Local Search

Follow this topic Ask about this topic

Momentum

0 papers in the last four weeks, with none the four weeks before. 0.0% of all new papers.

Jul 13Week of Sep 28

Latest papers 12

All topics

Fork-and-Flush: Escaping Idea Basins in Autoresearch Agents
Oct 5, 2026Ziyang Cai, Christos Ziakas, Vasilis Kontonis +5Local Search Long-Horizon Agent Tasks
Microsoft Research
Two-Point Local Optimality in $k$ -Means via Boundary-Point Screening
Oct 5, 2026Wenlong Lyu, Xujie Xiao, Yuheng JiaClustering Local Search
School of Computer Science and Engineering Southeast University Nanjing, Jiangsu 211189, China
DE-2LS: Differential Evolution with Lightweight Late Local Search for Constrained Numerical Optimization
Jun 26, 2026Dikshit Chauhan, Anupam TrivediDifferential Evolution Evolutionary Optimization
Department of Electrical and Computer Engineering, National University of Singapore
DE-2LS: Differential Evolution with Late-Stage local-search for Unconstrained Single-Objective Numerical Optimization
Jun 26, 2026Dikshit ChauhanDifferential Evolution Evolutionary Optimization
Department of Electrical and Computer Engineering, National University of Singapore
Local Search on Vertex Coloring for Bipartite Graphs
Jun 8, 2026Johanna GasseCombinatorial Optimization Local Search
Fachgebiet Algorithm Engineering der Digital-Engineering-Fakultät der Universität Potsdam
Leveraging Structural Constraints for Diffusion-based Neural TSP Solvers
Jun 8, 2026Mickaël Basson, Philippe PreuxTraveling Salesperson Problem Inference-Time Optimization
CRIStAL, Scool · Université de Lille, France · CNRS, France +2
Aperon Technical Report: Hierarchical No-Pointer Tangent-Local Search for High-Dimensional Approximate Nearest Neighbors
Jun 7, 2026Yong FuLocal Search Approximate Nearest Neighbor Search
Substratum Labs
Combinatorial Landscape Analysis for Dominating Set and Vertex Coloring
Jun 5, 2026Johanna Gasse, Antonia Heinen, Felix Knöfel +2Combinatorial Optimization Local Search
Hasso Plattner Institute, Potsdam, Germany
Local Preferential Bayesian Optimization
Jun 1, 2026Johanna Menn, Miriam Kober, Paul Brunzema +2Pairwise Preference Learning Pairwise Preference Evaluation
Institute for Data Science in Mechanical Engineering, RWTH Aachen University, Aachen, Germany · Department of Clinical Research, University of Bern, Bern, Switzerland · Center for Reproducible Science and Research Synthesis, University of Zurich, Zurich, Switzerland +1
Transforming Constraint Programs to Input for Local Search
May 19, 2026Jo Devriendt, Patrick De Causmaecker, Marc DeneckerCombinatorial Optimization Local Search
University of Leuven
A penalised Saito functional for heuristic search of free line arrangements
Apr 3, 2026Tomás S. R. SilvaLocal Search
Instituto de Matem´atica, Estat´ıstica e Computa¸c˜ao Cient´ıfica (IMECC), Universidade Estadual de Campinas (UNICAMP), 13083-859, Campinas, S˜ao Paulo, Brazil.
Unbiased Binning for Fairness-aware Attribute Representation
Sep 26, 2025Abolfazl Asudeh, Zeinab Asoodeh, Bita Asoodeh +1Dynamic Programming Algorithmic Fairness
University of Illinois Chicago · Independent Researcher · University of Edinburgh +1

Oct 5, 2026, cs.LGJ/K move · Enter open · S save

Fork-and-Flush: Escaping Idea Basins in Autoresearch Agents

Ziyang Cai, Christos Ziakas, Vasilis Kontonis, Tim Pearce+4

Microsoft Research

Local Search Long-Horizon Agent Tasks Autonomous Experimentation Autonomous Research Agents

Open paper AskPDF arXiv

Autoresearch agents tackle open-ended problems by repeatedly proposing candidate solutions, evaluating them, and using feedback to guide subsequent experiments. We show that independent runs of the same agent on the same task often plateau at substantially different scores, with gaps that persist even after considerable additional compute. Embedding their candidate artifacts by functional similarity provides further evidence that trajectories remain in localized regions of the solution space, which we call idea basins. To help agents escape these basins, we study a simple periodic intervention, fork-and-flush. Our method forks the agent into parallel trajectories, each inheriting the accumulated workspace but starting with a fresh chat context. After running each trajectory for a fixed horizon, the agent continues from the highest-scoring one. Across 13 long-horizon research and engineering tasks, with individual agent runs lasting up to several days, fork-and-flush outperformed the single-run and best-of-N baselines by a relative improvement of 66.0% and 44.4%, respectively, on the min-max normalized average score under an equal compute budget.

Momentum

Related topics

Latest papers 12