Stop overkilling simple tasks with black-box models, use more transparent models instead

Matteo Rizzo, Matteo Marcuzzo, Alessandro Zangari, Andrea Gasparetto, Andrea Albarelli

Keywords: Explainable AI, Computer Vision, Natural Language Processing, Transformer, Decision Tree

Abstract: The ability of deep learning-based approaches to extract features autonomously from raw data while outperforming traditional methods has led to several breakthroughs in artificial intelligence. However, it is well-known that deep learning models suffer from an intrinsic opacity, making it difficult to explain why they produce specific predictions. This is problematic not only because it hinders debugging but, most importantly, because it negatively affects the perceived trustworthiness of the systems. What is often overlooked is that many relatively simple tasks can be solved efficiently and effectively with data processing strategies paired with traditional models that are inherently more transparent. This work highlights the frequently neglected perspective of using knowledge-based and explainability-driven problem-solving in ML. To support our guidelines, we propose a simple strategy for solving the task of classifying the ripeness of banana crates. This is done by planning explainability and model design together. We showcase how the task can be solved using opaque deep learning models and more transparent strategies. Notably, there is a minimal loss of accuracy but a significant gain in explainability, which is truthful to the model’s inner workings. Additionally, we perform a user study to evaluate the perception of explainability by end users and discuss our findings.

Primary Area: visualization or interpretation of learned representations

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2024/AuthorGuide.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors' identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 7172