{"id":16765091,"url":"https://github.com/dataplayer12/fly-lsh","last_synced_at":"2025-03-21T23:33:16.162Z","repository":{"id":111555086,"uuid":"112754310","full_name":"dataplayer12/Fly-LSH","owner":"dataplayer12","description":"An implementation of efficient LSH inspired by fruit fly brain","archived":false,"fork":false,"pushed_at":"2018-12-23T09:26:27.000Z","size":219,"stargazers_count":87,"open_issues_count":0,"forks_count":27,"subscribers_count":6,"default_branch":"master","last_synced_at":"2024-10-14T05:28:22.168Z","etag":null,"topics":["locality-sensitive-hashing","machine-learning-algorithms","nearest-neighbor-search","retrieval"],"latest_commit_sha":null,"homepage":null,"language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/dataplayer12.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null}},"created_at":"2017-12-01T15:19:16.000Z","updated_at":"2024-07-28T14:37:48.000Z","dependencies_parsed_at":"2023-05-29T09:15:23.159Z","dependency_job_id":null,"html_url":"https://github.com/dataplayer12/Fly-LSH","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/dataplayer12%2FFly-LSH","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/dataplayer12%2FFly-LSH/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/dataplayer12%2FFly-LSH/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/dataplayer12%2FFly-LSH/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/dataplayer12","download_url":"https://codeload.github.com/dataplayer12/Fly-LSH/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":221820599,"owners_count":16886222,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["locality-sensitive-hashing","machine-learning-algorithms","nearest-neighbor-search","retrieval"],"created_at":"2024-10-13T05:28:17.491Z","updated_at":"2024-10-28T11:14:54.484Z","avatar_url":"https://github.com/dataplayer12.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"[![Build status](https://travis-ci.org/dataplayer12/Fly-LSH.svg?master)](https://travis-ci.org/dataplayer12)\n# Paper\nCode accompanying our [paper](https://arxiv.org/abs/1812.01844) **Improving Similarity Search with High-dimensional Locality sensitive hashing**\n\n# Summary\nWe make three important contributions:\n1. We present a new data independent approximate nearest neighbor (ANN) search algorithm inspired by the fruit fly olfactory circuit introduced by [Dasgupta et. al.](http://science.sciencemag.org/content/358/6364/793/tab-article-info). Named *DenseFly*, the proposed algorithm performs significantly better than several existing data independent algorithms on six benchmark datasets. (figures 2 and 3)\n2. We prove several theoretical results about the original *FlyHash* as well as the proposed *DenseFly* algorithms. In particular, we show that *FlyHash* preserves rank similarity under any *Lp* norm and that *DenseFly* approximates a *SimHash* in very high dimensions at a much lower computational cost. (Lemmas 1 and 2)\n3. We develop a multi-probe binning scheme for *FlyHash* and *DenseFly* algorithms, which are indispensable for practical applications of ANN algorithms. Remarkably, the proposed multi-probe binning scheme does not require additional computation over and above those used to create the high dimensional *Fly* or *DenseFly* hashes. Thus, the multi-probe versions of *FlyHash* and *DenseFly* result in a significant increase in mAP scores for a given query time. (figure 4)\n\n# Code\nThe code for all the new algorithms described are present in one large file. Helper scripts to compare different algorithms will be added soon.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fdataplayer12%2Ffly-lsh","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fdataplayer12%2Ffly-lsh","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fdataplayer12%2Ffly-lsh/lists"}