Multi-Scale Fusion with Explicit Word-Level Alignment for Multimodal Sentiment Analysis
Bibliographic record
Abstract
Multimodal sentiment analysis (MSA) integrates and processes data from multiple sources, like audio and text, to better understand human emotions through cross-modal interactions. Previous research generally obtain a feature embedding from different modalities and fuse at the utterance level directly, making it difficult to capture fine-grained multimodal representations and susceptible to irrelevant information from heterogeneous modalities. Moreover, these methods often overlook explicit word-level alignment between modalities, which limits effective representation learning and multimodal fusion. Therefore, we propose a novel method, Fine-grained Multimodal Fusion Network(MMTA) for MSA. Specifically, a fine-grained representation alignment module is introduced to extract representation at the word level by montreal forced aligner (MFA). A local-global multi-scale fusion method is then designed to further capture more fine-grained interaction from local features, i.e., word-level features. Finally, we also investigate the effectiveness of incorporating word-level contextual information into the fine-grained fusion process. Extensive experiments on the public MSA datasets, i.e., MOSI and MOSEI, show that our approach surpass previous baselines.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.004 | 0.002 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".