paper-with-me

홈 › Papers

Scaling Up Influence Functions

2021-12-06 · Andrea Schioppa, Polina Zablotskaia, David Vilar, Artem Sokolov

We address efficient calculation of influence functions for tracking predictions back to the training data. We propose and analyze a new approach to speeding up the inverse Hessian calculation based on Arnoldi iteration. With this improvement, we achieve, to the best of our knowledge, the first successful implementation of influence functions that scales to full-size (language and vision) Transformer models with several hundreds of millions of parameters. We evaluate our approach on image classification and sequence-to-sequence tasks with tens to a hundred of millions of training examples. Our code will be available at https://github.com/google-research/jax-influence.

📄 PDF Abstract BibTeX arXiv:2112.03052

Code (2)

google-research/jax-influence 공식 구현 jax
aai-institute/pyDVL pytorch

Tasks

image-classificationImage Classification

Methods 이 논문이 사용한 방법론

Attention 설명 없음
Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
Absolute Position Encodings Absolute Position Encodings are a type of position embeddings for [Transformer-based models] where positional encodings are…
Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…
Residual Connection 설명 없음
Adam 설명 없음
Dropout Dropout is a regularization technique for neural networks that drops a unit (along with connections) at training time with a specified probability $p$ (a common value is…
Position-Wise Feed-Forward Layer 설명 없음

Similar Papers 제목 키워드 기반

Revisiting inverse Hessian vector products for calculating influence functions

2024-09-25 · Yegor Klochkov, Yang Liu

Influence functions are a popular tool for attributing a model's output to training data. The traditional approach relies on the calculation of inverse Hessian-vector products (iHVP), but the classical solver "Linear tim…

The Central Role of the Loss Function in Reinforcement Learning

2024-09-19 · Kaiwen Wang, Nathan Kallus, Wen Sun

This paper illustrates the central role of loss functions in data-driven decision making, providing a comprehensive survey on their influence in cost-sensitive classification (CSC) and reinforcement learning (RL). We dem…

Decision Makingreinforcement-learningReinforcement LearningReinforcement Learning (RL)

ATHENA: Accelerated Multi-Task Heterogeneous Influence Functions for Robot Data Curation

2026-06-15 · Tao Xu, Jiaxin Wang, Runhao Zhang, Jiayi Guan 외 arxiv

In robot imitation learning, influence functions provide a principled approach to quantify each demonstration's effect on robot task outcomes, yet scaling them to billion-parameter Vision-Language-Action (VLA) models is …

On the Interaction of Adaptive Population Control with Cumulative Step-Size Adaptation

2024-10-01 · Amir Omeradzic, Hans-Georg Beyer

Three state-of-the-art adaptive population control strategies (PCS) are theoretically and empirically investigated for a multi-recombinative, cumulative step-size adaptation Evolution Strategy $(\mu/\mu_I, \lambda)$-CSA-…

The Banach-Butterfly Invariant: Influence-Adaptive Walsh Geometry for Ternary Polynomial Threshold Functions

2026-05-02 · Gorgi Pavlov arxiv

We introduce the Banach-Butterfly Invariant (BBT), an influence-adaptive Banach geometry on the Walsh-Hadamard butterfly factorization. For a Boolean function $f:\{-1,+1\}^n\to\{-1,+1\}$ with coordinate influences $\math…