Explainable AI in NLP: Interpretable Models for Transparent Decision Making
Gitanjali Shrivastava, Vivek Veeraiah, S. Praveenkumar, Punit Pathak, Tripti Sharma, Ankur Gupta · 2024
In the realm of Natural Language Processing NLP, the (XAI) initiative aims to build interpretable models that provide light on decision-making procedures. A challenge in this field is the development of models that not only perform admirably but also explain their reasoning behind certain forecasts or assessments. The problem arises from the fact that top-notch models are required. s for this kind of study often include topics like feature visualisation, attention processes, and model-agnostic techniques like LIME. Many believe that these approaches will make NLP models easier to understand and work with. Users may have trust in and understanding of the outcomes thanks to XAI in natural language processing, which explains how these models function. This facilitates the incorporation of these models into real-world applications that prioritise openness and interpretability, while also promoting accountability. The study compares the accuracy and error rates of traditional and proposed treatments. The proposed model reduces processing time and time taken for training, while the previous approach was prolonged by unfiltered data. The study also compares the time taken for the proposed approach and the time taken for the conventional approach. The results show that the proposed approach is more efficient and time-efficient than traditional treatments, resulting in better results.