Transferable Contextual Bandit for Cross-Domain Recommendation
Authors: Bo Liu, Ying Wei, Yu Zhang, Zhixian Yan, Qiang Yang
AAAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We perform both theoretical regret analysis and empirical experiments. The empirical results show that TCB outperforms the state-of-the-art algorithms over time. |
| Researcher Affiliation | Collaboration | The Hong Kong University of Science and Technology, Hong Kong Cheetah Mobile USA |
| Pseudocode | Yes | Algorithm 1 Transferable Contextual Bandit |
| Open Source Code | No | The paper states that "h Lin UCB and Factor UCB are both available at github.com/ huazhengwang/Bandit Lib" for baselines, and that "Supplementary material is available at http://www.cse.ust.hk/ bliuab". However, it does not explicitly state that the source code for the proposed TCB method itself is openly available or linked. |
| Open Datasets | Yes | Finally, we verify that TCB can also be applied to the homogeneous problem using the public Delicious data... Dataset is available at grouplens.org/datasets/hetrec-2011 |
| Dataset Splits | No | The paper describes using synthetic and real-world datasets for experiments but does not explicitly provide details about training/validation/test splits, such as percentages, counts, or specific pre-defined splits for all datasets. It mentions cumulative reward for a number of steps, implying a sequential evaluation rather than distinct splits for training, validation, and testing as typically defined. |
| Hardware Specification | No | The paper does not provide any specific details about the hardware used for running the experiments (e.g., GPU/CPU models, memory, or cloud instance types). |
| Software Dependencies | No | The paper does not specify any software dependencies with version numbers (e.g., Python 3.x, specific library versions like PyTorch 1.x or TensorFlow 2.x). |
| Experiment Setup | Yes | For all experiments, we fix TCB s hyperparameter as β = 0.1 and γ = 0.1 such that all terms in Eq. 4 are in the same order of magnitude. To be consistent with the canonical Lin UCB, we set the regularization hyperparameter of all methods to be one and exploration hyperparameter to be 0.2. |