DOVA: Deliberation-First Multi-Agent Orchestration for Autonomous Research Automation

Aaron Shen, Alfred Shen · 2026

Read DOVA: Deliberation-First Multi-Agent Orchestration for Autonomous Research Automation by Aaron Shen, Alfred Shen (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-13327.txt.

Description

Aaron Shen and Alfred Shen present DOVA, a multi-agent platform for autonomous research tasks. Single agents struggle with work that needs many sources, adversarial checks, and tailored delivery. DOVA puts deliberation first, so meta-reasoning runs before any tool is called. A persistent user model and entity-aware context guide that first step. A three-phase pipeline blends ensemble diversity, blackboard transparency, and iterative refinement. A six-level token-budget scheme cuts inference cost on simple tasks while keeping deep reasoning for hard ones.

Catalog record

CREATORAaron Shen, Alfred Shen
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE28806
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.13327

Browse the full library · About the Tower of Babel