$\nabla$SD: Differentiable Programming for Sparse Tensors

TL;DR

This paper introduces $ abla$SD, a new differentiable programming framework that efficiently handles sparse tensors, overcoming existing limitations related to irregular sparsity patterns and enabling scalable gradient computations in various data-intensive applications.

Contribution

The paper presents a novel framework for automatic differentiation of sparse tensors, improving performance and scalability over existing methods.

Findings

Outperforms state-of-the-art frameworks in efficiency and scalability.

Demonstrates effectiveness on synthetic and real-world datasets.

Enables scalable differentiable programming with sparse tensors.

Abstract

Sparse tensors are prevalent in many data-intensive applications, yet existing differentiable programming frameworks are tailored towards dense tensors. This presents a significant challenge for efficiently computing gradients through sparse tensor operations, as their irregular sparsity patterns can result in substantial memory and computational overheads. In this work, we introduce a novel framework that enables the efficient and automatic differentiation of sparse tensors, addressing this fundamental issue. Our experiments demonstrate the effectiveness of the proposed framework in terms of performance and scalability, outperforming state-of-the-art frameworks across a range of synthetic and real-world datasets. Our approach offers a promising direction for enabling efficient and scalable differentiable programming with sparse tensors, which has significant implications for numerous applications in machine learning, natural language processing, and scientific computing.