{"paper":{"title":"Transformed Latent Variable Multi-Output Gaussian Processes","license":"http://creativecommons.org/licenses/by/4.0/","headline":"T-LVMOGP scales multi-output Gaussian processes to over 10,000 outputs by embedding inputs and per-output latent variables through a Lipschitz-regularised neural network.","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Magnus Rattray, Mauricio A \\'Alvarez, Sokratia Georgaka, Xiaoyu Jiang, Xinxing Shi","submitted_at":"2026-05-06T17:05:50Z","abstract_excerpt":"Multi-Output Gaussian Processes (MOGPs) provide a principled probabilistic framework for modelling correlated outputs but face scalability bottlenecks when applied to datasets with high-dimensional output spaces. To maintain tractability, existing methods typically resort to restrictive assumptions, such as employing low-rank or sum-of-separable kernels, which can limit expressiveness. We propose the Transformed Latent Variable MOGP (T-LVMOGP), a novel framework that scales MOGPs to a massive number of outputs while preserving the capacity to capture meaningful inter-output dependencies. T-LVM"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"T-LVMOGP constructs a flexible multi-output deep kernel by mapping inputs and output-specific latent variables into an embedding space using a Lipschitz-regularised neural network, scales MOGPs to a massive number of outputs while preserving inter-output dependencies, and outperforms baselines in predictive accuracy and computational efficiency on climate modelling with over 10,000 outputs and zero-inflated spatial transcriptomics data.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The Lipschitz-regularised neural network mapping of inputs and output-specific latent variables into an embedding space is sufficient to capture meaningful inter-output dependencies without excessive loss of expressiveness or introduction of new fitting artefacts when combined with stochastic variational inference.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"T-LVMOGP scales multi-output Gaussian processes to massive output dimensions using transformed latent variables, deep kernels, and stochastic variational inference while capturing inter-output dependencies.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"T-LVMOGP scales multi-output Gaussian processes to over 10,000 outputs by embedding inputs and per-output latent variables through a Lipschitz-regularised neural network.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"846807fe91fd874c497240f8a08551258bb3a6257dbb78e529adbe58a3a863b2"},"source":{"id":"2605.05133","kind":"arxiv","version":2},"verdict":{"id":"f375785c-5144-4bd4-ae03-27243cd9ea54","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-08T16:54:49.717103Z","strongest_claim":"T-LVMOGP constructs a flexible multi-output deep kernel by mapping inputs and output-specific latent variables into an embedding space using a Lipschitz-regularised neural network, scales MOGPs to a massive number of outputs while preserving inter-output dependencies, and outperforms baselines in predictive accuracy and computational efficiency on climate modelling with over 10,000 outputs and zero-inflated spatial transcriptomics data.","one_line_summary":"T-LVMOGP scales multi-output Gaussian processes to massive output dimensions using transformed latent variables, deep kernels, and stochastic variational inference while capturing inter-output dependencies.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The Lipschitz-regularised neural network mapping of inputs and output-specific latent variables into an embedding space is sufficient to capture meaningful inter-output dependencies without excessive loss of expressiveness or introduction of new fitting artefacts when combined with stochastic variational inference.","pith_extraction_headline":"T-LVMOGP scales multi-output Gaussian processes to over 10,000 outputs by embedding inputs and per-output latent variables through a Lipschitz-regularised neural network."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.05133/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-20T10:36:52.567061Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_title_agreement","ran_at":"2026-05-19T21:31:19.502767Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T13:49:12.863193Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"656f63ac54edb5b67eab7d02597a3e0742d8eda224d94314aa73fb2251d8e85e"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}