https://arxiv.org/abs/2005.01831

https://images.app.goo.gl/a4JNVxkbha1LioTSA