Inherent Interpretability

Inherent interpretability in machine learning focuses on designing models and methods that are inherently transparent and understandable, aiming to reduce the "black box" nature of many AI systems. Current research emphasizes developing intrinsically interpretable model architectures, such as those based on decision trees, rule-based systems, and specific neural network designs (e.g., Kolmogorov-Arnold Networks), alongside techniques like feature attribution and visualization methods to enhance understanding of model behavior. This pursuit is crucial for building trust in AI, particularly in high-stakes applications like healthcare and finance, where understanding model decisions is paramount for responsible deployment and effective human-AI collaboration.

Papers

September 29, 2023

PRIME: Prioritizing Interpretability in Failure Mode Extraction
Keivan Rezaei, Mehrdad Saberi, Mazda Moayeri, Soheil Feizi
Inherent Interpretability Interpretable Feature Image Classification Model Failure Analysis Interpretability Research Essence Prime Language

September 28, 2023

Navigating Healthcare Insights: A Birds Eye View of Explainability with Knowledge Graphs
Satvik Garg, Shivam Parikh, Somya Garg
Knowledge Graph Inherent Interpretability High Explainability Eye View Medical AI Explainable AI Model Knowledge Infused Learning

September 25, 2023

Prior Bilinear Based Models for Knowledge Graph Completion
Jiayi Li, Ruilin Luo, Jiaqi Sun, Jing Xiao, Yujiu Yang
Inherent Interpretability Knowledge Graph Completion Posterior Mode Bilinear Model

September 24, 2023

September 23, 2023

Towards LLM-guided Causal Explainability for Black-box Text Classifiers
Amrita Bhattacharjee, Raha Moraffah, Joshua Garland, Huan Liu
Natural Language Processing Inherent Interpretability High Explainability Black Box Counterfactual Explanation Explainable Causal

September 21, 2023

Impact of architecture on robustness and interpretability of multispectral deep neural networks
Charles Godfrey, Elise Bishoff, Myles McKay, Eleanor Byler
Deep Learning Native Robustness Deep Learning Model Global Impact Inherent Interpretability Architecture Design Fusion Method Fusion Strategy

September 18, 2023

Interpretability is not Explainability: New Quantitative XAI Approach with a focus on Recommender Systems in Education
Riccardo Porcedda
Machine Learning Inherent Interpretability Recommender System High Explainability Human Understanding XAI Method Human Driving Focus eXplainable Artificial Intelligence

September 16, 2023

Extracting Interpretable Local and Global Representations from Attention on Time Series
Leonid Schwenke, Martin Atzmueller
Time Series Inherent Interpretability High Explainability Human Attention Interpretability Method Transformer Attention Global Representation Local Interpretability

September 15, 2023

Sparse Autoencoders Find Highly Interpretable Features in Language Models
Hoagy Cunningham, Aidan Ewart, Logan Riggs, Robert Huben, Lee Sharkey
Language Model Inherent Interpretability Supervised Autoencoder Mechanistic Interpretability Interpretable Feature Sparse Autoencoders Polysemous Word

September 14, 2023

September 11, 2023

Experimenting with UD Adaptation of an Unsupervised Rule-based Approach for Sentiment Analysis of Mexican Tourist Texts
Olga Kellert, Mahmud Uz Zaman, Nicholas Hill Matlis, Carlos Gómez-Rodríguez
Inherent Interpretability Sentiment Analysis Unsupervised Setting UNsupervised Approach Universal Dependency Sentiment Lexicon State of the Art Unsupervised

September 7, 2023

Automatic Concept Embedding Model (ACEM): No train-time concepts, No issue!
Rishabh Jain
Full Model Inherent Interpretability High Explainability Proper Issue Labeling Semantic Annotation Concept Embeddings

September 5, 2023

A Context-Sensitive Approach to XAI in Music Performance
Nicola Privato, Jack Armitage
Inherent Interpretability High Explainability Explainable Artificial Intelligence Context Aware xAI Community Music Performance

September 1, 2023

TExplain: Explaining Learned Visual Features via Pre-trained (Frozen) Language Models
Saeid Asgari Taghanaki, Aliasghar Khani, Ali Saheb Pasand, Amir Khasahmadi, Aditya Sanghi, Karl D. D. Willis, Ali Mahdavi-Amiri
Language Model Inherent Interpretability Vision Model Visual Feature Image Classifier Pre Trained Convolutional Neural Network

August 31, 2023

Unsupervised discovery of Interpretable Visual Concepts
Caroline Mazini Rodrigues, Nicolas Boutry, Laurent Najman
Inherent Interpretability XAI Method Unsupervised Discovery Activation Maximization Global Interpretability

August 29, 2023

August 28, 2023

Attention Visualizer Package: Revealing Word Importance for Deeper Insight into Encoder-Only Transformer Models
Ala Alam Falaki, Robin Gras
Inherent Interpretability Self Attention Encoder Only Transformer Attention Visualization Word Importance

Inherent Interpretability

Papers

PRIME: Prioritizing Interpretability in Failure Mode Extraction

Navigating Healthcare Insights: A Birds Eye View of Explainability with Knowledge Graphs

Prior Bilinear Based Models for Knowledge Graph Completion

A Text Classification-Based Approach for Evaluating and Enhancing the Machine Interpretability of Building Codes

I-AI: A Controllable & Interpretable AI System for Decoding Radiologists' Intense Focus for Accurate CXR Diagnoses

Towards LLM-guided Causal Explainability for Black-box Text Classifiers

Impact of architecture on robustness and interpretability of multispectral deep neural networks

Interpretability is not Explainability: New Quantitative XAI Approach with a focus on Recommender Systems in Education

Extracting Interpretable Local and Global Representations from Attention on Time Series

Sparse Autoencoders Find Highly Interpretable Features in Language Models

Interpretability-Aware Vision Transformer

Interpretability is in the Mind of the Beholder: A Causal Framework for Human-interpretable Representation Learning

Experimenting with UD Adaptation of an Unsupervised Rule-based Approach for Sentiment Analysis of Mexican Tourist Texts

Automatic Concept Embedding Model (ACEM): No train-time concepts, No issue!

A Context-Sensitive Approach to XAI in Music Performance

TExplain: Explaining Learned Visual Features via Pre-trained (Frozen) Language Models

Unsupervised discovery of Interpretable Visual Concepts

Probabilistic Dataset Reconstruction from Interpretable Models

How Faithful are Self-Explainable GNNs?

Attention Visualizer Package: Revealing Word Importance for Deeper Insight into Encoder-Only Transformer Models