Research map / Vision-language and multimodal / Cross-modal learning
LIMSSR: LLM-Driven Sequence-to-Score Reasoning under Training-Time Incomplete Multimodal Observations
Huangbiao Xu, Huanqi Wu, Xiao Ke, Yuxin Peng
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add LLM backbone to baseline (ablation) | LIMSSR baseline · FS1000 average | Spearman correlation: 0.528 -> 0.601 | Improved |
| add PCMI prompt-guided imputation (ablation) | LIMSSR · FS1000 average | Spearman correlation: 0.601 -> 0.687 | Improved |
| add LMRF fusion (ablation) | LIMSSR · FS1000 average | Spearman correlation: 0.687 -> 0.768 | Improved |
| add MDA dual-path aggregation (ablation) | LIMSSR · FS1000 average | MSE: 18.18 -> 14.08 | Improved |
| add LMRF (ablation) | LIMSSR · FS1000 average | Spearman correlation: 0.732 -> 0.789 | Improved |
| add PCMI (ablation) | LIMSSR · Fis-V average | Spearman correlation: 0.699 -> 0.743 | Improved |
| add contrastive loss L_con (ablation) | LIMSSR · FS1000 average | Spearman correlation: 0.759 -> 0.789 | Improved |
| add regularization loss L_reg (ablation) | LIMSSR · FS1000 average | Spearman correlation: 0.773 -> 0.789 | Improved |
| add both MDA paths instead of Path 1 only (ablation) | LIMSSR · FS1000 average | Spearman correlation: 0.778 -> 0.789 | Improved |
| add all prompt components (ablation) | LIMSSR · FS1000 average | Spearman correlation: 0.710 -> 0.789 | Improved |
| increase fusion tokens K from 2 to 4 (ablation) | LIMSSR · FS1000 average | Spearman correlation: 0.762 -> 0.789 | Improved |
| increase fusion tokens K from 4 to 6 (ablation) | LIMSSR · FS1000 average | Spearman correlation: 0.789 -> 0.779 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- MCMoE — MCMoE: Completing Missing Modalities with Mixture of Experts for Incomplete Multimodal Action Quality Assessment (AAAI 2026) · better on 9 of 9 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- TokenSwap: Benchmarking and Reducing the Modality Gap in Multimodal LLMs (NeurIPS 2026 · similarity 0.88)
- Revisiting the Data Sampling in Multimodal Post-training from a Difficulty-Distinguish View (AAAI 2026 · similarity 0.88)
- Towards General Continuous Memory for Vision-Language Models (NeurIPS 2025 · similarity 0.88)
- Retrieval-Augmented Dynamic Prompt Tuning for Incomplete Multimodal Learning (AAAI 2025 · similarity 0.87)
- MSP: Modality Self-Play for Training Multimodal LLMs Without Paired Cross-Modal Data (NeurIPS 2026 · similarity 0.87)
- LENS: Multi-level Evaluation of Multimodal Reasoning with Large Language Models (ICLR 2026 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
