Post Hoc

Post-hoc analysis in machine learning refers to techniques applied *after* a model is trained to improve its performance, interpretability, or robustness. Current research focuses on enhancing model explainability using methods like LIME and SHAP, improving out-of-distribution detection by combining existing methods, and addressing issues like data privacy risks associated with explanations. These advancements are crucial for building trustworthy and reliable AI systems, particularly in high-stakes applications like healthcare and autonomous systems, by increasing transparency and mitigating potential biases or vulnerabilities.

Papers

October 30, 2023

Uncertainty Quantification in Machine Learning Based Segmentation: A Post-Hoc Approach for Left Ventricle Volume Estimation in MRI
F. Terhag, P. Knechtges, A. Basermann, R. Tempone
Uncertainty Quantification Learning Based Biomedical Image Segmentation Post Hoc Robust Segmentation Left Ventricular Volume Estimation

October 25, 2023

This Reads Like That: Deep Learning for Interpretable Natural Language Processing
Claudio Fanconi, Moritz Vandenhirtz, Severin Husmann, Julia E. Vogt
Deep Learning Natural Language Processing Read V Prototypical Network Post Hoc Prototype Learning Pre Trained Sentence

July 27, 2023

Discriminative Feature Attributions: Bridging Post Hoc Explainability and Inherent Interpretability
Usha Bhalla, Suraj Srinivas, Himabindu Lakkaraju
Inherent Interpretability Interpretable Model Black Box Model Post Hoc Distractor Generation

July 1, 2023

CMA-ES for Post Hoc Ensembling in AutoML: A Great Success and Salvageable Failure
Lennart Purucker, Joeran Beel
Ensemble Model Online autoML Financial Success Post Hoc CMA E Greedy Learning Auto Sklearn AutoML Benchmark

May 24, 2023

How to Fix a Broken Confidence Estimator: Evaluating Post-hoc Methods for Selective Classification with Deep Neural Networks
Luís Felipe P. Cattelan, Danilo Silva
Deep Neural Network Confidence Estimation Post Hoc Selective Classification Confidence Estimator

May 22, 2023

If at First You Don't Succeed, Try, Try Again: Faithful Diffusion-based Text-to-Image Generation by Selection
Shyamgopal Karthik, Karsten Roth, Massimiliano Mancini, Zeynep Akata
Text to Image Generation View Selection Generative Process Post Hoc First Attempt T2I Diffusion Model Diffusion Based Text to Image Faithful Model

May 18, 2023

SimOAP: Improve Coherence and Consistency in Persona-based Dialogue Generation via Over-sampling and Post-evaluation
Junkai Zhou, Liang Pang, Huawei Shen, Xueqi Cheng
Language Model Pre Trained Language Model Open Domain Dialogue Discourse Coherence Post Hoc Persona Based Dialogue Sampling Algorithm End to End Response Generation

February 27, 2023

CrystalBox: Future-Based Explanations for Input-Driven Deep RL Systems
Sagar Patel, Sangeetha Abdu Jyothi, Nina Narodytska
Deep Reinforcement Learning Deep RL Post Hoc Quality Explanation Explainability Technique Future Based Explanation

October 21, 2022

Turning Fixed to Adaptive: Integrating Post-Evaluation into Simultaneous Machine Translation
Shoutao Guo, Shaolei Zhang, Yang Feng
Adaptive Importance View Translation Translation Task Simultaneous Machine Translation Post Hoc Row Turning Adaptive Policy

June 29, 2022

Causality for Inherently Explainable Transformers: CAT-XPLAIN
Subash Khanal, Benjamin Brodie, Xin Xing, Ai-Ling Lin, Nathan Jacobs
Causal Pattern Post Hoc Explanation Explainable Model Post Hoc

February 14, 2022

Counterfactual inference for sequential experiments
Raaz Dwivedi, Katherine Tian, Sabina Tomkins, Predrag Klasnja, Susan Murphy, Devavrat Shah
Estimation Task Counterfactual Inference Post Hoc Counterfactual Feature Sequential Experiment

January 25, 2022

Post-Hoc Explanations Fail to Achieve their Purpose in Adversarial Contexts
Sebastian Bordt, Michèle Finck, Eric Raidl, Ulrike von Luxburg
Line by Line Explanation High Explainability Post Hoc Explanation Post Hoc Explanation Capability Adversarial Context

Post Hoc

Papers

Uncertainty Quantification in Machine Learning Based Segmentation: A Post-Hoc Approach for Left Ventricle Volume Estimation in MRI

This Reads Like That: Deep Learning for Interpretable Natural Language Processing

Discriminative Feature Attributions: Bridging Post Hoc Explainability and Inherent Interpretability

CMA-ES for Post Hoc Ensembling in AutoML: A Great Success and Salvageable Failure

How to Fix a Broken Confidence Estimator: Evaluating Post-hoc Methods for Selective Classification with Deep Neural Networks

If at First You Don't Succeed, Try, Try Again: Faithful Diffusion-based Text-to-Image Generation by Selection

SimOAP: Improve Coherence and Consistency in Persona-based Dialogue Generation via Over-sampling and Post-evaluation

CrystalBox: Future-Based Explanations for Input-Driven Deep RL Systems

Turning Fixed to Adaptive: Integrating Post-Evaluation into Simultaneous Machine Translation

Causality for Inherently Explainable Transformers: CAT-XPLAIN

Counterfactual inference for sequential experiments

Post-Hoc Explanations Fail to Achieve their Purpose in Adversarial Contexts