{"agent_url":"https://mathhub.eggtea.org/works/complex-corrected-stochastic-gradient-descent-3ac3e220/v/1/files/Agent.md","archived":false,"change_note":"Initial publication authorized by eggtea; CC BY 4.0; original manuscript, available experiment source and explicit review scope.","created_at":"2026-10-17T09:22:24+00:00","files":[{"bytes":220914,"path":"1_paper.pdf","sha256":"3e222f197266ca0283b9618f4eb4d4cc1818c0e90a916d724c3981404da6eb02"},{"bytes":14059,"path":"2.1_AI_paper.md","sha256":"08317dd90d8813c1b4838b52c7693e547e76c47868e90597c75d01291976f47f"},{"bytes":58909,"path":"2_AI_paper.pdf","sha256":"347bdb281df4dac1c05f4fcf85e928993ddc3076ffab844a503550b7c60ef0a6"},{"bytes":34250,"path":"3_guide_en.pdf","sha256":"d14555cef88109104bbf6ef8f2bccf49add7cd4841f54dee1341cfcce854843d"},{"bytes":199717,"path":"4_guide_zh.pdf","sha256":"e6f3fd1ca91b465906e355512ea99206067c27361d471a8eb245b38a791ceec2"},{"bytes":111493,"path":"5_review_en.pdf","sha256":"08554e18d5980609042a9e366849530f68f3bfbc0d599c66ab1b9db9396e00dc"},{"bytes":192585,"path":"6_review_zh.pdf","sha256":"f45726a369fb1704a8ebec233655cb077fe556956950d5fea2ef1b5f1a4fa9f4"},{"bytes":34719,"path":"7_theorem_graph.pdf","sha256":"b58df8564684388bff7fcc9ebcf2df2844ba805a099cd870abb0b8440b16c0cc"},{"bytes":2245,"path":"Agent.md","sha256":"768e7e005404cef38d4def0b69aa2418d8ffd69b16029a4073fad89a7591b313"},{"bytes":326,"path":"LICENSE.txt","sha256":"c18762d18fac8ab4f109b0da15919fe68116124ab16e12a68a964154afcdb2a9"},{"bytes":1089,"path":"README.md","sha256":"c0160e6a7436153bc6231f09e47a82ad57e5b70d35fd9a67d85a90bb973afa8b"},{"bytes":1541,"path":"experiments/README.md","sha256":"df17c3f8b0eba99a44b7e51d84c0a4f1a89552d9bbf362f35458484d456b786e"},{"bytes":2082,"path":"experiments/neural_network/run_long.py","sha256":"fe67bcc160c184d17913ba939e547acb96dc28be8a49c2561b87d8616b58353e"},{"bytes":11643,"path":"experiments/neural_network/run_nn.py","sha256":"aea4a0a0e163d788a4e59d12acd14944edaf1e48770872ff36232d5e24b21271"},{"bytes":23,"path":"experiments/requirements.txt","sha256":"7d4421a90405b6fdd5662ef4373601d6e75e4b463ffbf4877b1e72ec744299db"},{"bytes":14988,"path":"experiments/run_mse.py","sha256":"d1c09ca5f33f6fa839cae8ed60a6526aeb8c7af1d96f76087f6aebc16f0970fc"},{"bytes":7305,"path":"experiments/simulate.cpp","sha256":"d40794c6695a67381d3e8fec83b4ff1a4218cf8b5311a883fcef868c03008126"},{"bytes":14567,"path":"latex/AI_paper.tex","sha256":"27bc248ca68b159fb69f31f3927e6f547771326e638d78221ee30c034c540041"},{"bytes":1461,"path":"latex/abstract.tex","sha256":"b0cc35040fba7473d59cff8cb03e0f7a89bdd7ebb76348604380360d80ceaf5a"},{"bytes":4605,"path":"latex/algorithm.tex","sha256":"0f87c4112baa38aece1402f8051ebd1218c0d790dad90fc692f990a97f785551"},{"bytes":1818,"path":"latex/appendix_experiments.tex","sha256":"7bc6cdbce0b5a9f0b4fcf2490cc7a32a543bb47b62ecb9ad0dda7781378d8a9a"},{"bytes":1267,"path":"latex/appendix_notation.tex","sha256":"8d769ce70105ab1dcfe4945a1c7167d9aaddf5b72787dfcbb6c6c0e49e18cd45"},{"bytes":18389,"path":"latex/appendix_proofs.tex","sha256":"a570622208146062fa945bb98976bb728dc72ae36edd0b64cac7ca25e1c9fa20"},{"bytes":2228,"path":"latex/examples.tex","sha256":"104a56f90eba7a8e9fe994aa9aef660078630b18ae27ffb69d67f8b3a3635da7"},{"bytes":4919,"path":"latex/experiments.tex","sha256":"614b15aed402431c036425a70f44626ec8a4c4b836dd889c78995603251d26f1"},{"bytes":35021,"path":"latex/figure_neural_network.pdf","sha256":"696a213b1f6771138be18710b224c5ce0fa2b555509a0ecc7eaf8fc5c0c150f9"},{"bytes":44431,"path":"latex/figure_regression.pdf","sha256":"4bfbe8f94a23958e74818ea94e1d75843b00fb152f647a2db8f6f5e7278731ac"},{"bytes":14155,"path":"latex/future_work.tex","sha256":"198d0d36be026683522c0448d4a9f331d1026044ed328b443509e58ffdb58fa1"},{"bytes":5750,"path":"latex/guide_en.tex","sha256":"fbc2bd514df2ec5f899155b4b9eb05c6527e58968fd20461dc37d2e93d32f97d"},{"bytes":5276,"path":"latex/guide_zh.tex","sha256":"9349059647387ed74df583f1137defc870573500ea34670a166ef738316a9783"},{"bytes":2776,"path":"latex/introduction.tex","sha256":"75e5da8a87e153e398440c50adab27fb921d8ceef76fadb8633a775ce638f7d7"},{"bytes":6685,"path":"latex/main_results.tex","sha256":"c38319df6540463b67e227ec451865b027c270c98bcea968a189e6f1e68fcd3e"},{"bytes":2744,"path":"latex/paper.tex","sha256":"a2bff5a012aba54d42259fa5bbc12b9bb65076e05aa04b3c29fc526400b68361"},{"bytes":4747,"path":"latex/references.bib","sha256":"1de9f2ce93d92334b9d7759150c4174b9bb558dbc9b7201a5e611d1569bbf1c1"},{"bytes":5186,"path":"latex/related_work.tex","sha256":"263d51958b322048745a0b2201d2241b8d06ab045b9cea456aa604afc6c8a71a"},{"bytes":7603,"path":"latex/theorem_graph.tex","sha256":"ad80b67203529be68ef6678008bc8c30f34c14583f89c442ca5d4f7b93a84a67"},{"bytes":3518,"path":"packaging.json","sha256":"f27324893455465b3385cc16fbf11247f4c3663e9e17186dc8f7720e42073e3a"},{"bytes":1001,"path":"takeaways.md","sha256":"5cf2a531a2fcb6dc41dc8c3aba1957e17ad1d3a77c89bcfdf3101e5e733bb64d"},{"bytes":2416,"path":"theorem_graph/README.md","sha256":"96730e32140f9423c17f9f8b5cd526d875514e9c90af2f7bed0c86d6d5b21d5e"},{"bytes":34382,"path":"theorem_graph/theorem_graph.json","sha256":"302a2b46f6c61ac2cc3cf3ce0cabdc90960083cc9e3829a9214200a5bb9bae1e"},{"bytes":5856,"path":"verification/5_review_en.txt","sha256":"0f1dfd47341a393a8b6d23e0097102669b453b122aca2e6b6dc768520b364513"},{"bytes":4760,"path":"verification/6_review_zh.txt","sha256":"ccdf5251302369e24e813fb31819deff4f521d1981112082ae37e4c2e497fda5"},{"bytes":2485,"path":"verification/review.md","sha256":"aa8800ef9a8345687b302127409b5ffb15b3e0dc4d77118d48346a8b373c0863"},{"bytes":3255,"path":"work.yaml","sha256":"d3d88afd37933308a658595d0c67624edcac144464c25398b94acc1ac975b214"}],"is_demo":false,"manifest":{"abstract":"For entire input scores under exponential derivative envelopes, amplitude control and suitable sampling moments, an imaginary Gaussian perturbation yields an unbiased clean-gradient oracle. The manuscript derives moment bounds, convergence rates and last-iterate and averaged CLTs, with a distinct averaging covariance at the inverse-time endpoint.","authors":["eggtea"],"field":"Optimization","keywords":["SGD","measurement error","complex correction"],"license":"CC-BY-4.0","main":"1_paper.pdf","references":[{"authors":"Nakamura, Takashi","doi":"10.1093/biomet/77.1.127","title":"Corrected score function for errors-in-variables models: Methodology and application to generalized linear models"},{"authors":"Novick, Steven J. and Stefanski, Leonard A.","doi":"10.1198/016214502760047005","title":"Corrected score estimation via complex variable simulation extrapolation"},{"authors":"Stefanski, Leonard A.","doi":"10.1080/03610928908830159","title":"Unbiased estimation of a nonlinear function of a normal mean with application to measurement-error models"},{"authors":"Konstantopoulos, Takis and Patie, Pierre and Sarkar, Rohan","doi":"10.5802/aif.3600","title":"A new class of solutions to the van Dantzig problem, the Lee--Yang property, and the Riemann hypothesis"},{"authors":"Woolsey, Nicholas and Huang, Xianzheng","doi":"10.1007/s10463-026-00984-x","title":"Debiased inference in errors-in-variables problems with non-Gaussian measurement error"},{"authors":"Cook, J. R. and Stefanski, Leonard A.","doi":"10.1080/01621459.1994.10476871","title":"Simulation-extrapolation estimation in parametric measurement error models"},{"authors":"Spicker, Dylan and Wallace, Michael P. and Yi, Grace Y.","doi":"10.1002/cjs.11777","title":"Nonparametric simulation extrapolation for measurement-error models"},{"authors":"Belomestny, Denis and Goldenshluger, Alexander","doi":"10.1214/20-AOS1969","title":"Density deconvolution under general assumptions on the distribution of measurement errors"},{"authors":"Loh, Po-Ling and Wainwright, Martin J.","doi":"10.1214/12-AOS1018","title":"High-dimensional regression with noisy and missing data: Provable guarantees with nonconvexity"},{"authors":"Carroll, Raymond J. and Ruppert, David and Stefanski, Leonard A. and Crainiceanu, Ciprian M.","doi":"10.1201/9781420010138","title":"Measurement Error in Nonlinear Models: A Modern Perspective"},{"authors":"Fabian, V\u00e1clav","doi":"10.1214/aoms/1177698258","title":"On Asymptotic Normality in Stochastic Approximation"},{"authors":"Polyak, B. T. and Juditsky, A. B.","doi":"10.1137/0330046","title":"Acceleration of Stochastic Approximation by Averaging"},{"authors":"Robbins, Herbert and Siegmund, David","doi":"10.1016/B978-0-12-604550-5.50015-8","title":"A Convergence Theorem for Non Negative Almost Supermartingales and Some Applications"}],"schema_version":1,"takeaways":"takeaways.md","title":"Complex-Corrected Stochastic Gradient Descent for Noisy Observations","type":"paper","verification":[{"date":"2026-10-17","model":"GPT-6 (Codex; exact runtime variant not exposed)","path":"verification/review.md","scope":"Statement, assumptions, public-summary and supplied-review provenance audit; not a complete fresh proof review.","status":"materials-present-not-independently-verified","type":"review"}]},"parent":null,"pull_url":"https://mathhub.eggtea.org/api/works/complex-corrected-stochastic-gradient-descent-3ac3e220/versions/1/bundle","resolves":[],"sha256":"79efcba8a5d420be1b102f8353f5412fa168cd4fd31835ace6ffab7fb410e293","submitted_by":"eggtea","url":"https://mathhub.eggtea.org/works/complex-corrected-stochastic-gradient-descent-3ac3e220/v/1","version":1,"work_id":"complex-corrected-stochastic-gradient-descent-3ac3e220"}