Skip to content
AI News HubLIVE
Original source2 min read

RoCo-ACE: Rollout-Conditioned Online Distillation for Retention-Aware Knowledge Injection

Summary

The paper introduces RoCo-ACE, a rollout-conditioned online distillation objective for knowledge injection in multimodal large language models. It uses same-rollout reference-free/reference-conditioned likelihood contrast to reallocate distillation weight to reference-supported rollout tokens, and adds sparse reference-side anchored correction for authoritative facts omitted from the rollout. Across three knowledge-injection settings, six retention benchmarks, multiple baselines, and multiple base models, RoCo-ACE achieves the best injected-knowledge accuracy while keeping retention close to the base model.

SourcearXiv AIAuthor: Yan Hong, Wei Li, Kedong Xiu, Jun Lan, Shuheng Zhou, Zhongcai Lyu, Huijia Zhu, Weiqiang Wang, Jianfu Zhang
RoCo-ACE: Rollout-Conditioned Online Distillation for Retention-Aware Knowledge Injection
Report an error

The correction channel is not available yet. You can copy the article reference below for later.

Correction instructions
Read article

[Submitted on 10 Jun 2026]

Title:RoCo-ACE: Rollout-Conditioned Online Distillation for Retention-Aware Knowledge Injection

View a PDF of the paper titled RoCo-ACE: Rollout-Conditioned Online Distillation for Retention-Aware Knowledge Injection, by Yan Hong and 8 other authors

View PDF HTML (experimental)

Abstract:Knowledge injection updates pretrained MLLMs with new factual or domain-specific knowledge, but fitting full authoritative answers can cause drift in non-updated behavior. Online distillation mitigates this drift by training on model-generated rollouts, yet uniform reference-conditioned distillation provides coarse supervision: it can under-emphasize reference-supported rollout tokens and supervise omitted facts only indirectly. We introduce RoCo-ACE, a rollout-conditioned online distillation objective for knowledge injection. RoCo uses same-rollout reference-free/reference-conditioned likelihood contrast to reallocate additional distillation weight to reference-supported rollout tokens, while ACE adds sparse reference-side anchored correction for authoritative anchors omitted from the rollout without full-answer imitation. Across three knowledge-injection settings, six retention benchmarks, multiple baselines, and multiple base models, RoCo-ACE achieves the best injected-knowledge accuracy among compared methods while keeping evaluated retention close to the base model.

Subjects:

Artificial Intelligence (cs.AI); Machine Learning (cs.LG)

Cite as: arXiv:2607.24771 [cs.AI]

(or arXiv:2607.24771v1 [cs.AI] for this version)

https://doi.org/10.48550/arXiv.2607.24771

arXiv-issued DOI via DataCite

Submission history

From: Yan Hong [view email] [v1] Wed, 10 Jun 2026 13:33:10 UTC (7,107 KB)

Full-text links:

Access Paper:

View a PDF of the paper titled RoCo-ACE: Rollout-Conditioned Online Distillation for Retention-Aware Knowledge Injection, by Yan Hong and 8 other authors

View PDF

HTML (experimental)

TeX Source

view license

Current browse context:

cs.AI

new | recent | 2026-07

Change to browse by:

cs cs.LG

References & Citations

NASA ADS

Google Scholar

Semantic Scholar

Loading...

Data provided by:

Bibliographic Tools

Bibliographic and Citation Tools

Bibliographic Explorer Toggle

Bibliographic Explorer (What is the Explorer?)

Connected Papers Toggle

Connected Papers (What is Connected Papers?)

Litmaps Toggle

Litmaps (What is Litmaps?)

scite.ai Toggle

scite Smart Citations (What are Smart Citations?)

Code, Data, Media

Code, Data and Media Associated with this Article

alphaXiv Toggle

alphaXiv (What is alphaXiv?)

Links to Code Toggle

CatalyzeX Code Finder for Papers (What is CatalyzeX?)

DagsHub Toggle

DagsHub (What is DagsHub?)

GotitPub Toggle

Gotit.pub (What is GotitPub?)

Huggingface Toggle

Hugging Face (What is Huggingface?)

ScienceCast Toggle

ScienceCast (What is ScienceCast?)

Demos

Demos

Replicate Toggle

Replicate (What is Replicate?)

Spaces Toggle

Hugging Face Spaces (What is Spaces?)

Spaces Toggle

TXYZ.AI (What is TXYZ.AI?)

Related Papers

Recommenders and Search Tools

Link to Influence Flower

Influence Flower (What are Influence Flowers?)

Core recommender toggle

CORE Recommender (What is CORE?)

Author

Venue

Institution

Topic

About arXivLabs

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)

Key points and analysis

Article intelligence

EngineersAdvanced

Key points

  • RoCo-ACE reallocates distillation weight to reference-supported rollout tokens via same-rollout reference-free/reference-conditioned likelihood contrast.
  • The ACE component applies sparse reference-side anchored correction for authoritative anchors omitted from the rollout without full-answer imitation.
  • It achieves the best injected-knowledge accuracy among compared methods while maintaining retention near the base model across six benchmarks.

Highlights and analysis are generated automatically and may contain errors. Check the original source.