Metraipolis › archive › dc_774483df0833

TTSD-FAR: Test-Time Self-Distillation with Fisher-Anchored Restoration for Missing-Modality Emotion Recognition in LVLMs — arXiv:2608.18386

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.18386
License (as archived)
CC0-1.0
Archived
2026-10-18T13:45:04.046Z
SHA-256 of the archived text
e058d94d932da35c2f146620cc375b9b0990f40464b18d321634688e4441e5b4
Stable URL of this page
https://metraipolis.com/documento/dc_774483df0833

How to cite

TTSD-FAR: Test-Time Self-Distillation with Fisher-Anchored Restoration for Missing-Modality Emotion Recognition in LVLMs — arXiv:2608.18386 — arXiv (en), https://arxiv.org/abs/2608.18386 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_774483df0833 (Metraipolis archive dc_774483df0833, sha256 e058d94d932da35c2f146620cc375b9b0990f40464b18d321634688e4441e5b4)
BibTeX
@misc{dc774483df0833,
  title = {TTSD-FAR: Test-Time Self-Distillation with Fisher-Anchored Restoration for Missing-Modality Emotion Recognition in LVLMs — arXiv:2608.18386},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.18386},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_774483df0833; sha256 e058d94d932da35c2f146620cc375b9b0990f40464b18d321634688e4441e5b4},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.18386. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large video-language models (LVLMs) have shown remarkable performance on multimodal tasks like multimodal emotion recognition (ER) in the wild. ER is inherently multimodal, requiring a joint understanding of facial expressions, vocalizations, language, biosignals, and gestures. However, real-world deployment remains challenging: modalities may be missing or noisy at test time. Partial observations can be viewed as a distribution shift relative to the complete-modality distribution. SOTA TTA methods based on entropy minimization or perplexity reduction do not transfer to autoregressive LVLMs, while retrieval augmented generation (RAG) degrades when the observed modality is weak. Because no ground-truth supervision exists to verify individual updates, adaptation across this stream risks accumulating drift and degrading once the model departs from a reliable solution. An effective solution must therefore adapt to arbitrary missing-modality patterns and remain effective during continual adaptation. We address both jointly with Test-Time Self-Distillation (TTSD), a parameter-efficient framework in which a frozen teacher, trained on complete modalities, guides an adaptive low-rank student via self-distillation, updating only a negligible number of parameters. Stability is built into this same loop through Fisher-Anchored Restoration (FAR), which monitors Fisher information stability to detect convergence versus drift and restores the student toward the teacher's anchor when distributional shifts are identified. Our experiments on MELD, DFEW, and BAH under 0%-50% missing modalities show that this unified adaptation-restoration design consistently outperforms entropy-based adaptation, RAG, and perplexity-based generation over long adaptation horizons, where baselines without restoration progressively degrade while TTSD-FAR remains consistent.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-20 in cs.AI).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2608.18386 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-18T13:41:17.936Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-20?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).