Abstract:Matrices are exceptionally useful in various fields of study as they provide a convenient framework to organize and manipulate data in a structured manner. However, modern matrices can involve billions of elements, making their storage and processing quite demanding in terms of computational resources and memory usage. Although prohibitively large, such matrices are often approximately low rank. We propose an algorithm that exploits this structure to obtain a low rank decomposition of any matrix $\mathbf{A}$ as $\mathbf{A} \approx \mathbf{L}\mathbf{R}$, where $\mathbf{L}$ and $\mathbf{R}$ are the low rank factors. The total number of elements in $\mathbf{L}$ and $\mathbf{R}$ can be significantly less than that in $\mathbf{A}$. Furthermore, the entries of $\mathbf{L}$ and $\mathbf{R}$ are quantized to low precision formats $--$ compressing $\mathbf{A}$ by giving us a low rank and low precision factorization. Our algorithm first computes an approximate basis of the range space of $\mathbf{A}$ by randomly sketching its columns, followed by a quantization of the vectors constituting this basis. It then computes approximate projections of the columns of $\mathbf{A}$ onto this quantized basis. We derive upper bounds on the approximation error of our algorithm, and analyze the impact of target rank and quantization bit-budget. The tradeoff between compression ratio and approximation accuracy allows for flexibility in choosing these parameters based on specific application requirements. We empirically demonstrate the efficacy of our algorithm in image compression, nearest neighbor classification of image and text embeddings, and compressing the layers of LlaMa-$7$b. Our results illustrate that we can achieve compression ratios as aggressive as one bit per matrix coordinate, all while surpassing or maintaining the performance of traditional compression techniques.

Interpolative Butterfly Factorization

Efficient Identification of Butterfly Sparse Matrix Factorizations

Butterfly factorization via randomized matrix-vector multiplications

Butterfly factorization by algorithmic identification of rank-one blocks

Butterfly factorization with error guarantees

Sparse Approximate Multifrontal Factorization with Butterfly Compression for High Frequency Wave Equations

A Linear-complexity Tensor Butterfly Algorithm for Compressing High-dimensional Oscillatory Integral Operators

Multidimensional Phase Recovery and Interpolative Decomposition Butterfly Factorization

Multivariate Polynomial Factorization by Interpolation Method

Low-Rank Matrix Iteration Using Polynomial-Filtered Subspace Extraction.

An Iterative Reweighted Method for Tucker Decomposition of Incomplete Multiway Tensors

A Butterfly-Based Direct Integral-Equation Solver Using Hierarchical LU Factorization for Analyzing Scattering from Electrically Large Conducting Objects

Optimal unit triangular factorization of symplectic matrices

A Butterfly‐based Direct Solver Using Hierarchical LU Factorization for Poggio‐Miller‐Chang‐Harrington‐Wu‐Tsai Equations

"Compress and eliminate" solver for symmetric positive definite sparse matrices

Efficient $1$-bit tensor approximations

Efficient algorithms for Tucker decomposition via approximate matrix multiplication

Matrix Compression via Randomized Low Rank and Low Precision Factorization

Hierarchical interpolative factorization for elliptic operators: differential equations

A Parallelization Technique Based on Factor Combination and Graph Partitioning for General Incomplete LU Factorization

An L-DEIM Induced High Order Tensor Interpolatory Decomposition