Cornelia Caragea’s scientific contributions

... Mean average precision (mAP) is a gold standard metric in information retrieval and evaluation systems [44]. This metric provides a robust way to evaluate models that produce ranked lists of results, so we use it to evaluate our scene graph generation models. ...

Reference:

Enabling Perspective-Aware Ai with Contextual Scene Graph Generation

... Furthermore, we also created calibration plots to evaluate the observed and predicted risk by the selected model. 29 In these calibration plots, the predicted probabilities of death by AI models were plotted against the observed probabilities of dying in bivariate (x-y) plots. ...

Reference:

APASL-ACLF Research Consortium-Artificial Intelligence (AARC-AI) model precisely predicts outcomes in acute-on-chronic liver failure patients

... Number of correct predictions in top k Total number of predictions (5) (3) Mean Reciprocal Rank (MRR) [7] is a statistical measure commonly used in information retrieval and questionanswering systems to assess how effectively a system ranks the correct answers. In the context of LLMs, MRR evaluates the quality of responses in classification tasks by measuring how high the correct answer is positioned in a ranked list of possible answers based on their probabilities. ...

Reference:

VulnLLMEval: A Framework for Evaluating Large Language Models in Software Vulnerability Detection and Patching