FAQ Question

Semantic Scholar uses a state-of-the-art paper embedding model trained using contrastive learning to find papers similar to those in each Library folder.

The papers you save within a Library folder are positive signals, and the more papers you save to a Library folder and rate as “not relevant” in your Research Feed, the better your recommended papers will be.

With these signals, our model is able to fine-tune your paper recommendations and suggest more relevant research over time.

Tip: To ensure you receive the best recommendations, try rating 5 relevant papers by adding them to a Library folder, and 3 non-relevant papers by marking them as "not relevant" in the corresponding Research Feed.