Multimodal Depression Detection from Speech and Text: Public Reproducibility Package for ASOC Submission

This record provides a public, share-safe reproducibility package for the ASOC submission on multimodal depression detection. The uploaded archive research_data.zip includes: (i) base manifests and subject-disjoint fold split files, (ii) processed result/summary tables, (iii) key Python and shell scripts used in the experiment workflow, (iv) selected experiment configuration files, and (v) environment requirements.Raw audio/text data are not redistributed because they originate from third-party datasets (EATD-Corpus, MODMA, and DAIC-WOZ) and remain subject to their original access and licensing terms.This package supports policy-compliant methodological transparency while avoiding redistribution of restricted source data.