{"id":49932698,"url":"https://github.com/ayan-cs/caufr-ts","last_synced_at":"2026-05-17T04:05:10.955Z","repository":{"id":352945347,"uuid":"1148235687","full_name":"ayan-cs/caufr-ts","owner":"ayan-cs","description":"Implementation of the paper \"CauFR-TS: Causal Time Series Identifiability via Factorized Representations.\" [Accepted with Minor Revisions at TMLR]","archived":false,"fork":false,"pushed_at":"2026-05-10T06:10:45.000Z","size":22392,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2026-05-10T08:18:36.054Z","etag":null,"topics":["causal-discovery","causal-inference","causal-reasoning","granger-causality","granger-causality-analysis","independent-component-analysis","time-series-analysis","time-series-forecasting"],"latest_commit_sha":null,"homepage":"https://openreview.net/forum?id=Al4OnLoQsp","language":"Jupyter Notebook","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/ayan-cs.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,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-02-02T18:26:11.000Z","updated_at":"2026-05-10T06:10:49.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/ayan-cs/caufr-ts","commit_stats":null,"previous_names":["ayan-cs/caufr-ts"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/ayan-cs/caufr-ts","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ayan-cs%2Fcaufr-ts","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ayan-cs%2Fcaufr-ts/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ayan-cs%2Fcaufr-ts/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ayan-cs%2Fcaufr-ts/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/ayan-cs","download_url":"https://codeload.github.com/ayan-cs/caufr-ts/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ayan-cs%2Fcaufr-ts/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":33127007,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-16T18:38:32.183Z","status":"online","status_checked_at":"2026-05-17T02:00:05.366Z","response_time":107,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"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":["causal-discovery","causal-inference","causal-reasoning","granger-causality","granger-causality-analysis","independent-component-analysis","time-series-analysis","time-series-forecasting"],"created_at":"2026-05-17T04:05:07.722Z","updated_at":"2026-05-17T04:05:10.947Z","avatar_url":"https://github.com/ayan-cs.png","language":"Jupyter Notebook","funding_links":[],"categories":[],"sub_categories":[],"readme":"# CauFR-TS\n\n\u003cp align=\"center\"\u003e\n  \u003cb\u003eCausal Time Series Identifiability via Factorized Representations\u003c/b\u003e\n  \u003cbr\u003e \u003ci\u003eAccepted \u0026 Published at TMLR\u003c/i\u003e \n\u003c/p\u003e\n\u003cp align=\"center\"\u003e\n  \u003ca href=\"https://openreview.net/forum?id=Al4OnLoQsp\"\u003e\n    \u003cimg src=\"https://img.shields.io/badge/TMLR-Paper-blue?logo=openreview\"\u003e\n  \u003c/a\u003e\n  \u003ca href=\"https://github.com/ayan-cs/CauFR-TS\"\u003e\n    \u003cimg src=\"https://visitor-badge.laobi.icu/badge?page_id=ayan-cs.CauFR-TS\"\u003e\n  \u003c/a\u003e\n\u003c/p\u003e\n\u003cp align=\"center\"\u003e\n  \u003cimg src=\"https://img.shields.io/badge/PyTorch-2.4.0-red\"\u003e\n  \u003cimg src=\"https://img.shields.io/badge/scikit--learn-1.8.0-F7931E\"\u003e\n  \u003cimg src=\"https://img.shields.io/badge/CUDA-12.5-green\"\u003e\n  \u003cimg src=\"https://img.shields.io/badge/License-MIT-blue\"\u003e\n  \u003cimg src=\"https://img.shields.io/badge/Reproducibility-Verified-success\"\u003e\n\u003c/p\u003e\n\n## 🥸 Overview\n\n**CauFR-TS** is a neural framework for causal discovery in multivariate time series that explicitly enforces mechanism modularity at the representation level.\n\nUnlike existing neural Granger causality methods that rely on shared latent encoders, CauFR-TS employs **dimension-wise factorized encoders**, preventing latent information leakage across variables. This design restores the conditional independence assumptions required for Granger causality and leads to **structurally identifiable causal graphs**.\n\nIn addition, CauFR-TS introduces a **data-driven, parameter-free adaptive thresholding** strategy based on Gaussian Mixture Models to robustly separate causal signals from noise.\n\n---\n## 🤓 Key Contributions\n- **Factorized Encoder Architecture**  \n  Each time series variable is encoded independently, eliminating latent confounding caused by shared representations.\n\n- **Group-Sparse Decoder for Granger Causality**  \n  Cross-variable dependencies are mediated exclusively through structured latent aggregation using group lasso regularization.\n\n- **Adaptive, Data-Driven Thresholding**  \n  A Gaussian Mixture Model is fit to decoder weight norms to automatically distinguish causal interactions from noise, avoiding heuristic cutoffs.\n\n- **Strong Empirical Performance**  \n  Consistent improvements over state-of-the-art baselines on synthetic chaotic systems and in silico biological benchmarks.\n\n---\n## 🧐 Methodology\n\n\u003cp align=\"center\"\u003e \u003cimg src=\"assets/architecture-v3.png\" width=\"85%\"\u003e \u003c/p\u003e\n\nCauFR-TS models the conditional distribution $p(x_t|x_{1:t-1})$ using a factorized variational architecture:\n**1.** Each variable is processed by an independent Transformer-based encoder.\n**2.** Latent variables are reparameterized and concatenated into a structured latent vector.\n**3.** Multi-head decoders predict future values using group-sparse weights.\n**4.** Causal adjacency is recovered via adaptive probabilistic thresholding on decoder weights.\n\nThis architecture ensures that all inter-variable information flow is explicitly mediated by the learned causal matrix.\n\n---\n## 😵‍💫 Results \u0026 Visualizations\n\n* 📈 **Training Dynamics:**\n    - Evolution of group-lasso weights\n    - Separation of causal vs non-causal mechanisms\n    - Convergence of adaptive thresholds\n\n\u003cp align=\"center\"\u003e \u003cimg src=\"assets/convergence_plots.png\" width=\"85%\"\u003e \u003c/p\u003e\n\n* 📊 **Causal Graph Recovery:**\n    - (Left) Ground-truth causal adjacency matrix\n    - (Middle) Estimated raw causal matrix obtained from the learned group-lasso decoder weights prior to thresholding\n    - (Right) The final binary causal graph after applying the proposed adaptive thresholding procedure\n\n\u003cp align=\"center\"\u003e \u003cimg src=\"assets/causal_matrix_quality.png\" width=\"85%\"\u003e \u003c/p\u003e\n\n---\n## 🤖 Execution\n\n```\ngit clone git@github.com:ayan-cs/CauFR-TS.git\ncd CauFR-TS\n```\n\nMake necessary changes based on your dataset. For a quick run on  H´enon maps,\n\n```\npython train.py\n```\n\n---\n\n## 🗣️ Contact\n\n\u003cp align=\"left\"\u003e\n  📧 Email: p23iot002@iitj.ac.in \u003cbr\u003e\n  📧 Email: ghoshayanabha@gmail.com \u003cbr\u003e\n  🎓 \u003ca href=\"https://scholar.google.com/citations?user=oB4N3H4AAAAJ\u0026hl=en\"\u003eGoogle Scholar\u003c/a\u003e \u003cbr\u003e\n  🌐 \u003ca href=\"https://www.linkedin.com/in/ayanabha-ghosh-cs/\"\u003eLinkedIn\u003c/a\u003e\n\u003c/p\u003e\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fayan-cs%2Fcaufr-ts","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fayan-cs%2Fcaufr-ts","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fayan-cs%2Fcaufr-ts/lists"}