Efficient Estimation of Influence of a Training Instance

2020-12-08 · EMNLP (sustainlp) 2020 11 · Sosuke Kobayashi, Sho Yokoi, Jun Suzuki, Kentaro Inui

Understanding the influence of a training instance on a neural network model leads to improving interpretability. However, it is difficult and inefficient to evaluate the influence, which shows how a model's prediction would be changed if a training instance were not used. In this paper, we propose an efficient method for estimating the influence. Our method is inspired by dropout, which zero-masks a sub-network and prevents the sub-network from learning each training instance. By switching between dropout masks, we can use sub-networks that learned or did not learn each training instance and estimate its influence. Through experiments with BERT and VGGNet on classification datasets, we demonstrate that the proposed method can capture training influences, enhance the interpretability of error predictions, and cleanse the training dataset for improving generalization.

📄 PDF Abstract BibTeX arXiv:2012.04207

Code (0)

등록된 구현이 없습니다.

Methods 이 논문이 사용한 방법론

Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.

Interpretability 설명 없음

Linear Warmup With Linear Decay Linear Warmup With Linear Decay is a learning rate schedule in which we increase the learning rate linearly for $n$ updates and then linearly decay afterwards.

WordPiece 설명 없음

Multi-Head Attention 설명 없음

Residual Connection 설명 없음

Adam 설명 없음

Dense Connections Dense Connections, or Fully Connected Connections, are a type of layer in a deep neural network that use a linear operation where every input is connected to every output…

Identifying a Training-Set Attack's Target Using Renormalized Influence Estimation

2022-01-25 · Zayd Hammoudeh, Daniel Lowd

Targeted training-set attacks inject malicious instances into the training set to cause a trained model to mislabel one or more specific test instances. This work proposes the task of target identification, which determi…

Influence Estimation for Generative Adversarial Networks

2021-01-20 · ICLR 2021 1 · Naoyuki Terashita, Hiroki Ohashi, Yuichi Nonaka, Takashi Kanemaru

Identifying harmful instances, whose absence in a training dataset improves model performance, is important for building better machine learning models. Although previous studies have succeeded in estimating harmful inst…

Training Data Influence Analysis and Estimation: A Survey

2022-12-09 · Zayd Hammoudeh, Daniel Lowd

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 de…

Survey

Instance Influence Estimation for Hyperspectral Target Signature Characterization using Extended Functions of Multiple Instances

2016-03-21 · Sheng Zou, Alina Zare

The Extended Functions of Multiple Instances (eFUMI) algorithm is a generalization of Multiple Instance Learning (MIL). In eFUMI, only bag level (i.e. set level) labels are needed to estimate target signatures from mixed…

Multiple Instance Learning

Benchmarking and Error Diagnosis in Multi-Instance Pose Estimation

2017-07-17 · ICCV 2017 10 · Matteo Ruggero Ronchi, Pietro Perona

We propose a new method to analyze the impact of errors in algorithms for multi-instance pose estimation and a principled benchmark that can be used to compare them. We define and characterize three classes of errors - l…

Benchmarking Pose Estimation