{"id":15833674,"url":"https://github.com/hackintoshrao/solving-physics-in-keras","last_synced_at":"2025-04-01T12:26:02.209Z","repository":{"id":149607140,"uuid":"122592663","full_name":"hackintoshrao/solving-physics-in-keras","owner":"hackintoshrao","description":"Solving a physics problem using keras ","archived":false,"fork":false,"pushed_at":"2018-02-23T08:49:22.000Z","size":9,"stargazers_count":2,"open_issues_count":0,"forks_count":0,"subscribers_count":3,"default_branch":"master","last_synced_at":"2024-10-06T13:41:33.337Z","etag":null,"topics":["deep-neural-networks","keras","keras-neural-networks","neural-networks","physics"],"latest_commit_sha":null,"homepage":null,"language":"Jupyter Notebook","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/hackintoshrao.png","metadata":{"files":{"readme":"Readme.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2018-02-23T08:18:53.000Z","updated_at":"2021-04-12T08:41:18.000Z","dependencies_parsed_at":null,"dependency_job_id":"a3ce74e5-6f45-4100-b03a-74fde5277049","html_url":"https://github.com/hackintoshrao/solving-physics-in-keras","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/hackintoshrao%2Fsolving-physics-in-keras","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/hackintoshrao%2Fsolving-physics-in-keras/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/hackintoshrao%2Fsolving-physics-in-keras/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/hackintoshrao%2Fsolving-physics-in-keras/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/hackintoshrao","download_url":"https://codeload.github.com/hackintoshrao/solving-physics-in-keras/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":246637699,"owners_count":20809661,"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":["deep-neural-networks","keras","keras-neural-networks","neural-networks","physics"],"created_at":"2024-10-05T13:41:24.060Z","updated_at":"2025-04-01T12:26:02.189Z","avatar_url":"https://github.com/hackintoshrao.png","language":"Jupyter Notebook","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Solving physics problem using Keras\n\n\nArtificial neural network regression on (Nsample, None, 6) -\u003e (Nsample, 3) datasets\n # Artificial neural network regression on (Nsample, None, 6) -\u003e (Nsample, 3) datasets , you're asked to use machine learning to fit a dataset of shape (Nsample, None, 6) -\u003e (Nsample, 3). It's a toy dataset, not a Netflix user group. Code for generating the dataset is part of the code . Here's a description of the 'bigger picture' of the dataset. \n \n The Coulomb's law says that force between 2 electrical charges is inversely proportional to the square of the distance between them. Imagine you're Coulomb, and you are asked to rediscover that law using Machine Learning. You have the three-dimensional distance between 2 electrical charges, a vector with shape (3), and the force, a vector of shape (3). You collect the force data on many different three-dimensional distances. You get a dataset of shape (Nsample, 3) -\u003e (Nsample, 3). You use straightforward multi-layer perceptron to fit this dataset and are successful. Imagine you're Coulomb and asked to use ML to regress the force-distance law again. But this time, you are bad at experiments, so whenever you try to make 2 electrically charged bodies, you end up making 1000-2000 of them. You can still measure the force on an electrical charge, but this force results not from a pair of electrical charges, but 1000-2000 pairs SUMMED together! What's worse, 1000-2000 is a VARIABLE range, maybe 1001 for one run and 1500 for another! So, for each sample, you have 1000-2000 three-dimensional distances (None,3), and a total force (3). You collect the force data on many such systems and end up with a dataset of shape (Nsample, None, 3) -\u003e (Nsample, 3). What should you do now? That is this project's question. It is definitely doable - Coulomb did it using mathematics, so we sure can do it using ML regression. A minor point: - Imagine the three-dimensional distances is 6 dimensional. Thus a (Nsample, None, 6) -\u003e (Nsample, None, 3) dataset. I want you to study this because it's harder. Coulomb did it using math, so we should be able to do it using ML. ---------------------------------------------------------------------------- My only criterion of success is accuracy. I'd say 5-6% relative RMSE error? You can either use my code for generating dataset. Alternatively, if you know what Coulomb's law is and understand the foregoing idea, you can roll your own code; but you should use the system size specified in my code. Too small a system and the question is too easy. \n\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fhackintoshrao%2Fsolving-physics-in-keras","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fhackintoshrao%2Fsolving-physics-in-keras","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fhackintoshrao%2Fsolving-physics-in-keras/lists"}