A semi-supervised model of peer review is introduced that is intended to overcome the bias and incompleteness of traditional peer review. Traditional approaches are reliant on human biases, while consensus decision-making is constrained by sparse information. Here, the architecture for one potential improvement (a semi-supervised, human-assisted classifier) to the traditional approach will be introduced and evaluated. To evaluate the potential advantages of such a system, hypothetical receiver operating characteristic (ROC) curves for both approaches will be assessed. This will provide more specific indications of how automation would be beneficial in the manuscript evaluation process. In conclusion, the implications for such a system on measurements of scientific impact and improving the quality of open submission repositories will be discussed.