We address the Ambivalence/Hesitancy (A/H) Video Recognition Challenge at the 10th ABAW Competition (CVPR 2026). We propose a divergence-based multimodal fusion that explicitly measures cross-modal conflict between visual, audio, and textual channels. Visual features are encoded as Action Units (AUs) extracted via Py-Feat, audio via Wav2Vec 2.0, and text via BERT. Each modality is processed by a BiLSTM with attention pooling and projected into a shared embedding space. The fusion module computes pairwise absolute differences between modality embeddings, directly capturing the incongruence that characterizes A/H. On the BAH dataset, our approach achieves a Macro F1 of 0.6808 on the validation test set, outperforming the challenge baseline of 0.2827. Statistical analysis across 1{,}132 videos confirms that temporal variability of AUs is the dominant visual discriminator of A/H.
@article{arxiv.2603.16939,
title = {Solution for 10th Competition on Ambivalence/Hesitancy (AH) Video Recognition Challenge using Divergence-Based Multimodal Fusion},
author = {Aislan Gabriel O. Souza and Agostinho Freire and Leandro Honorato Silva and Igor Lucas B. da Silva and João Vinícius R. de Andrade and Gabriel C. de Albuquerque and Lucas Matheus da S. Oliveira and Mário Stela Guerra and Luciana Machado},
journal= {arXiv preprint arXiv:2603.16939},
year = {2026}
}