{"paper":{"title":"Iskra: A System for Inverse Geometry Processing","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"Iskra differentiates existing geometry processing algorithms by applying the adjoint method directly to user-written imperative code.","cross_cats":["cs.CV","cs.LG"],"primary_cat":"cs.GR","authors_text":"Ahmed H. Mahmoud, Ana Dodik, Justin Solomon","submitted_at":"2026-02-12T15:59:06Z","abstract_excerpt":"We propose a system for differentiating through solutions to geometry processing problems. Our system differentiates a broad class of geometric algorithms, exploiting existing fast problem-specific schemes common to geometry processing, including local-global and ADMM solvers. It is compatible with machine learning frameworks, opening doors to new classes of inverse geometry processing applications. We marry the scatter-gather approach to mesh processing with tensor-based workflows and rely on the adjoint method applied to user-specified imperative code to generate an efficient backward pass b"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Our system differentiates a broad class of geometric algorithms, exploiting existing fast problem-specific schemes common to geometry processing, including local-global and ADMM solvers.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the adjoint method applied to user-specified imperative code will generate an efficient and accurate backward pass for the targeted geometry processing algorithms without requiring reformulation or introducing significant overhead.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Iskra provides a differentiable interface to common geometry processing algorithms by marrying scatter-gather mesh operations with tensor workflows and adjoint differentiation, enabling inverse applications.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Iskra differentiates existing geometry processing algorithms by applying the adjoint method directly to user-written imperative code.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"a4db964500774fa1258081b80f2cfc91d66bce5881bd4b7c59e4c80a99f5cb18"},"source":{"id":"2602.12105","kind":"arxiv","version":2},"verdict":{"id":"df34bcae-7da4-4510-8aa4-d063e9e5792c","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T05:35:43.512699Z","strongest_claim":"Our system differentiates a broad class of geometric algorithms, exploiting existing fast problem-specific schemes common to geometry processing, including local-global and ADMM solvers.","one_line_summary":"Iskra provides a differentiable interface to common geometry processing algorithms by marrying scatter-gather mesh operations with tensor workflows and adjoint differentiation, enabling inverse applications.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the adjoint method applied to user-specified imperative code will generate an efficient and accurate backward pass for the targeted geometry processing algorithms without requiring reformulation or introducing significant overhead.","pith_extraction_headline":"Iskra differentiates existing geometry processing algorithms by applying the adjoint method directly to user-written imperative code."},"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"}