Black Box Model

Black box models, characterized by their opaque internal workings, pose challenges in understanding their decision-making processes, hindering trust and accountability. Current research focuses on improving interpretability through methods like generalized additive models (GAMs) and surrogate models, as well as addressing vulnerabilities to adversarial attacks and biases through techniques such as explanation-driven attacks and robust defense mechanisms. This work is crucial for building trust in AI systems across various applications, from medical diagnosis to autonomous driving, by enhancing transparency and mitigating potential risks associated with unpredictable model behavior.

Papers

April 13, 2023

Certified Zeroth-order Black-Box Defense with Robust UNet Denoiser
Astha Verma, Siddhesh Bangar, A V Subramanyam, Naman Lal, Rajiv Ratn Shah, Shin'ichi Satoh
Adversarial Perturbation Black Box Model Black Box Defense Residual UNet

March 25, 2023

Learning with Explanation Constraints
Rattana Pukdee, Dylan Sam, J. Zico Kolter, Maria-Florina Balcan, Pradeep Ravikumar
LeArning Abstract Line by Line Explanation Black Box Model Large Deep Learning Model

March 21, 2023

March 15, 2023

Interpretable Ensembles of Hyper-Rectangles as Base Models
Andrei V. Konstantinov, Lev V. Utkin
Black Box Model Ensemble Model Base Model Interpretable Ensemble Effective Ensemble

March 8, 2023

Learning Hybrid Interpretable Models: Theory, Taxonomy, and Methods
Julien Ferry, Gabriel Laberge, Ulrich Aïvodji
NCD Method Theoretical Understanding Comprehensive Taxonomy Interpretable Model Black Box Model Hybrid Model Interpretable Hybrid Hybrid Reasoning

March 7, 2023

Bootstrap The Original Latent: Learning a Private Model from a Black-box Model
Shuai Wang, Daoan Zhang, Jianguo Zhang, Weiwei Zhang, Rui Li
Black Box Model Intermediate Latent Private Model Model Privacy

March 2, 2023

Towards Trustable Skin Cancer Diagnosis via Rewriting Model's Decision
Siyuan Yan, Zhen Yu, Xuelin Zhang, Dwarikanath Mahapatra, Shekhar S. Chandra, Monika Janda, Peter Soyer, Zongyuan Ge
Deep Neural Network Full Model Black Box Model Skin Lesion Decision Relevant Information Model Inference Melanoma Diagnosis

February 17, 2023

Black-Box Batch Active Learning for Regression
Andreas Kirsch
Active Learning Novel Regression Black Box Model Regression Task Batch Active Learning

February 15, 2023

February 6, 2023

Stop overkilling simple tasks with black-box models and use transparent models instead
Matteo Rizzo, Matteo Marcuzzo, Alessandro Zangari, Andrea Gasparetto, Andrea Albarelli
Deep Learning Artificial Intelligence Machine Learning Model Black Box Model Model Transparency

January 13, 2023

Uncertainty Quantification for Local Model Explanations Without Model Access
Surin Ahn, Justin Grana, Yafet Tamene, Kristian Holsheimer
Uncertainty Quantification Black Box Model Post Hoc Explanation Confidence Interval Local Interpretable Model Agnostic Explanation Explanation Uncertainty Model Sharing

December 18, 2022

December 16, 2022

Interpretable models for extrapolation in scientific machine learning
Eric S. Muckley, James E. Saal, Bryce Meredig, Christopher S. Roper, John H. Martin
Interpretable Model Model Performance Black Box Model Extrapolation Framework Scientific Machine Learning Trend Extrapolation Poor Extrapolation Performance

December 8, 2022

XRand: Differentially Private Defense against Explanation-Guided Attacks
Truc Nguyen, Phung Lai, NhatHai Phan, My T. Thai
Explainable AI Black Box Model RGB Image Local Differential Privacy Backdoor Trigger Explanation Guided

November 24, 2022

ML Interpretability: Simple Isn't Easy
Tim Räz
Machine Learning Model Inherent Interpretability Explainable AI Black Box Model

November 17, 2022

Supervised Feature Compression based on Counterfactual Analysis
Veronica Piccialli, Dolores Romero Morales, Cecilia Salvatore
Pre Trained Model Counterfactual Explanation Black Box Model Counterfactual Analysis Feature Compression

November 14, 2022

Global Performance Guarantees for Neural Network Models of AC Power Flow
Samuel Chevalier, Spyros Chatzivasileiadis
Neural Network Model Black Box Model Neural Network Verification Power Flow Constraint Tightening Global Guarantee