Rigorous roundoff error analysis of probabilistic floating-point computations
From MaRDI portal
Publication:832297
DOI10.1007/978-3-030-81688-9_29zbMATH Open1493.65087arXiv2105.13217OpenAlexW3186933279MaRDI QIDQ832297FDOQ832297
Publication date: 25 March 2022
Abstract: We present a detailed study of roundoff errors in probabilistic floating-point computations. We derive closed-form expressions for the distribution of roundoff errors associated with a random variable, and we prove that roundoff errors are generally close to being uncorrelated with their generating distribution. Based on these theoretical advances, we propose a model of IEEE floating-point arithmetic for numerical expressions with probabilistic inputs and an algorithm for evaluating this model. Our algorithm provides rigorous bounds to the output and error distributions of arithmetic expressions over random variables, evaluated in the presence of roundoff errors. It keeps track of complex dependencies between random variables using an SMT solver, and is capable of providing sound but tight probabilistic bounds to roundoff errors using symbolic affine arithmetic. We implemented the algorithm in the PAF tool, and evaluated it on FPBench, a standard benchmark suite for the analysis of roundoff errors. Our evaluation shows that PAF computes tighter bounds than current state-of-the-art on almost all benchmarks.
Full work available at URL: https://arxiv.org/abs/2105.13217
Recommendations
Cites Work
- Certification of bounds on expressions involving rounded operators
- dReal: An SMT Solver for Nonlinear Theories over the Reals
- An Abstract Interpretation Framework for the Round-Off Error Analysis of Floating-Point Programs
- Title not available (Why is that?)
- Title not available (Why is that?)
- Title not available (Why is that?)
- Accuracy and Stability of Numerical Algorithms
- Title not available (Why is that?)
- Numerical inverting of matrices of high order
- Semantics of probabilistic programs
- A generalization of \(p\)-boxes to affine arithmetic
- Title not available (Why is that?)
- A guide to Monte Carlo simulations in statistical physics.
- Rigorous Estimation of Floating-Point Round-off Errors with Symbolic Taylor Expansions
- A New Approach to Probabilistic Rounding Error Analysis
- Uncertainty propagation using probabilistic affine forms and concentration of measure inequalities
- Certified Roundoff Error Bounds Using Semidefinite Programming
- Probabilistic Error Analysis for Inner Products
- Rigorous floating-point mixed-precision tuning
- Sound compilation of reals
Cited In (10)
- Probability Analysis of Round-Off Errors in Floating-Point Arithmetic
- Title not available (Why is that?)
- Stochastic Rounding Variance and Probabilistic Bounds: A New Approach
- A priori worst case error bounds for floating-point computations
- On roundoff error distributions in floating point and logarithmic arithmetic
- Title not available (Why is that?)
- Probabilistic analysis of floating-point addition
- Precision-aware deterministic and probabilistic error bounds for floating point summation
- Title not available (Why is that?)
- Probabilistic Error Analysis for Inner Products
Uses Software
This page was built for publication: Rigorous roundoff error analysis of probabilistic floating-point computations
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q832297)
