{"id":13738424,"url":"https://github.com/cwkx/GON","last_synced_at":"2025-05-08T16:34:00.315Z","repository":{"id":62082147,"uuid":"276976873","full_name":"cwkx/GON","owner":"cwkx","description":"Gradient Origin Networks - a new type of generative model that is able to quickly learn a latent representation without an encoder","archived":false,"fork":false,"pushed_at":"2021-02-04T11:49:43.000Z","size":2843,"stargazers_count":161,"open_issues_count":2,"forks_count":20,"subscribers_count":12,"default_branch":"master","last_synced_at":"2024-11-15T07:34:29.133Z","etag":null,"topics":["autoencoders","generative-models","implicit","machine-learning","networks","neural-networks","representation","siren","unsupervised-learning"],"latest_commit_sha":null,"homepage":"https://cwkx.github.io/data/GON/","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/cwkx.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}},"created_at":"2020-07-03T19:51:46.000Z","updated_at":"2024-07-23T01:51:02.000Z","dependencies_parsed_at":"2022-10-26T04:49:18.849Z","dependency_job_id":null,"html_url":"https://github.com/cwkx/GON","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/cwkx%2FGON","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/cwkx%2FGON/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/cwkx%2FGON/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/cwkx%2FGON/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/cwkx","download_url":"https://codeload.github.com/cwkx/GON/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":253105570,"owners_count":21855057,"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":["autoencoders","generative-models","implicit","machine-learning","networks","neural-networks","representation","siren","unsupervised-learning"],"created_at":"2024-08-03T03:02:22.076Z","updated_at":"2025-05-08T16:33:59.844Z","avatar_url":"https://github.com/cwkx.png","language":"Python","readme":"# Gradient Origin Networks\n\nThis paper has been accepted at ICLR 2021.\n\nThis paper proposes a new type of generative model that is able to quickly learn a latent representation without an encoder. This is achieved using empirical Bayes to calculate the expectation of the posterior, which is implemented by initialising a latent vector with zeros, then using the gradient of the log-likelihood of the data with respect to this zero vector as new latent points. The approach has similar characteristics to autoencoders, but with a simpler architecture, and is demonstrated in a variational autoencoder equivalent that permits sampling. This also allows implicit representation networks to learn a space of implicit functions without requiring a hypernetwork, retaining their representation advantages across datasets. The experiments show that the proposed method converges faster, with significantly lower reconstruction error than autoencoders, while requiring half the parameters.\n\n[![GON](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/gist/samb-t/fbac83a2ec9312616ed61cd74dac50ce/gon.ipynb) **(GON)** \u003cbr\u003e\n\n[![Variational GON](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/gist/samb-t/5181643d0a5aeef7830b50dc4e84f659/variational-gon.ipynb) **(Variational GON)** \u003cbr\u003e\n\n[![Implicit GON](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/gist/cwkx/8c3a8b514f3bdfe123edc3bb0e6b7eca/gon.ipynb) **(Implicit GON)**\u003cbr\u003e\n\nThe code is available in [GON.py](GON.py) and licensed under the MIT license. For more information, please visit the [Project Page](https://cwkx.github.io/data/GON/). Here is a [link to the paper](https://arxiv.org/pdf/2007.02798.pdf). The implicit GON version uses a SIREN (*Implicit Neural Representations with Periodic Activation Functions*, Sitzmann et al., 2020).\n\n[![YouTube Preview](VIDEO.gif)](https://www.youtube.com/watch?v=ro7t98Q1gXg)\n\n\n## Citation\nIf you find this useful, please cite:\n```\n@inproceedings{bond2020gradient,\n   title     = {Gradient Origin Networks},\n   author    = {Sam Bond-Taylor and Chris G. Willcocks},\n   booktitle = {International Conference on Learning Representations},\n   year      = {2021},\n   url       = {https://openreview.net/pdf?id=0O_cQfw6uEh}\n}\n```\n","funding_links":[],"categories":["Python"],"sub_categories":[],"project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fcwkx%2FGON","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fcwkx%2FGON","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fcwkx%2FGON/lists"}