Abstract
This extended abstract accompanies an invited talk at CASC 2024, which surveys recent developments in Real Quantifier Elimination (QE) and Cylindrical Algebraic Decomposition (CAD). After introducing these concepts we will first consider adaptations of CAD inspired by computational logic, in particular the algorithms which underpin modern SAT solvers. CAD theory has found use in collaboration with these via the Satisfiability Modulo Theory (SMT) paradigm; while the ideas behind SAT/SMT have led to new algorithms for Real QE. Second we will consider the optimisation of CAD through the use of Machine Learning (ML). The choice of CAD variable ordering has become a key case study for the use of ML to tune algorithms in computer algebra. We will also consider how explainable AI techniques might give insight for improved computer algebra software without any reliance on ML in the final code.
Original language | English |
---|---|
Title of host publication | Computer Algebra in Scientific Computation |
Editors | François Boulier, Chenqi Mou, Timur M. Sadykov, Evgenii V. Vorozhtsov |
Publisher | Springer |
Chapter | 1 |
Pages | 1-10 |
Number of pages | 10 |
Volume | 14938 |
ISBN (Electronic) | 978-3-031-69070-9 |
ISBN (Print) | 978-3-031-69069-3 |
DOIs | |
Publication status | Published - 21 Aug 2024 |
Event | 26th International Computer Algebra in Scientific Computing - Rennes, France Duration: 2 Sept 2024 → 6 Sept 2024 https://casc-conference.org/ |
Publication series
Name | Lecture Notes in Computer Science |
---|---|
Publisher | Springer |
ISSN (Print) | 0302-9743 |
ISSN (Electronic) | 1611-3349 |
Conference
Conference | 26th International Computer Algebra in Scientific Computing |
---|---|
Abbreviated title | CASC 2024 |
Country/Territory | France |
City | Rennes |
Period | 2/09/24 → 6/09/24 |
Internet address |
Funding
The author is currently supported by UKRI EPSRC grant EP/T015748/1: Pushing Back the Doubly-Exponential Wall of Cylindrical Algebraic Decomposition (the DEWCAD Project). Some of the work surveyed was supported by UKRI EPSRC grant EP/R019622/1, Embedding Machine Learning within Quantifier Elimination Procedures; and EU H2020 FET CSA 712689, Satisfiability Checking and Symbolic Computation (the SC Project).
Funders | Funder number |
---|---|
Engineering and Physical Sciences Research Council | EP/T015748/1, EP/R019622/1 |
European Horizon 2020 | FET CSA 712689 |
Keywords
- Quantifier elimination
- Cylindrical algebraic decomposition
- SAT
- SMT
- Machine learning
- Explainable AI