All notable changes to this project will be documented in this file
- fix AMF: insert of perturbation evaluation, insert multi-step adversarial perturbations
- fix AMR: insert of perturbation evaluation, insert multi-step adversarial perturbations, improved the data pipeline
- start experiments with command line parameters (
python start_experiments.py --conf name_config_file) - released configuration files for DEMO paper (Under review at RecSys 2021)
- readme file in
/data/directory with links to datasets
- early stopping strategies
- offline recommendation files evaluation (ProxyRecommender, RecommendationFolder)
- negative sampling evaluation
- improved Microsoft Windows compatibility
- binarization of explicit dataset
- automatic loading of implicit datasets
- multiple prefiltering
- managing side information with modular loaders
- alignment of side information with training data
- improved Documentation: Model creation, Side Information loading, Early Stopping, Negative Sampling
- added nDCG as formulated in Rendle's 2020 KDD paper
- visual loader with tensorflow pipeline
- added and fixing visual recsys method:
- DVBPR
- VBPR
- DeepStyle
- ACF
- VNPR
- added new recommender method
- MF (Rendle's 2020 RecSys reproducibility paper)
- EASER
- RP3beta
- iALS
requirements.txtfor Pillow vulnerabilities, change version to >=8.1.1- Adversarial features for ECIR tutorial "AL4Rec"
- Hands-on example for ECIR Tutorial "AL4Rec"
- new dataloader ItemCategoryLoader
- Enabled FunSVD with batch_size
- Generalized feature-aware Factorization Machines
- batch_size field in recommendation snippet examples
- MultiDAE documentation
- setup issues by @deklanw
- indentation in README config by @deklanw
- separated sample main and sample devel scripts
- paper experiment coherence
- combination Full recommendation metrics + argpartition
- Dataset data mock-ups
- long names namespace model generation during hyperspace exploration
- default validation k setup
First Release