Listen, Look, and Learn: Learning Without Forgetting through SAM-Audio

Abstract

Class-Incremental Learning (CIL) aims to continuously learn new classes without forgetting previously acquired knowledge. While recent CIL advances have spurred significant interest across various modalities, the audio-visual setting remains underexplored. Furthermore, although foundational multimodal models like SAM-Audio encapsulate rich static priors, our empirical analysis reveals that these representations struggle in incremental settings. This work bridges this gap by integrating SAM-Audio's audio-visual priors into the CIL setting. Specifically, we leverage its dense audio and visual representations and employ a novel guided attention strategy where the audio features contextually guide the visual representations. To further mitigate catastrophic forgetting, we introduce dual-level distillation objectives at both the feature and logit levels. Extensive evaluations on audio-visual CIL benchmarks demonstrate that our approach consistently outperforms state-of-the-art methods.

Workflow Status

Review status

pending

Role

unreviewed

Read priority

now

Vote

Not set.

Saved

no

Collections

Not filed yet.

Next action

Not filled yet.

Reading Brief

No structured notes yet. Add `summary_sections`, `why_relevant`, `claim_impact`, or `next_action` in `papers.jsonl` to enrich this view.

Why It Surfaced

No ranking explanation is available yet.

Tags

No tags.

Similar Papers

BibTeX

@article{gupta2026listen,
  title = {Listen, Look, and Learn: Learning Without Forgetting through SAM-Audio},
  author = {Avi Gupta and Nilotpal Sinha and Vishnu Raj and Sambuddha Saha and Pratik Joshi and Koteswar Rao Jerripothula and Tammam Tillo},
  year = {2026},
  abstract = {Class-Incremental Learning (CIL) aims to continuously learn new classes without forgetting previously acquired knowledge. While recent CIL advances have spurred significant interest across various modalities, the audio-visual setting remains underexplored. Furthermore, although foundational multimodal models like SAM-Audio encapsulate rich static priors, our empirical analysis reveals that these representations struggle in incremental settings. This work bridges this gap by integrating SAM-Audio},
  url = {https://arxiv.org/abs/2606.10887},
  keywords = {cs.CV},
  eprint = {2606.10887},
  archiveprefix = {arXiv},
}

Metadata

{}