Pith Number

pith:ETZAOYRG

pith:2026:ETZAOYRG4KQ3R3VFPO6JBRNXAE

not attested not anchored not stored refs resolved

SpurAudio: A Benchmark for Studying Shortcut Learning in Few-Shot Audio Classification

Giries Abu Ayoub, Loay Mualem, Morad Tukan

A new benchmark shows few-shot audio classifiers suffer sharp drops when background correlations are broken, even in large pretrained models.

arxiv:2605.13672 v1 · 2026-05-13 · cs.CV

Open paper page JSON Open Graph Bundle Merged state Verified badge What is a Pith Number?

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{ETZAOYRG4KQ3R3VFPO6JBRNXAE}

Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge

Record completeness

1 Bitcoin timestamp

2 Internet Archive

3 Author claim open · sign in to claim

4 Citations open

5 Replications open

✓ Portable graph bundle live · download bundle · merged state

The bundle contains the canonical record plus signed events. A mirror can host it anywhere and recompute the same current state with the deterministic merge algorithm.

Claims

C1strongest claim

many state-of-the-art few-shot methods suffer severe performance degradation when background correlations are disrupted, despite achieving similar accuracy under standard evaluation protocols. Crucially, this vulnerability persists even in large pretrained audio foundation models.

C2weakest assumption

The assumption that foreground events and background environments in audio are naturally separable in a way that permits controlled, multi-level contextual shifts representative of real-world conditions.

C3one line summary

SpurAudio benchmark shows state-of-the-art few-shot audio classifiers suffer large performance drops when background correlations are disrupted, even in large pretrained models.

References

78 extracted · 78 resolved · 4 Pith anchors

[1] Classifying sounds in polyphonic urban sound scenes.AES E-Library 2022

[2] How robust are audio embeddings for polyphonic sound event tagging?IEEE/ACM Transactions on Audio, Speech, and Language Processing, 31:2658–2667, 2023 2023

[3] wav2vec 2.0: A framework for self-supervised learning of speech representations.Advances in neural information processing systems, 33:12449–12460 2020

[4] Meta-learning with task-adaptive loss function for few-shot learning 2021

[5] Meta-learning with differentiable closed-form solvers 2018 · arXiv:1805.08136

Formal links

2 machine-checked theorem links

Receipt and verification

First computed	2026-05-18T02:44:17.158648Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

24f2076226e2a1b8eea57bbc90c5b7011722e4c1f8ca587baea068a9e6471c8b

Aliases

arxiv: 2605.13672 · arxiv_version: 2605.13672v1 · doi: 10.48550/arxiv.2605.13672 · pith_short_12: ETZAOYRG4KQ3 · pith_short_16: ETZAOYRG4KQ3R3VF · pith_short_8: ETZAOYRG

Agent API

Resolver JSON Graph JSON Events JSON Schema Signing key

Verify this Pith Number yourself

curl -sH 'Accept: application/ld+json' https://pith.science/pith/ETZAOYRG4KQ3R3VFPO6JBRNXAE \
  | jq -c '.canonical_record' \
  | python3 -c "import sys,json,hashlib; b=json.dumps(json.loads(sys.stdin.read()), sort_keys=True, separators=(',',':'), ensure_ascii=False).encode(); print(hashlib.sha256(b).hexdigest())"
# expect: 24f2076226e2a1b8eea57bbc90c5b7011722e4c1f8ca587baea068a9e6471c8b

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "62c1ff2a536c654ae17342b7d78ac5388014e230cf35d67c6710aec7eb2d0ec2",
    "cross_cats_sorted": [],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.CV",
    "submitted_at": "2026-05-13T15:32:57Z",
    "title_canon_sha256": "831c85f859235355edf55da23da924a86619c6a93fb02c04343a993b1b33dc8a"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.13672",
    "kind": "arxiv",
    "version": 1
  }
}