Pith Number

pith:6NIWHZOO

pith:2026:6NIWHZOOM67FZJ3SCGCKM7LKM4

not attested not anchored not stored refs resolved

Angel or Demon: Investigating the Plasticity Interventions' Impact on Backdoor Threats in Deep Reinforcement Learning

Chunyi Zhou, Jiahao Chen, Linkang Du, Oubo Ma, Ruixiao Lin, Shouling Ji, Yang Dai

Most plasticity interventions reduce backdoor threats in deep reinforcement learning, but SAM makes them worse by amplifying gradients.

arxiv:2605.14587 v1 · 2026-05-14 · cs.LG · cs.AI · cs.CR

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

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{6NIWHZOOM67FZJ3SCGCKM7LKM4}

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

only one intervention (i.e., SAM) exacerbates backdoor threats, while other interventions mitigate them. Pathological analysis identifies that the exacerbation is attributed to backdoor gradient amplification, while the mitigation stems from activation pathway disruption and representation space compression.

C2weakest assumption

The 14,664 tested cases sufficiently represent the space of practical DRL deployments and attack scenarios so that the observed patterns generalize beyond the chosen environments and models.

C3one line summary

Most plasticity interventions in DRL reduce backdoor attack success rates while SAM increases them via gradient amplification; the work introduces an SCC framework and loss-sharpness detection indicator.

References

75 extracted · 75 resolved · 0 Pith anchors

[1] Continuous control with deep reinforcement learning , author=. ICLR , year=

[2] Multi-agent actor-critic for mixed cooperative-competitive environments , author=. NeurIPS , year=

[3] CLIBE: detecting dynamic backdoors in transformer-based NLP models , author=. NDSS , year=

[4] IEEE transactions on neural networks and learning systems , year=

[5] Understanding black-box predictions via influence functions , author=. ICML , year=

Formal links

2 machine-checked theorem links

Receipt and verification

First computed	2026-05-17T23:39:05.291373Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

f35163e5ce67be5ca7721184a67d6a67050855ee3daa3715030ad55375c26cf2

Aliases

arxiv: 2605.14587 · arxiv_version: 2605.14587v1 · doi: 10.48550/arxiv.2605.14587 · pith_short_12: 6NIWHZOOM67F · pith_short_16: 6NIWHZOOM67FZJ3S · pith_short_8: 6NIWHZOO

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/6NIWHZOOM67FZJ3SCGCKM7LKM4 \
  | 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: f35163e5ce67be5ca7721184a67d6a67050855ee3daa3715030ad55375c26cf2

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "c1a66aaeed12fe61c691610188d7728b89f113ee2e29c05e8f1b7dc175dde62e",
    "cross_cats_sorted": [
      "cs.AI",
      "cs.CR"
    ],
    "license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
    "primary_cat": "cs.LG",
    "submitted_at": "2026-05-14T08:58:24Z",
    "title_canon_sha256": "8fc5c0bd06529dc31057da8642c41e4b715bcb901e7e6b93b57560bab1a2208b"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.14587",
    "kind": "arxiv",
    "version": 1
  }
}