Research map / Learning theory and algorithms / Combinatorial and search optimization
Novel Is Not Always Better: On the Relation between Novelty and Dominance Pruning
Joschka Gross, Alvaro Torralba, Maximilian Fickert
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use LD-simulation dominance relation instead of identity in IW | IW(k) · IPC single-goal tasks | effective width 1: 40.42% -> 40.35% | No clear effect |
| use LD-simulation dominance instead of identity | IW(2) · IPC | expansions | Improved |
| use unsafe relation LD2 instead of LD (ablation) | IW(k) · IPC single-goal tasks | effective width >2: 13.83% -> 22.21% | Hurt |
| use unsafe relation LD1 (ablation) | IW(k) · IPC single-goal tasks | effective width >2: 13.83% -> 41.87% | Hurt |
| use unsafe relation LD0 (ablation) | IW(k) · IPC single-goal tasks | effective width >2: 13.83% -> 76.18% | Hurt |
| use inverse relation LDinv (ablation) | IW(k) · IPC single-goal tasks | effective width >2: 13.83% -> 50.34% | Hurt |
| use LD-simulation instead of identity relation | hQB · IPC coverage | coverage | Hurt |
| use LD-simulation instead of identity relation | hQB · IPC domains | expansions | Mixed |
| use new priority for non-novel states with atomic projections | hQB · IPC | coverage | Improved |
| use new non-novel priority with larger projections Qpre | hQB · IPC | coverage | No clear effect |
| replace atomic projections Q1 with Qpre (ablation) | hQB · IPC | total coverage: 1564 -> 1634 | Improved |
| replace Q1 with Qcg (ablation) | hQB · IPC | total coverage: 1564 -> 1630 | Improved |
Closest papers by meaning
- Novelty vs. Potential Heuristics: A Comparison of Hardness Measures for Satisficing Planning (AAAI 2024 · similarity 0.81)
- Novelty Detection in Reinforcement Learning with World Models (ICML 2025 · similarity 0.80)
- Evaluating and Understanding Scheming Propensity in LLM Agents (NeurIPS 2026 · similarity 0.80)
- How new data permeates LLM knowledge and how to dilute it (ICLR 2025 · similarity 0.80)
- Rejecting Hallucinated State Targets during Planning (ICML 2025 · similarity 0.79)
- Implicit Search via Discrete Diffusion: A Study on Chess (ICLR 2025 · similarity 0.79)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
