arXiv:2608.06411v2 Announce Type: replace Abstract: Multimodal large language models (MLLMs) achieve strong performance across diverse vision-language tasks, but their efficiency is limited by the cost of processing numerous visual tokens. Visual token pruning can reduce this cost, but requires accurate token importance estimates. Recent studies have demonstrated that text-to-vision attention from middle language model layers can effectively guide visual token pruning, typically using attention from a predefined middle layer to select the visual tokens to retain. Two problems therefore remain. First, our analysis shows that the layer whose attention is most responsive to the question varies substantially across samples, making a fixed layer suboptimal. Second, obtaining attention from the appropriate middle layer requires processing numerous visual tokens through several language model layers, by which point considerable computation has already been spent. To address both problems, we propose Middle-layer Attention Prediction (MAP), which uses Question Contrastive Teacher Selection to identify a sample-specific teacher layer by contrasting attention under the original and reference questions, and distills attention from the selected layer into a lightweight predictor that estimates visual token importance from multi-modal input features. During inference, MAP combines the predicted importance scores with a diversity criterion to prune visual tokens before the first language model layer. Thus, MAP requires no attention maps for pruning and remains compatible with existing inference acceleration techniques. Across ten benchmarks on LLaVA-NeXT-7B, MAP retains 97.5% of the unpruned model performance with only 5.56% of the visual tokens, yielding a 3.09x end-to-end speedup.

npj Digital Medicine, Published online: 09 September 2026; doi:10.1038/s41746-026-03235-8

DystoniaDBSNet as a novel deep learning biomarker of predictive deep brain stimulation outcome in dystonia

Cell Death Discovery, Published online: 11 April 2026; doi:10.1038/s41420-026-03102-1

CX3CR1-mediated immune networks in sepsis: implications for precision therapy

Oncogene, Published online: 02 April 2026; doi:10.1038/s41388-026-03761-5

Strand-asymmetric G-runs and G4s downstream of TSS modulate tumor suppressor gene transcription

Nature, Published online: 01 April 2026; doi:10.1038/s41586-026-10446-2

Structural modifications in strain-engineered bilayer nickelate thin films

Nature, Published online: 31 March 2026; doi:10.1038/s41586-026-10335-8

Author Correction: Signatures of ambient pressure superconductivity in thin film La₃Ni₂O₇

arXiv:2603.08034v1 Announce Type: cross Abstract: Emotion recognition in real-world environments is hindered by partial occlusions, missing modalities, and severe class imbalance. To address these issues, particularly for the Affective Behavior Analysis in-the-wild (ABAW) Expression challenge, we propose a multimodal framework that dynamically fuses visual and audio representations. Our approach uses a dual-branch Transformer architecture featuring a safe cross-attention mechanism and a modality dropout strategy. This design allows the network to rely on audio-based predictions when visual cues are absent. To mitigate the long-tail distribution of the Aff-Wild2 dataset, we apply focal loss optimization, combined with a sliding-window soft voting strategy to capture dynamic emotional transitions and reduce frame-level classification jitter. Experiments demonstrate that our framework effectively handles missing modalities and complex spatiotemporal dependencies, achieving an accuracy of 60.79% and an F1-score of 0.5029 on the Aff-Wild2 validation set.