{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/arxiv-2601-18005","title":"Flow-based Extremal Mathematical Structure Discovery","arxiv_id":"2601.18005","date":"2026-01-25","proceeding":null,"authors":["Gergely Bérczi","Baran Hashemi","Jonas Klüver"],"abstract":"The discovery of extremal structures in mathematics requires navigating vast and nonconvex landscapes where analytical methods offer little guidance and brute-force search becomes intractable. We introduce FlowBoost, a closed-loop generative framework that learns to discover rare and extremal geometric structures by combining three components: (i) a geometry-aware conditional flow-matching model that learns to sample high-quality configurations, (ii) reward-guided policy optimization with action exploration that directly optimizes the generation process toward the objective while maintaining diversity, and (iii) stochastic local search for both training-data generation and final refinement. Unlike prior open-loop approaches, such as PatternBoost that retrains on filtered discrete samples, or AlphaEvolve which relies on frozen Large Language Models (LLMs) as evolutionary mutation operators, FlowBoost enforces geometric feasibility during sampling, and propagates reward signal directly into the generative model, closing the optimization loop and requiring much smaller training sets and shorter training times, and reducing the required outer-loop iterations by orders of magnitude, while eliminating dependence on LLMs. We demonstrate the framework on four geometric optimization problems: sphere packing in hypercubes, circle packing maximizing sum of radii, the Heilbronn triangle problem, and star discrepancy minimization. In several cases, FlowBoost discovers configurations that match or exceed the best known results. For circle packings, we improve the best known lower bounds, surpassing the LLM-based system AlphaEvolve while using substantially fewer computational resources.","url_abs":"https://arxiv.org/abs/2601.18005","url_pdf":"https://arxiv.org/pdf/2601.18005","source":{"archive":null,"snapshot":"2025-07-28","note":"not in the Papers with Code archive (frozen at the snapshot)","row_kind":"graph","title_abstract_authors_date":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)"},"code_links":[],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2601.18005","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2601.18005"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"mentioned_in_github":null,"is_official":null,"provenance":"deterministic:regex_extraction","mentioned_in_paper":null,"url":"https://github.com/berczig/FlowBoost","reach":null}],"summary":{"unverified":18},"by_repo_kind":{"found_in_text":{"samples":18,"ran":0,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":18,"samples":[{"code_sha256_prefix":"391da1089de662c5","entry":"all_triangle_areas","repo":"berczig/FlowBoost","repo_kind":"found_in_text","path":"flow_boost/heilbronn_square/flow_matching_heilbronn.py","file_url":"https://github.com/berczig/FlowBoost/blob/HEAD/flow_boost/heilbronn_square/flow_matching_heilbronn.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"391da1089de662c5"}},{"code_sha256_prefix":"58a710501d5d8ba3","entry":"circle_pair_penalty","repo":"berczig/FlowBoost","repo_kind":"found_in_text","path":"flow_boost/circles_in_square/flow_matching_sumradii.py","file_url":"https://github.com/berczig/FlowBoost/blob/HEAD/flow_boost/circles_in_square/flow_matching_sumradii.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"58a710501d5d8ba3"}},{"code_sha256_prefix":"2bff1ffa367143da","entry":"circle_wall_penalty","repo":"berczig/FlowBoost","repo_kind":"found_in_text","path":"flow_boost/circles_in_square/flow_matching_sumradii.py","file_url":"https://github.com/berczig/FlowBoost/blob/HEAD/flow_boost/circles_in_square/flow_matching_sumradii.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"2bff1ffa367143da"}},{"code_sha256_prefix":"8e2adb27913602ba","entry":"clamp_unit_square","repo":"berczig/FlowBoost","repo_kind":"found_in_text","path":"flow_boost/heilbronn_square/flow_matching_heilbronn.py","file_url":"https://github.com/berczig/FlowBoost/blob/HEAD/flow_boost/heilbronn_square/flow_matching_heilbronn.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"8e2adb27913602ba"}},{"code_sha256_prefix":"2f705b01f997764b","entry":"compute_metrics","repo":"berczig/FlowBoost","repo_kind":"found_in_text","path":"flow_boost/heilbronn_square/plot_data_heilbronn.py","file_url":"https://github.com/berczig/FlowBoost/blob/HEAD/flow_boost/heilbronn_square/plot_data_heilbronn.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"2f705b01f997764b"}},{"code_sha256_prefix":"15fc8a32d3849637","entry":"compute_metrics","repo":"berczig/FlowBoost","repo_kind":"found_in_text","path":"flow_boost/heilbronn_square/plot_data_points.py","file_url":"https://github.com/berczig/FlowBoost/blob/HEAD/flow_boost/heilbronn_square/plot_data_points.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"15fc8a32d3849637"}},{"code_sha256_prefix":"5bfedbd698572809","entry":"compute_sum_radii","repo":"berczig/FlowBoost","repo_kind":"found_in_text","path":"flow_boost/circles_in_square/plot_data_sumradii.py","file_url":"https://github.com/berczig/FlowBoost/blob/HEAD/flow_boost/circles_in_square/plot_data_sumradii.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"5bfedbd698572809"}},{"code_sha256_prefix":"ca467f6c359703a2","entry":"extract_bstart_bend","repo":"berczig/FlowBoost","repo_kind":"found_in_text","path":"flow_boost/output/vmerge.py","file_url":"https://github.com/berczig/FlowBoost/blob/HEAD/flow_boost/output/vmerge.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"ca467f6c359703a2"}},{"code_sha256_prefix":"0a09310d7745fcc3","entry":"find_min_area_and_triangles","repo":"berczig/FlowBoost","repo_kind":"found_in_text","path":"flow_boost/heilbronn_square/sample_generation.py","file_url":"https://github.com/berczig/FlowBoost/blob/HEAD/flow_boost/heilbronn_square/sample_generation.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"0a09310d7745fcc3"}},{"code_sha256_prefix":"5e7dba141ade472e","entry":"init_radii","repo":"berczig/FlowBoost","repo_kind":"found_in_text","path":"flow_boost/circles_in_square/sample_generation.py","file_url":"https://github.com/berczig/FlowBoost/blob/HEAD/flow_boost/circles_in_square/sample_generation.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"5e7dba141ade472e"}},{"code_sha256_prefix":"96942d0b9f12bda2","entry":"load_dataset","repo":"berczig/FlowBoost","repo_kind":"found_in_text","path":"flow_boost/circles_in_square/plot_data_sumradii.py","file_url":"https://github.com/berczig/FlowBoost/blob/HEAD/flow_boost/circles_in_square/plot_data_sumradii.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"96942d0b9f12bda2"}},{"code_sha256_prefix":"c0a6c1d41247a2b3","entry":"load_dataset","repo":"berczig/FlowBoost","repo_kind":"found_in_text","path":"flow_boost/heilbronn_square/plot_data_heilbronn.py","file_url":"https://github.com/berczig/FlowBoost/blob/HEAD/flow_boost/heilbronn_square/plot_data_heilbronn.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"c0a6c1d41247a2b3"}},{"code_sha256_prefix":"cfbfce6523b26b63","entry":"load_dataset","repo":"berczig/FlowBoost","repo_kind":"found_in_text","path":"flow_boost/heilbronn_square/plot_data_points.py","file_url":"https://github.com/berczig/FlowBoost/blob/HEAD/flow_boost/heilbronn_square/plot_data_points.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"cfbfce6523b26b63"}},{"code_sha256_prefix":"0d73bc22ded8a72e","entry":"sample_t","repo":"berczig/FlowBoost","repo_kind":"found_in_text","path":"flow_boost/circles_in_square/flow_matching_sumradii.py","file_url":"https://github.com/berczig/FlowBoost/blob/HEAD/flow_boost/circles_in_square/flow_matching_sumradii.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"0d73bc22ded8a72e"}},{"code_sha256_prefix":"cafea8dbef362e75","entry":"sample_uniform_centers","repo":"berczig/FlowBoost","repo_kind":"found_in_text","path":"flow_boost/circles_in_square/sample_generation.py","file_url":"https://github.com/berczig/FlowBoost/blob/HEAD/flow_boost/circles_in_square/sample_generation.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"cafea8dbef362e75"}},{"code_sha256_prefix":"91481f0658ae85d9","entry":"sample_uniform_points","repo":"berczig/FlowBoost","repo_kind":"found_in_text","path":"flow_boost/heilbronn_square/sample_generation.py","file_url":"https://github.com/berczig/FlowBoost/blob/HEAD/flow_boost/heilbronn_square/sample_generation.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"91481f0658ae85d9"}},{"code_sha256_prefix":"903d7d4685ff1fda","entry":"sample_uniform_x1_like","repo":"berczig/FlowBoost","repo_kind":"found_in_text","path":"flow_boost/heilbronn_square/flow_matching_heilbronn.py","file_url":"https://github.com/berczig/FlowBoost/blob/HEAD/flow_boost/heilbronn_square/flow_matching_heilbronn.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"903d7d4685ff1fda"}},{"code_sha256_prefix":"bc905e1b5e297d92","entry":"validate_circle_samples","repo":"berczig/FlowBoost","repo_kind":"found_in_text","path":"flow_boost/circles_in_square/flow_matching_sumradii_centers_only.py","file_url":"https://github.com/berczig/FlowBoost/blob/HEAD/flow_boost/circles_in_square/flow_matching_sumradii_centers_only.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"bc905e1b5e297d92"}}]},"arxiv_metadata":{"licence":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)","fields":["title","abstract","authors","date"],"primary_category":"cs.LG","source":"arxiv_2026.jsonl"},"syntology_extracted_results":{"kind":"leaderboard_placements","source":"Syntology's leaderboard-shaped extractor over the paper's own arXiv-HTML tables: a model pointed at a cell, the number was read from that cell and checked against the board's metric, dataset, split and scale, and an independent check accepted the entry; not reviewed by the paper's authors or the archive's editors","extractor_model":"global.anthropic.claude-sonnet-4-5-20250929-v1:0","verifier_model":null,"prompt_sha":"fa63d4bb9d755694","coverage":{"sentence":"Syntology has checked 6,264 of the 9,581 papers on this site that are newer than the archive; results from the others appear after they are checked.","papers_newer_than_archive":9581,"papers_checked":6264},"entries":[],"not_placed":{"boards":0,"rejected_by_independent_check":0,"refused_by_a_rule":0,"check_did_not_answer":0,"proposed_without_a_cell":0,"declined_by_site":0}}}