Training Data Influence Analysis and Estimation: A Survey

Abstract

Good models require good training data. For overparameterized deep models, the causal relationship between training data and model predictions is increasingly opaque and poorly understood. Influence analysis partially demystifies training's underlying interactions by quantifying the amount each training instance alters the final model. Measuring the training data's influence exactly can be provably hard in the worst case; this has led to the development and use of influence estimators, which only approximate the true influence. This paper provides the first comprehensive survey of training data influence analysis and estimation. We begin by formalizing the various, and in places orthogonal, definitions of training data influence. We then organize state-of-the-art influence analysis methods into a taxonomy; we describe each of these methods in detail and compare their underlying assumptions, asymptotic complexities, and overall strengths and weaknesses. Finally, we propose future research directions to make influence analysis more useful in practice as well as more theoretically and empirically sound. A curated, up-to-date list of resources related to influence analysis is available at https://github.com/ZaydH/influence_analysis_papers.

arXiv commentPublished in Springer journal "Machine Learning"

References in corpus (44)

• An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
• A Unified Approach to Interpreting Model Predictions
• Stochastic Backpropagation and Approximate Inference in Deep Generative Models
• Kernel methods in machine learning
• Understanding Black-box Predictions via Influence Functions
• Understanding deep learning requires rethinking generalization
• A Survey on the Explainability of Supervised Machine Learning
• Poisoning Attacks against Support Vector Machines
• Benign Overfitting in Linear Regression
• Sketching as a Tool for Numerical Linear Algebra
• Towards Efficient Data Valuation Based on the Shapley Value
• Data Shapley: Equitable Valuation of Data for Machine Learning
• What Neural Networks Memorize and Why: Discovering the Long Tail via Influence Estimation
• Why Do Adversarial Attacks Transfer? Explaining Transferability of Evasion and Poisoning Attacks
• Coresets via Bilevel Optimization for Continual Learning and Streaming
• On the Accuracy of Influence Functions for Measuring Group Effects
• A Survey of Machine Unlearning
• Neuron Shapley: Discovering the Responsible Neurons
• Deduplicating Training Data Mitigates Privacy Risks in Language Models
• Identifying Mislabeled Data using the Area Under the Margin Ranking
• Second-Order Stochastic Optimization for Machine Learning in Linear Time
• Coresets for Data-efficient Training of Machine Learning Models
• HYDRA: Hypergradient Data Relevance Analysis for Interpreting Deep Neural Networks
• "Influence Sketching": Finding Influential Samples In Large-Scale Regressions
• Rank List Sensitivity of Recommender Systems to Interaction Perturbations
• Beta Shapley: a Unified and Noise-reduced Data Valuation Framework for Machine Learning
• From Shapley Values to Generalized Additive Models and back
• Influence-guided Data Augmentation for Neural Tensor Completion
• Identifying a Training-Set Attack's Target Using Renormalized Influence Estimation
• Improved Certified Defenses against Data Poisoning with (Deterministic) Finite Aggregation
• Rethinking Influence Functions of Neural Networks in the Over-parameterized Regime
• On Second-Order Group Influence Functions for Black-Box Predictions
• Data Banzhaf: A Robust Data Valuation Framework for Machine Learning
• RIM: Reliable Influence-based Active Learning on Graphs
• Adapting and Evaluating Influence-Estimation Methods for Gradient-Boosted Decision Trees
• TracInAD: Measuring Influence for Anomaly Detection
• Characterizing Structural Regularities of Labeled Data in Overparameterized Models
• Run-Off Election: Improved Provable Defense against Data Poisoning Attacks
• Nondeterminism and Instability in Neural Network Optimization
• Provable Robustness Against a Union of ${\ell }_0$ Adversarial Attacks
• On Collective Robustness of Bagging Against Data Poisoning
• Measuring the Effect of Training Data on Deep Learning Predictions via Randomized Experiments
• Influence Estimation for Generative Adversarial Networks
• Representer Point Selection for Explaining Regularized High-dimensional Models

Cited by in corpus (4)

• Creative Writers' Attitudes on Writing as Training Data for Large Language Models
• Assessing the Efficacy of Deep Learning Approaches for Facial Expression Recognition in Individuals with Intellectual Disabilities
• Leveraging Influence Functions for Resampling Data in Physics-Informed Neural Networks
• SwarmSearch: Decentralized Search Engine with Self-Funding Economy