Dimensions overall score 7.0

• The PROClaim framework achieves 81.7% accuracy on the Check-COVID benchmark, outperforming standard multi-agent debate by 10.0 percentage points.result 95%
• The Progressive RAG (P-RAG) component is the primary driver of performance gains, contributing +7.5 percentage points to accuracy.result 90%
• The framework uses a courtroom-style structure with specialized roles (Plaintiff, Defense, Judge, Critic, Expert Witness) for adversarial deliberation.method 90%
• Progressive RAG (P-RAG) performs iterative, query-adaptive retrieval to continuously expand and refine the evidence pool during the debate.method 90%
• The framework employs evidence negotiation, self-reflection, and heterogeneous multi-judge aggregation to enforce calibration, robustness, and diversity.method 85%
• Structural deliberation and model heterogeneity effectively mitigate systematic biases in claim verification.result 85%
• Standard multi-agent debate (MAD) is limited by unstructured debate dynamics, leading to premature convergence and shared bias reinforcement.limitation 80%
• The framework uses an evidence novelty threshold of 0.20 for admitting new evidence, selected as a redundancy-noise trade-off heuristic.technical 90%

Competitive landscape

Competitor map is still being generated for this paper. Enable generation or check back soon.

Startup potential card