Lluna: Collaborative Unimodal Models for Missing-Modality Inference
Summary
Lluna improves multimodal predictions when some input modalities are missing by combining private unimodal predictions with a graph-based residual collaboration branch and progressive consistency training, achieving a 4.26% average improvement in Mean-Missing over baselines across eight datasets.
View Cached Full Text
Cached at: 10/15/26, 11:00 PM
Paper page - Lluna: Collaborative Unimodal Models for Missing-Modality Inference
Source: https://huggingface.co/papers/2610.20526
Abstract
Lluna improves multimodal predictions under missing inputs by combining private unimodal predictions with graph-based residual collaboration and progressive consistency training.
Pretrainedunimodal modelsoffer a flexible route tomultimodal predictionthrough collaboration, but this collaboration can become less accurate than a surviving unimodal model when modalities are missing. We study this setting using fixed surviving-model anchors and find that collaborative gains can weaken or reverse as inputs disappear, even when the anchor’s own input is unchanged. Our diagnostics further reveal reduced recovery of modality-specific information after aggregation and state-dependentcollaborative representations. We proposeLluna, which combines a direct private prediction path from adapted unimodal features with a commongraph-based collaborationbranch that provides a residual correction.Progressive Collaboration Consistencyorganizes richer-to-poorer guidance along single-modality removals, aligning normalized common representations while task supervision adapts predictions to the available evidence. Across eight datasets spanning affective computing, scene understanding, and medical prediction,Llunaachieves the best evaluated average performance over non-full input states, improving Mean-Missing by 4.26\% on average relative to the strongest evaluated baseline. The fixed-anchor diagnostic also shows positive gains in mean collaboration under minimal inputs on all eight tasks. Code is available at https://github.com/AIFrontierLab/Lluna.
View arXiv pageView PDFGitHub1Add to collection
Get this paper in your agent:
hf papers read 2610\.20526
Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash
Models citing this paper0
No model linking this paper
Cite arxiv.org/abs/2610.20526 in a model README.md to link it from this page.
Datasets citing this paper0
No dataset linking this paper
Cite arxiv.org/abs/2610.20526 in a dataset README.md to link it from this page.
Spaces citing this paper0
No Space linking this paper
Cite arxiv.org/abs/2610.20526 in a Space README.md to link it from this page.
Collections including this paper0
No Collection including this paper
Add this paper to acollectionto link it from this page.
Similar Articles
Large Language Models as Unified Multimodal Learners for Clinical Prediction
The paper proposes converting multimodal patient data (text, labs, vitals) into a single natural language sequence and fine-tuning LLMs for clinical prediction, achieving comparable or better performance than specialized fusion architectures across three tasks.
LoMo: Local Modality Substitution for Deeper Vision-Language Fusion
LoMo proposes a data curation method that reformulates single-modality prompts into interleaved multimodal sequences to improve cross-modal representation alignment in vision-language models, achieving consistent gains on multiple benchmarks.
Lance: Unified Multimodal Modeling by Multi-Task Synergy
Lance is a unified multimodal model that leverages a dual-stream mixture-of-experts architecture and collaborative multi-task training to achieve strong performance in understanding, generation, and editing of both images and videos, outperforming existing open-source unified models.
MSP: Modality Self-Play for Training Multimodal LLMs Without Paired Cross-Modal Data
The paper introduces Modality Self-Play (MSP), a training framework that enables multimodal foundation models to compose across modalities without real paired cross-modal data, using proxy textual descriptions as semantic bridges and achieving state-of-the-art or competitive results on multiple multimodal benchmarks.
C3-UniMM: Causal Cycle-Consistent Unified Multimodal Modeling via Super Alignment and Shared Decoding Space
This paper introduces C3-UniMM, a unified multimodal modeling framework based on causal cycle consistency and super alignment to improve cross-modal understanding and generation with enhanced structural consistency.