Coracle is an R package for a conformalized framework for multimodal AI with continuous outcomes.

It has several advantages:

• adapts to early, late, and intermediate fusion.
• provides theoretical marginal confidence guarantees and achieves valid finite-sample coverage without relying on distributional assumptions.

You can install the development version directly from GitHub:

install.packages("devtools")
devtools::install_github("himelmallick/Coracle")
library(Coracle)

Here is a minimal call to Coracle():

cml <- Coracle::Conformal(
  fit = NULL,           # A fitted \emph{IntegratedLearner} object from early/late fusion
  fit_coop = NULL,      # A fitted \emph{BayesCOOP} object for intermediate fusion
  data_calib,           # list with feature_table, sample_metadata and feature_metadata (calibration set)
  data_valid,           # list with feature_table, sample_metadata and feature_metadata (validation set)
  fusion_choice = c("late", "early", "intermediate"), # choice of different fusion schemes
  conf_level = 0.95     # desired marginal confidence level
)

A list with components:

• df: A dataframe containing the Coracle output for different fusion choices
• coverage: A list containing the coverage for different fusion choices
• conf_level: Desired marginal confidence level

For an in-depth workflow including real data preprocessing, baseline comparisons, and performance benchmarking, please see the full tutorial:

📄 View the Coracle Tutorial

We are happy to troubleshoot any issues with the package. Please contact the authors via email or open an issue in the GitHub repository.