[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85782-en":3,"doc-seo-85782-105":28,"detail-sidebar-cat-0-en-105":90},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":11,"language":21,"language_code":22,"site_id":23,"html_lang":22,"table_of_contents":24,"faqs":25,"seo_title":13,"seo_description":14,"update_tm":26,"read_time":27},85782,8796095462418,"Noah","https://ap-avatar.wpscdn.com/avatar/80000253c1241d02b47?x-image-process=image/resize,m_fixed,w_180,h_180&k=1778826106357471780",8,"Research & Report","ARIMA Reconstruction-Grounded Predictive Representation Learning for Symbolic Music","Self-supervised learning for symbolic music has advanced through token-level pretraining, yet such representations remain dependent on tokenizer-specific sequences and often yield time-span embeddings only indirectly. This paper introduces ARIMA, a reconstruction-grounded latent predictive framework that learns compact window-based representations from data. ARIMA encodes each fixed-duration window into a continuous latent, trains a causal predictor via contrastive next-latent prediction, and grounds learning through structured reconstruction of music elements. Evaluation shows efficiency and strong performance on tasks covering harmony, timing, and cross-performance retrieval, with ablations highlighting next-latent prediction and stable structured reconstruction without explicit variance regularization.","ARIMA: RECONSTRUCTION-GROUNDED PREDICTIVE REPRESENTATION LEARNING FOR SYMBOLIC MUSIC  \nMingyang Yao Zhaoxiang Feng  \nUniversity of California, San Diego  \n{m5yao, [zhf004}@ucsd.edu](zhf004}@ucsd.edu)  \narXiv :2607 . 10003v 1 [ cs . SD] 10 Jul 2026  \nABSTRACT  \nSelf-supervised learning for symbolic music has advanced largely through token-level pretraining, but such representations remain tied to tokenizer-specific sequences and often provide time-span-level embeddings only indirectly. In this paper, we propose ARIMA, a reconstruction-grounded latent predictive framework for symbolic music that learns compact window-based representations directly from data. ARIMA encodes each fixed-duration window into a continuous latent representation, trains a causal predictor with contrastive next-latent prediction, and grounds the encoder through structured reconstruction of music elements. This design preserves local musical details while modeling temporal progression across windows. We evaluate ARIMA on downstream tasks spanning various levels of music understanding. Results show that ARIMA is particularly efficient and effective on tasks involving harmonic, timing, and cross-performance retrieval, while remaining competitive with much larger baselines on other tasks. Ablations further show that next-latent prediction is essential for temporally integrated representations, and that structured reconstruction stabilizes latent learning without requiring explicit variance regularization. The code is here 1  \n1. INTRODUCTION  \nSelf-supervised learning (SSL) has become an important approach for learning transferable representations from unlabeled symbolic music. Recent works such as MidiBERTPiano [1], MusicBERT [2], PianoBART [3], and Aria [4] adapt masked language modeling or autoregressive objectives to MIDI tokenizations [5], including REMI [6], CP [7], OctupleMIDI [2] and other variants. In parallel, contrastive models such as CLaMP [8] learn joint representations between symbolic music and natural language, enabling semantic retrieval and zero-shot music classification. These studies demonstrate the value of large-scale SSL for music understanding and generation.  \nDespite this progress, most existing symbolic SSL remain closely tied to token-level or cross-modal alignment objectives. Token-based models learn from tokenizerspecific event sequences. While this is effective for generation and note-level classification, time-span-level representations are usually obtained only indirectly through pooling or task-specific adaptation. Moreover, the token counts  \n1 [https://github.com/AndyWeasley2004/symbolic_music_wm](https://github.com/AndyWeasley2004/symbolic_music_wm)  \ncorresponding to the same temporal span can vary substantially with note density, time resolution, and tokenizer design, making musical timings less explicit. Cross-modal models such as CLaMP provide higher-level semantic supervision, but their objective primarily captures languageassociated musical semantics rather than the temporal evolution of music. This leaves a gap for self-supervised objectives that learn compact time-span-level representations while directly modeling progression across time.  \nA related direction has been explored through jointembedding predictive architectures (JEPA) [9–11] under the world model domain. Instead of predicting the next frame as raw images or discrete tokens, JEPA-style models learn by predicting the next state latent from context and actions. Specifically, an encoder maps each input image to an embedding, while a predictor models causal dependencies among embeddings in latent space. This provides a natural analogy for symbolic music: local windows can be encoded as musical states, and self-supervised prediction across windows can encourage the model to learn how these states progress over time.  \nHowever, symbolic music also differs from vision and video in an important aspect. Low-level information such as pitch, onset patterns, sus","cbCainUeIKGuByNQ","https://ap.wps.com/l/cbCainUeIKGuByNQ","pdf",523701,1,"English","en",105,"# Abstract\n# Introduction\n# Related Work\n## Self-supervised learning for symbolic music","[{\"question\":\"What problem does ARIMA address in symbolic music self-supervised learning?\",\"answer\":\"Existing token-level objectives tie representations to tokenizer-specific event sequences and often provide time-span embeddings only indirectly. ARIMA targets compact window-level representations while explicitly modeling progression across time.\"},{\"question\":\"How does ARIMA learn representations from symbolic music windows?\",\"answer\":\"ARIMA encodes each fixed-duration window into a continuous latent using an encoder grounded by reconstruction of pianoroll, chroma, and velocity. It then models temporal progression with a causal predictor using contrastive next-latent prediction.\"},{\"question\":\"What do the evaluation and ablation results indicate?\",\"answer\":\"ARIMA is efficient and effective on tasks involving harmony, timing, and cross-performance retrieval, and remains competitive elsewhere. Ablations show next-latent prediction is essential for temporally integrated representations, while structured reconstruction stabilizes latent learning without explicit variance regularization.\"}]",1784206249,20,{"code":4,"msg":29,"data":30},"ok",{"site_id":23,"language":22,"slug":31,"title":13,"keywords":32,"description":14,"schema_data":33,"social_meta":85,"head_meta":87,"extra_data":89,"updated_unix":26},"arima-reconstruction-grounded-predictive-representation-learning-for-symbolic-music","",{"@graph":34,"@context":84},[35,52,67],{"@type":36,"itemListElement":37},"BreadcrumbList",[38,42,46,49],{"item":39,"name":40,"@type":41,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":43,"name":44,"@type":41,"position":45},"https://docshare.wps.com/document/","Document",2,{"item":47,"name":12,"@type":41,"position":48},"https://docshare.wps.com/document/research-report/",3,{"item":50,"name":13,"@type":41,"position":51},"https://docshare.wps.com/document/arima-reconstruction-grounded-predictive-representation-learning-for-symbolic-music/85782/",4,{"url":50,"name":13,"@type":53,"author":54,"headline":13,"publisher":56,"fileFormat":59,"inLanguage":22,"description":14,"dateModified":60,"datePublished":61,"encodingFormat":59,"isAccessibleForFree":62,"interactionStatistic":63},"DigitalDocument",{"name":9,"@type":55},"Person",{"url":39,"name":57,"@type":58},"DocShare","Organization","application/pdf","2026-07-17","2026-07-16",true,{"@type":64,"interactionType":65,"userInteractionCount":20},"InteractionCounter",{"@type":66},"ViewAction",{"@type":68,"mainEntity":69},"FAQPage",[70,76,80],{"name":71,"@type":72,"acceptedAnswer":73},"What problem does ARIMA address in symbolic music self-supervised learning?","Question",{"text":74,"@type":75},"Existing token-level objectives tie representations to tokenizer-specific event sequences and often provide time-span embeddings only indirectly. ARIMA targets compact window-level representations while explicitly modeling progression across time.","Answer",{"name":77,"@type":72,"acceptedAnswer":78},"How does ARIMA learn representations from symbolic music windows?",{"text":79,"@type":75},"ARIMA encodes each fixed-duration window into a continuous latent using an encoder grounded by reconstruction of pianoroll, chroma, and velocity. It then models temporal progression with a causal predictor using contrastive next-latent prediction.",{"name":81,"@type":72,"acceptedAnswer":82},"What do the evaluation and ablation results indicate?",{"text":83,"@type":75},"ARIMA is efficient and effective on tasks involving harmony, timing, and cross-performance retrieval, and remains competitive elsewhere. Ablations show next-latent prediction is essential for temporally integrated representations, while structured reconstruction stabilizes latent learning without explicit variance regularization.","https://schema.org",{"og:url":50,"og:type":86,"og:title":13,"og:site_name":57,"og:description":14},"article",{"robots":88,"canonical":50},"index,follow",{"doc_id":7,"site_id":23},{"code":4,"msg":5,"data":91},[92,96,100,104,109,114,119,122,126,129,133],{"id":20,"doc_module":4,"doc_module_name":44,"category_name":93,"show_sort_weight":94,"slug":95},"Story & Novel",90,"story-novel",{"id":45,"doc_module":4,"doc_module_name":44,"category_name":97,"show_sort_weight":98,"slug":99},"Literature",80,"literature",{"id":51,"doc_module":4,"doc_module_name":44,"category_name":101,"show_sort_weight":102,"slug":103},"Exam",70,"exam",{"id":105,"doc_module":4,"doc_module_name":44,"category_name":106,"show_sort_weight":107,"slug":108},5,"Comic",60,"comic",{"id":110,"doc_module":4,"doc_module_name":44,"category_name":111,"show_sort_weight":112,"slug":113},6,"Technology",50,"technology",{"id":115,"doc_module":4,"doc_module_name":44,"category_name":116,"show_sort_weight":117,"slug":118},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":44,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":44,"category_name":124,"show_sort_weight":27,"slug":125},9,"Religion & Spirituality","religion-spirituality",{"id":27,"doc_module":4,"doc_module_name":44,"category_name":127,"show_sort_weight":27,"slug":128},"World Cup","world-cup",{"id":130,"doc_module":4,"doc_module_name":44,"category_name":131,"show_sort_weight":130,"slug":132},10,"Lifestyle","lifestyle",{"id":134,"doc_module":4,"doc_module_name":44,"category_name":135,"show_sort_weight":105,"slug":136},19,"General","general"]