Details

Autor(en) / Beteiligte

• Chaturvedi, Manish
• Patel, Pankesh
• Yadav, Ramnarayan

Titel

Explainable AI for ICT: System and Software Architecture

Ist Teil von

• Recent Advancements in ICT Infrastructure and Applications, 2022, p.189-208

Ort / Verlag

Singapore: Springer

Erscheinungsjahr

2022

Link zum Volltext

Quelle

Alma/SFX Local Collection

Beschreibungen/Notizen

• Artificial Intelligence (AI) has become a revolution in the ICT domain due to the swift progress of analytical techniques and the availability of structured/unstructured data. With the indispensable role of AI in different applications, there are growing concerns over the lack of transparency and explainability. In addition, potential bias may affect the predictions of a model. This is where Explainable Artificial Intelligence (XAI) comes into the picture. XAI increases the trust placed in an AI system by researchers, medical practitioners, and others. Thus, it leads to widespread deployment of AI in healthcare, agriculture, online mart, and many more. The aim is to enlighten practitioners on the understandability and interpretability of EAI systems using a variety of techniques available which can be very advantageous in the ICT domain. In this chapter, we present two different techniques leveraging EAI where a user has to make the right choices based on his requirements. The software architecture of the first techniques is based on a medical diagnosis model where we need to be confident enough to treat a patient as instructed by a black-box model. Another approach presents an online Mart where a reliable pricing method can be developed by ML models that can read through historical sales data. The objective here is to match buyers and sellers, to weigh animals, and to oversee their sale. However, when AI models suggest or recommend a decision, that in itself does not reveal too much (i.e., it acts as a black box). Hence, a model capable of explaining the different factors that impact the price point is essential for the needs of a user.