Modular Delta Merging with Orthogonal Constraints: A Scalable Framework for Continual and Reversible Model Composition

[2507.20997] Modular Delta Merging with Orthogonal Constraints: A Scalable Framework for Continual and Reversible Model Composition Skip to main content Search Submit Donate Log in Search arXiv Press Enter to search · Advanced search Computer Science > Machine Learning arXiv:2507.20997 (cs) [Submitted on 28 Jul 2025 ( v1 ), last revised 11 Apr 2026 (this version, v4)] Title: Modular Delta Merging with Orthogonal Constraints: A Scalable Framework for Continual and Reversible Model Composition Authors: Haris Khan , Sadia Asif , Shumaila Asif , Muhammad Zeeshan Karamat , Rajesh Upadhayaya View a PDF of the paper titled Modular Delta Merging with Orthogonal Constraints: A Scalable Framework for Continual and Reversible Model Composition, by Haris Khan and 4 other authors View PDF HTML (experimental) Abstract: In real-world machine learning deployments, models must be continually updated, composed, and when required, selectively undone. However, existing approaches to model merging and continual learning often suffer from task interference, catastrophic forgetting, or lack of reversibility. We propose Modular Delta Merging with Orthogonal Constraints (MDM-OC), a novel framework that enables scalable, interference-free, and reversible composition of fine-tuned models. Each task-specific model is encoded as a delta from a shared base and projected into an orthogonal subspace to eliminate conflict. These projected deltas are then merged via gradient-based optimization to form a unified model that retains performance across tasks. Our approach supports continual integration of new models, structured unmerging for compliance such as GDPR requirements, and model stability via elastic weight consolidation and synthetic replay. Extensive experiments on vision and natural language processing benchmarks demonstrate that MDM-OC outperforms prior baselines in accuracy, backward transfer, and unmerge fidelity, while remaining memory-efficient and computationally tractable. This framework offers a principled solution for modular and compliant AI system design. Subjects: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI) Cite as: arXiv:2507.20997 [cs.LG] (or arXiv:2507.20997v4 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2507.20997 Focus to learn more arXiv-issued DOI via DataCite Submission history From: Haris Khan [ view email ] [v1] Mon, 28 Jul 2025 17:08:49 UTC (1,172 KB) [v2] Thu, 7 Aug 2025 06:44:42 UTC (1,172 KB) [v3] Sat, 21 Feb 2026 10:25:48 UTC (1,246 KB) [v4] Sat, 11 Apr 2026 17:15:56 UTC (1,245 KB) Full-text links: Access Paper: View a PDF of the paper titled Modular Delta Merging with Orthogonal Constraints: A Scalable Framework for Continual and Reversible Model Composition, by Haris Khan and 4 other authors View PDF HTML (experimental) TeX Source view license Current browse context: cs.LG < prev | next > new | recent | 2025-07 Change to browse by: cs cs.AI References & Citations NASA ADS Google Scholar Semantic Scholar export BibTeX citation Loading... BibTeX formatted citation × loading... Data provided by: Bookmark Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer ( What is the Explorer? ) Connected Papers Toggle Connected Papers ( What is Connected Papers? ) Litmaps Toggle Litmaps ( What is Litmaps? ) scite.ai Toggle scite Smart Citations ( What are Smart Citations? ) Code, Data, Media Code, Data and Media Associated with this Article alphaXiv Toggle alphaXiv ( What is alphaXiv? ) Links to Code Toggle CatalyzeX Code Finder for Papers ( What is CatalyzeX? ) DagsHub Toggle DagsHub ( What is DagsHub? ) GotitPub Toggle Gotit.pub ( What is GotitPub? ) Huggingface Toggle Hugging Face ( What is Huggingface? ) ScienceCast Toggle ScienceCast ( What is ScienceCast? ) Demos Demos Replicate Toggle Replicate ( What is Replicate? ) Spaces Toggle Hugging Face Spaces ( What is Spaces? ) Spaces Toggle TXYZ.AI ( What is TXYZ.AI? ) Related Papers Recommenders and Search Tools Link to Influence Flower Influence Flower ( What are Influence Flowers? ) Core recommender toggle CORE Recommender ( What is CORE? ) IArxiv recommender toggle IArxiv Recommender ( What is IArxiv? ) Author Venue Institution Topic About arXivLabs arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs . Which authors of this paper are endorsers? | Disable MathJax ( What is MathJax? ) We gratefully acknowledge support from our major funders , member institutions , , and all contributors. About · Help · Contact · Subscribe · Copyright · Privacy · Accessibility · Operational Status (opens in new tab) Major funding support from

Related documents

• Linear time mixing and pre-cutoff for oriented kinetically constrained models#1104784
• Validating Hybrid-State Cache Recovery for GLM-5.3-Flash with vLLM and LMCache#1084925
• Validating Hybrid-State Cache Recovery for GLM-5.3-Flash with vLLM and LMCache#919349
• Information geometry of emergent symmetry quotients and their weak unfoldings#1077189
• Long-Range Interacting Many-Body Systems in the Irrep Basis#1070433