Awesome Auto Research Hub Papers · Datasets · Projects
← Back to papers

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

arXiv 2026 52.5 method

TLDR

DOVA is a multi-agent platform for autonomous research automation using deliberation-first orchestration, hybrid reasoning, and adaptive thinking.

Reasoning

The paper introduces a novel multi-agent architecture with meta-reasoning and adaptive token allocation, supported by an ablation study. However, it lacks explicit real-world evaluation or benchmark results, limiting empirical validation.

Read-first score

Read-first score 52.5, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 64.

Recency 8%
100

Uses a gentle age decay so recent papers surface without erasing older foundations. 2026

Topical relevance 42%
53.3

Uses existing LLM keyword relevance scores normalized to 0-100. AI scientist,automated scientific discovery,autonomous research agent,automated research,literature review agent,survey generation,automated experimentation,experiment design agent,AI for scientific research,paper writing agent,research automation,scientific discovery agent

Methodology quality 25%
50

Screens visible abstract and analysis fields for experiment, dataset, baseline, metric, and limitation evidence. markers=ablation

Reproducibility 25%
38

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=False; dataset=False; markers=code

Field roles

Frontier

Rank sensitivity

Stability: volatile; rank range: 75.

Keyword Scores

automated research
10
research automation
10
autonomous research agent
9
AI for scientific research
7
automated scientific discovery
6
scientific discovery agent
6
literature review agent
4
AI scientist
3
automated experimentation
3
survey generation
2
experiment design agent
2
paper writing agent
2

Deep Analysis

Innovations

  • Deliberation-first orchestration with explicit meta-reasoning before tool invocation, informed by a persistent user model and entity-aware conversation context
  • Hybrid collaborative reasoning: a composable three-phase pipeline unifying ensemble diversity, blackboard transparency, and iterative refinement
  • Adaptive multi-tiered thinking: a six-level token-budget allocation scheme that reduces inference cost by 40-60% on simple tasks while preserving deep reasoning capacity

Methodology

The paper presents DOVA, a multi-agent platform that formalizes core algorithms for deliberation-first orchestration, hybrid collaborative reasoning, and adaptive multi-tiered thinking. An architectural ablation study across seven system configurations evaluates the contribution of each component to answer confidence, source coverage, and token efficiency.

Key Results

Adaptive multi-tiered thinking reduces inference cost by 40-60% on simple tasks while preserving deep reasoning capacity; the ablation study quantifies each component's impact on answer confidence, source coverage, and token efficiency.

Tags

AI