Self Interpretable

Self-interpretable machine learning models aim to build transparency and understanding directly into the model's architecture and training process, rather than relying on post-hoc explanations. Current research focuses on developing such models across various architectures, including convolutional neural networks and those tailored for specific applications like gigapixel image analysis and time series forecasting (e.g., using generalized additive models). This approach addresses the limitations of post-hoc methods by providing more reliable and user-friendly interpretations, improving trust and facilitating better decision-making in high-stakes domains like healthcare and power grid management. The ultimate goal is to enhance the usability and accountability of complex machine learning systems.

Papers

October 29, 2024

Gnothi Seauton: Empowering Faithful Self-Interpretability in Black-Box Models
Shaobo Wang, Hongxuan Tang, Mingyang Wang, Hongrui Zhang, Xuyang Liu, Weiya Li, Xuming Hu, Linfeng Zhang
Full Model Explainable AI Black Box Black Box Model Self Interpretable Model Self Interpretable

June 30, 2024

Towards Understanding Sensitive and Decisive Patterns in Explainable AI: A Case Study of Model Interpretation in Geometric Deep Learning
Jiajun Zhu, Siqi Miao, Rex Ying, Pan Li
Inherent Interpretability Explainable AI Sensitive Data Geometric Deep Learning Model Interpretation Self Interpretable

December 22, 2023

SI-MIL: Taming Deep MIL for Self-Interpretability in Gigapixel Histopathology
Saarthak Kapse, Pushpak Pati, Srijan Das, Jingwei Zhang, Chao Chen, Maria Vakalopoulou, Joel Saltz, Dimitris Samaras, Rajarsi R. Gupta, Prateek Prasanna
Inherent Interpretability Multiple Instance Learning Instance Learning Global Interpretability Tame Function Self Interpretable

June 14, 2023

SaDI: A Self-adaptive Decomposed Interpretable Framework for Electric Load Forecasting under Extreme Events
Hengbo Liu, Ziqing Ma, Linxiao Yang, Tian Zhou, Rui Xia, Yi Wang, Qingsong Wen, Liang Sun
Interpretable Model Load Forecasting Extreme Event Generalized Additive Model Novel Forecasting Self Interpretable

March 13, 2023

A Test Statistic Estimation-based Approach for Establishing Self-interpretable CNN-based Binary Classifiers
Sourya Sengupta, Mark A. Anastasio
Inherent Interpretability Binary Classifier Estimation Method Self Interpretable Model Test Statistic Self Interpretable

January 27, 2022

LAP: An Attention-Based Module for Concept Based Self-Interpretation and Knowledge Injection in Convolutional Neural Networks
Rassa Ghavami Modegh, Ahmad Salimi, Alireza Dizaji, Hamid R. Rabiee
Convolutional Neural Network Deep Convolutional Neural Network Concept Based Knowledge Injection Attention Based Pooling Self Interpretable

Self Interpretable

Papers

Gnothi Seauton: Empowering Faithful Self-Interpretability in Black-Box Models

Towards Understanding Sensitive and Decisive Patterns in Explainable AI: A Case Study of Model Interpretation in Geometric Deep Learning

SI-MIL: Taming Deep MIL for Self-Interpretability in Gigapixel Histopathology

SaDI: A Self-adaptive Decomposed Interpretable Framework for Electric Load Forecasting under Extreme Events

A Test Statistic Estimation-based Approach for Establishing Self-interpretable CNN-based Binary Classifiers

LAP: An Attention-Based Module for Concept Based Self-Interpretation and Knowledge Injection in Convolutional Neural Networks