Metraipolis › archive › dc_6e69839f5eea

ActICL: Post-Training for Agentic Multimodal In-Context Learning — arXiv:2610.19475

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.19475
License (as archived)
CC0-1.0
Archived
2026-10-15T20:36:27.266Z
SHA-256 of the archived text
2033f11be00d85969ac575d0c8f4e4fda3c18e41db17845b69337833dadaaecc
Stable URL of this page
https://metraipolis.com/documento/dc_6e69839f5eea

How to cite

ActICL: Post-Training for Agentic Multimodal In-Context Learning — arXiv:2610.19475 — arXiv (en), https://arxiv.org/abs/2610.19475 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_6e69839f5eea (Metraipolis archive dc_6e69839f5eea, sha256 2033f11be00d85969ac575d0c8f4e4fda3c18e41db17845b69337833dadaaecc)
BibTeX
@misc{dc6e69839f5eea,
  title = {ActICL: Post-Training for Agentic Multimodal In-Context Learning — arXiv:2610.19475},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.19475},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_6e69839f5eea; sha256 2033f11be00d85969ac575d0c8f4e4fda3c18e41db17845b69337833dadaaecc},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.19475. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Multimodal in-context learning (ICL) is an important capability of multimodal large language models (MLLMs). Its performance is closely tied to the demonstrations it uses, making demonstration retrieval a core problem for ICL. Common methods use fixed scoring methods to retrieve all the demonstrations at once, which may introduce redundancy and noise. Context-aware retrieval methods have emerged to address this problem by allowing incremental retrieval from a growing context. However, none of them lets the MLLM learn how to judge what new demonstrations it needs and then carry out progressive retrieval on its own. To this end, we introduce ActICL, which trains an MLLM agent to acquire and use an ICL context via tool-augmented reasoning. An ActICL agent analyzes its own information gaps during iterative reasoning and tool-use actions. Based on these gaps, it performs targeted retrieval to acquire a set of initial candidates. It can then use the existing context to further filter the candidates and refine the context, or it can iteratively perform a new retrieval. When the agent judges that the context is sufficient for the task, or when the resource limit is reached with the user request met, it sends the acquired context to an answering model, either the agent itself or another MLLM. We use a two-stage post-training procedure that jointly optimizes the agent's acquisition and answering. With an adaptive number of demonstrations, ActICL improves average performance by 11.6 percentage points over standard similarity-based retrieval across seven benchmarks and three MLLMs, and by 4.4 points compared to the strongest prior baseline. The contexts acquired by ActICL agents can also be used effectively by frontier models. When GPT-5.6-Sol performs 4-shot ICL, ActICL's contexts improve its performance by 3.5 points over standard retrieval, averaged over three benchmarks.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2610.19475 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-15T19:36:18.291Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).