Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].
Label Disentanglement in Partition-based Extreme Multilabel Classification
Authors: Xuanqing Liu, Wei-Cheng Chang, Hsiang-Fu Yu, Cho-Jui Hsieh, Inderjit Dhillon
NeurIPS 2021 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experimental results on synthetic and real datasets show that our method can successfully disentangle multi-modal labels, leading to state-of-the-art (SOTA) results on four XMC benchmarks. |
| Researcher Affiliation | Collaboration | Xuanqing Liu Department of Computer Science UCLA EMAIL Wei-Cheng Chang Amazon Inc. EMAIL Hsiang-Fu Yu Amazon Inc. EMAIL Cho-Jui Hsieh Department of Computer Science UCLA EMAIL Inderjit Dhillon Amazon Inc. & UT Austin EMAIL |
| Pseudocode | Yes | Algorithm 1 Our proposed framework. |
| Open Source Code | Yes | Code will be released at https://github.com/amzn/pecos. |
| Open Datasets | Yes | We consider four publicly available XMC benchmark datasets [2, 10] for our experiments. See Table 1 for data statistics...These four datasets and the sparse TF-IDF features are downloaded from https://github.com/yourh/Attention XML which are the same as used in Attention XML [10] and X-Transformer [12]. |
| Dataset Splits | No | Table 1 specifies 'ntrn' (number of instances in training splits) and 'ntst' (number of instances in testing splits) but does not provide details on a validation split, nor does it specify exact percentages for training, validation, and test splits. |
| Hardware Specification | No | The paper mentions 'Extra training time and inference time' and 'negligible overhead' but does not specify any particular hardware components like CPU models, GPU models, or memory specifications used for the experiments. |
| Software Dependencies | No | The paper states: 'Our implementation is based on the PECOS software library.' However, it does not provide specific version numbers for PECOS or any other software dependencies. |
| Experiment Setup | Yes | Hyper-parameters. Our technique depends on hyperparameter λ, which is tested in a standalone section. For hyperparameters in the XMC model, we mostly follow the default settings in the corresponding software. The details about hyperparameters are listed in the Appendix A.6. |