A Case Study in Functional Conversion and Mode Inference in miniKanren
summary
Claims a concrete pipeline that turns miniKanren relations into deterministic functions via functional conversion plus mode inference, implemented as a conversion scheme in hk and evaluated on sorting and permutation generators. Reports a dramatic gap: a slow relational direction taking about 300 seconds on a 5‑element list versus the converted functional version handling up to 262 elements per second; inversion removes search generators and yields single-answer functions for many relations. Shows where conversion fails (nonfunctional predicates) and targets partial deduction and type checks for future work.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 1:12Functional conversion of addition
- 3:21Functional conversion examples
- 5:29Generators in functional conversion
- 7:29Mode inference and conversion
- 10:28Performance comparison
- 12:50Advantages of functional conversion
- 14:28Program inversion and work
- 16:18Discussion on generators
- 17:28Limitations of conversion
- 19:28Converting functions to relations
- 20:50Q&A
related talks
23mTyped Embedding of miniKanren for Functional Conversion
15mAn Empirical Study of Partial Deduction for miniKanren
16mMulti-stage Relational Programming
20mKanren Light
14mOn Fair Relational Conjunction
26mtypedKanren: Statically Typed Relational Programming with Exhaustive Matching in(…)
40mminiKanren Philosophy
14mRelational Synthesis for Pattern Matching