Pith Number

pith:M5WLHRDP

pith:2026:M5WLHRDP3GFADTXBMCQMJESRJV

not attested not anchored not stored refs resolved

Echoes in Filter Bubble: Diagnosing and Curing Popularity Bias in Generative Recommenders

Bangguo Zhu, Chengqi Zhang, Hao Chen, Jun Yin, Peng Huo, Ruochen Liu, Senzhang Wang, Shirui Pan

Generative recommenders develop severe popularity bias from token-level optimization flaws and uniform item tokenization, which Ghost corrects using asymmetric unlikelihood optimization and skeleton-founded tokenization.

arxiv:2605.16825 v1 · 2026-05-16 · cs.IR · cs.AI

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

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{M5WLHRDP3GFADTXBMCQMJESRJV}

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

The severe popularity bias emerges from the confluence of a token-level optimization flaw and the undifferentiated property of item tokenization. Ghost substantially alleviates popularity bias and promotes fairer recommendations, while incurring slight degradation to the overall recommendation utility.

C2weakest assumption

The theoretical analyses correctly pinpoint the root causes of popularity bias in generative recommenders, and the proposed asymmetric unlikelihood optimization together with skeleton-founded tokenization effectively mitigate these causes without introducing new biases or requiring extensive post-hoc adjustments.

C3one line summary

Ghost reduces popularity bias in generative recommenders through asymmetric unlikelihood optimization and skeleton-founded tokenization.

References

48 extracted · 48 resolved · 2 Pith anchors

[1] Deep interest network for click-through rate prediction, 2018

[2] Deep neural networks for youtube recommendations, 2016

[3] Deepinf: Social influence prediction with deep learning, 2018

[4] Recommender systems with generative retrieval, 2023

[5] Adapt- ing large language models by integrating collaborative semantics for recommendation, 2024

Receipt and verification

First computed	2026-05-20T00:03:24.627724Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

676cb3c46fd98a01cee160a0c492514d7660a873087c51de92f5e696d2dba7f4

Aliases

arxiv: 2605.16825 · arxiv_version: 2605.16825v1 · doi: 10.48550/arxiv.2605.16825 · pith_short_12: M5WLHRDP3GFA · pith_short_16: M5WLHRDP3GFADTXB · pith_short_8: M5WLHRDP

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/M5WLHRDP3GFADTXBMCQMJESRJV \
  | 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: 676cb3c46fd98a01cee160a0c492514d7660a873087c51de92f5e696d2dba7f4

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "5184a92f5d0609a10357796c50d4d871d39989d238f575b13e056458448765ec",
    "cross_cats_sorted": [
      "cs.AI"
    ],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.IR",
    "submitted_at": "2026-05-16T06:02:40Z",
    "title_canon_sha256": "d46c42a553fbd03f14ece5762727663a84b0b96a3d84253fd847ef0ef9fc3438"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.16825",
    "kind": "arxiv",
    "version": 1
  }
}