{"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/a-fast-and-scalable-joint-estimator-for-1","title":"A Fast and Scalable Joint Estimator for Integrating Additional Knowledge in Learning Multiple Related Sparse Gaussian Graphical Models","arxiv_id":"1806.00548","date":"2018-06-01","proceeding":"ICML 2018 7","authors":["Beilun Wang","Arshdeep Sekhon","Yanjun Qi"],"abstract":"We consider the problem of including additional knowledge in estimating\nsparse Gaussian graphical models (sGGMs) from aggregated samples, arising often\nin bioinformatics and neuroimaging applications. Previous joint sGGM estimators\neither fail to use existing knowledge or cannot scale-up to many tasks (large\n$K$) under a high-dimensional (large $p$) situation. In this paper, we propose\na novel \\underline{J}oint \\underline{E}lementary \\underline{E}stimator\nincorporating additional \\underline{K}nowledge (JEEK) to infer multiple related\nsparse Gaussian Graphical models from large-scale heterogeneous data. Using\ndomain knowledge as weights, we design a novel hybrid norm as the minimization\nobjective to enforce the superposition of two weighted sparsity constraints,\none on the shared interactions and the other on the task-specific structural\npatterns. This enables JEEK to elegantly consider various forms of existing\nknowledge based on the domain at hand and avoid the need to design\nknowledge-specific optimization. JEEK is solved through a fast and entry-wise\nparallelizable solution that largely improves the computational efficiency of\nthe state-of-the-art $O(p^5K^4)$ to $O(p^2K^4)$. We conduct a rigorous\nstatistical analysis showing that JEEK achieves the same convergence rate\n$O(\\log(Kp)/n_{tot})$ as the state-of-the-art estimators that are much harder\nto compute. Empirically, on multiple synthetic datasets and two real-world\ndata, JEEK outperforms the speed of the state-of-arts significantly while\nachieving the same level of prediction accuracy. Available as R tool @\nhttp://jointnets.org/","url_abs":"http://arxiv.org/abs/1806.00548v4","url_pdf":"http://arxiv.org/pdf/1806.00548v4.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"a-fast-and-scalable-joint-estimator-for-1","repo_url":"https://github.com/QData/JEEK","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null},{"paper_slug":"a-fast-and-scalable-joint-estimator-for-1","repo_url":"https://github.com/QData/JointNets","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"2k","task_name":"2k"},{"task_slug":"computational-efficiency","task_name":"Computational Efficiency"},{"task_slug":"structured-prediction","task_name":"Structured Prediction"}],"methods":[{"method_slug":"speed","method_name":"SPEED"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}