PLDI 2026 (series) / EGRAPHS 2026 (series) / EGRAPHS / Predicate E-Graphs with Symbolic Conditional Rewriting
Predicate E-Graphs with Symbolic Conditional RewritingRemote
Mon 15 Jun 2026 09:46 - 10:10 at Meadows CD - Session 1
To represent the possible conditionally optimised versions of a program term, we define predicate e-graphs as an extension of coloured e-graphs. In predicate e-graphs: 1) Colours are associated with sets of Boolean terms denoting assumptions, and the directed graph structure on colours is compatible with the implication structure on these sets. 2) Colours form a rooted directed graph instead of a tree, meaning that one colour can have multiple direct “parents”. 3) The directed graph structure can add edges dynamically. 4) Colours can be identified, requiring the corresponding e-graphs to be merged and identified.
This extended abstract articulates the core ideas, invariants, and operations of the model.
| Presentation slides (26.EGRAPHS-predicateEgraphs.pdf) | 1.0MiB |
| Extended Abstract (Verified_Optimization_by_Conditional_Rewriting (11).pdf) | 344KiB |
Mon 15 JunDisplayed time zone: Mountain Time (US & Canada) change
Mon 15 Jun
Displayed time zone: Mountain Time (US & Canada) change
09:00 - 10:10 | |||
09:00 23mTalk | From Rewriting to Fixpoints: Solving Recursive Equations with E-Graphs EGRAPHS Samuel Coward University College London (UCL), Cheng Zhang Worcester Polytechnic Institute, Philip Zucker Draper, Alexandra Silva Cornell University | ||
09:23 23mTalk | Lifting E-Graphs: A Function Isn’t a Constant EGRAPHS Philip Zucker Draper Pre-print File Attached | ||
09:46 23mTalk | Predicate E-Graphs with Symbolic Conditional RewritingRemote EGRAPHS Anders Ågren Thuné Uppsala University, Johannes Borgström Uppsala University, Sweden, Lars-Henrik Eriksson Uppsala University, John Högberg Ericsson, Tjark Weber Uppsala University, Tobias Wrigstad Uppsala University File Attached | ||