Pith Number

pith:UKKBYDNJ

pith:2026:UKKBYDNJF5SZAW3DCGCYIEGUJT

not attested not anchored not stored refs resolved

ICED: Concept-level Machine Unlearning via Interpretable Concept Decomposition

Jing Lin, Junhao Dong, Li Xu, Piotr Koniusz, Shen Lin

Vision-language models can unlearn specific concepts by decomposing images into sparse semantic combinations and suppressing only the targets.

arxiv:2605.14309 v1 · 2026-05-14 · cs.CV · cs.AI · cs.LG

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

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{UKKBYDNJF5SZAW3DCGCYIEGUJT}

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

Based on this decomposition, our method formulates unlearning as concept-level optimization, where target concepts are selectively suppressed while intra-instance non-target semantics and global cross-modal knowledge are preserved. Extensive experiments across both in-domain and out-of-domain forgetting settings demonstrate that our method enables more comprehensive target forgetting, better preserves non-target knowledge within the same image, and maintains competitive model utility compared with existing VLM unlearning methods.

C2weakest assumption

Visual representations can be decomposed into sparse, nonnegative combinations of semantic concepts from a compact task-specific vocabulary, providing an explicit interface for fine-grained knowledge manipulation without affecting unrelated semantics.

C3one line summary

ICED decomposes visual features into interpretable concepts to enable selective unlearning of target knowledge in VLMs while preserving non-target semantics and model utility.

References

36 extracted · 36 resolved · 0 Pith anchors

[1] Learning transferable visual models from natural language supervi- sion, 2021

[2] Trustworthy ai: From principles to practices, 2023

[3] Allies teach better than enemies: Inverse adversaries for robust knowledge distillation, 2026

[4] Tug-of-war no more: Harmonizing accuracy and robustness in vision-language models via stability-aware task vector merging, 2026

[5] Can bad teaching induce forgetting? unlearning in deep networks using an incompetent teacher, 2023

Receipt and verification

First computed	2026-05-17T23:39:10.014658Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

a2941c0da92f65905b6311858410d44cc8efdce34fef30be70d3bcac65a4ed5c

Aliases

arxiv: 2605.14309 · arxiv_version: 2605.14309v1 · doi: 10.48550/arxiv.2605.14309 · pith_short_12: UKKBYDNJF5SZ · pith_short_16: UKKBYDNJF5SZAW3D · pith_short_8: UKKBYDNJ

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/UKKBYDNJF5SZAW3DCGCYIEGUJT \
  | 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: a2941c0da92f65905b6311858410d44cc8efdce34fef30be70d3bcac65a4ed5c

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "99c9dc26a4ca38913bd9b44aebe2d10028cfc4139aff60b8fa3bda69419b7327",
    "cross_cats_sorted": [
      "cs.AI",
      "cs.LG"
    ],
    "license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
    "primary_cat": "cs.CV",
    "submitted_at": "2026-05-14T03:22:12Z",
    "title_canon_sha256": "4d0b7cc78c52ec7a3ca78468448974d19ad064808de7e286312404d4b3a69c8a"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.14309",
    "kind": "arxiv",
    "version": 1
  }
}