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

June 16, 2022

Benchmarking Heterogeneous Treatment Effect Models through the Lens of Interpretability
Jonathan Crabbé, Alicia Curth, Ioana Bica, Mihaela van der Schaar
Inherent Interpretability Camera Lens Treatment Effect Heterogeneous Treatment Effect Heterogeneous Treatment Effect Estimation

June 14, 2022

June 12, 2022

A Functional Information Perspective on Model Interpretation
Itai Gat, Nitay Calderon, Roi Reichart, Tamir Hazan
Inherent Interpretability Deep Network Model Interpretability Fisher Information Model Interpretation Differential Entropy

June 9, 2022

June 7, 2022

On the balance between the training time and interpretability of neural ODE for time series modelling
Yakov Golovanev, Alexander Hvatov
Time Series Inherent Interpretability Neural ODE Differential Equation Balancing Strategy Data Driven Modeling Training Time

June 5, 2022

Interpretable Mixture of Experts
Aya Abdelsalam Ismail, Sercan Ö. Arik, Jinsung Yoon, Ankur Taly, Soheil Feizi, Tomas Pfister
Inherent Interpretability Expert Knowledge Interpretable Model Model Explanation

June 1, 2022

Assessing the trade-off between prediction accuracy and interpretability for topic modeling on energetic materials corpora
Monica Puerto, Mason Kellett, Rodanthi Nikopoulou, Mark D. Fuge, Ruth Doherty, Peter W. Chung, Zois Boukouvalas
Inherent Interpretability Energy Policy Research Topic Detection Prediction Accuracy Energy Data Energetic Material

May 31, 2022

GlanceNets: Interpretabile, Leak-proof Concept-based Models
Emanuele Marconato, Andrea Passerini, Stefano Teso
Representation Learning Inherent Interpretability Glance Annotation Interpretable Way High Level Single Learning Concept Based Model

May 27, 2022

Neural Basis Models for Interpretability
Filip Radenovic, Abhimanyu Dubey, Dhruv Mahajan
Inherent Interpretability Single Neural Network Generalized Additive Model Complex Machine Learning

May 25, 2022

May 24, 2022

May 23, 2022

May 21, 2022

Scaling Laws and Interpretability of Learning from Repeated Data
Danny Hernandez, Tom Brown, Tom Conerly, Nova DasSarma, Dawn Drain, Sheer El-Showk, Nelson Elhage, Zac Hatfield-Dodds, Tom Henighan, Tristan Hume, Scott Johnston, Ben Mann, Chris Olah, Catherine Olsson, Dario Amodei, Nicholas Joseph, Jared Kaplan, Sam McCandlish
Large Language Model Strong Generalization Inherent Interpretability Scaling Law Information Redundancy Recent Large Language Model Exercise Repetition

May 20, 2022

Inherent Interpretability

Papers

Benchmarking Heterogeneous Treatment Effect Models through the Lens of Interpretability

Can Foundation Models Talk Causality?

Combining Counterfactuals With Shapley Values To Explain Image Models

A Functional Information Perspective on Model Interpretation

Process Knowledge-Infused AI: Towards User-level Explainability, Interpretability, and Safety

There is no Accuracy-Interpretability Tradeoff in Reinforcement Learning for Mazes

On the balance between the training time and interpretability of neural ODE for time series modelling

Interpretable Mixture of Experts

Assessing the trade-off between prediction accuracy and interpretability for topic modeling on energetic materials corpora

GlanceNets: Interpretabile, Leak-proof Concept-based Models

Neural Basis Models for Interpretability

On the Interpretability of Regularisation for Neural Networks Through Model Gradient Similarity

Logical Satisfiability of Counterfactuals for Faithful Explanations in NLI

Improving Shape Awareness and Interpretability in Deep Networks Using Geometric Moments

Functional Network: A Novel Framework for Interpretability of Deep Neural Networks

Towards Opening the Black Box of Neural Machine Translation: Source and Target Interpretations of the Transformer

A Fine-grained Interpretability Evaluation Benchmark for Neural NLP

Scaling Laws and Interpretability of Learning from Repeated Data

B-cos Networks: Alignment is All We Need for Interpretability

Constructive Interpretability with CoLabel: Corroborative Integration, Complementary Features, and Collaborative Learning