Better Training Data Attribution via Better Inverse Hessian-Vector Products
This work addresses a computational bottleneck in training data attribution for machine learning practitioners, offering an incremental improvement in efficiency and accuracy.
The paper tackled the challenge of efficiently approximating inverse Hessian-vector products for training data attribution, introducing ASTRA, which improved accuracy and reduced iterations compared to existing methods, leading to significant performance gains in TDA.
Training data attribution (TDA) provides insights into which training data is responsible for a learned model behavior. Gradient-based TDA methods such as influence functions and unrolled differentiation both involve a computation that resembles an inverse Hessian-vector product (iHVP), which is difficult to approximate efficiently. We introduce an algorithm (ASTRA) which uses the EKFAC-preconditioner on Neumann series iterations to arrive at an accurate iHVP approximation for TDA. ASTRA is easy to tune, requires fewer iterations than Neumann series iterations, and is more accurate than EKFAC-based approximations. Using ASTRA, we show that improving the accuracy of the iHVP approximation can significantly improve TDA performance.