New paper · SkillAudit: From Fixed-Suite Benchmarking to Skill-Centered Assessment

Guider

Type
Open source
Year
2025
Status
active

The official implementation of a guided-calibration framework for denoising multimodal recommender systems.

Guider addresses noisy visual and textual signals in multimodal recommendation. Its guided-calibration design can be attached to different recommender backbones rather than requiring a single fixed architecture.

The repository contains the PyTorch implementation used for the WSDM 2025 paper, along with training configuration and instructions for reproducing the experiments.