Update importance.rst

docs/source/_xai/importance.rst

@@ -18,7 +18,7 @@ For Science
 
 The goal of scientific discovery is to understand. Data collection and analysis is a core element of scientific methods, and scientists have long used statistical techniques to aid their work. As early as the 1900s the development of the t-test gave researchers a new tool to extract insights from data to test the veracity of their hypotheses. Today, machine learning has become a vital tool for researchers across domains to analyze large datasets, detect previously unforeseen patterns, or extract unexpected insights. 
 
-For example, machine learning is used in areas like computational biology to analyze genomic data or to predict the three-dimensional structures of proteins, and in computational climate science to understand the effects of climate change on cities and regions (e.g. by combining local observational data to large-scale climate models, researchers hope to acquire more detailed pictures of the local impacts of climate change), in computational physics to find patterns in vast amounts of astronomical data that can be very noisy data.
+For example, machine learning is used in areas like computational biology to analyze genomic data or to predict the three-dimensional structures of proteins, in computational climate science to understand the effects of climate change on cities and regions (e.g. by combining local observational data to large-scale climate models, researchers hope to acquire more detailed pictures of the local impacts of climate change), or in computational physics to find patterns in vast amounts of astronomical data that can be very noisy data.
 
 In some contexts, the accuracy of these methods alone is sufficient to make AI useful – filtering telescope observations to identify likely targets for further study, for example. However, researchers want to know not just what the answer is but why! Explainable AI can help researchers to understand the insights that come from research data, by providing accessible explanations of which features have been particularly relevant in the creation of the system’s output. A project that is working on explainable AI for science is, for example, the Automated Statistician project. They have created a system that can generate an explanation of its forecasts or predictions, by breaking complicated datasets into interpretable sections and explaining its findings in accessible language. Together, the machine learning system and the explanations help researchers analyze large amounts of data and enhance their understanding of the features of that data (*The Royal Society, 2019*).