Socialized Word Embeddings
Authors: Ziqian Zeng, Yichun Yin, Yangqiu Song, Ming Zhang
IJCAI 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | To demonstrate the effectiveness, we used the latest large-scale Yelp data to train our vectors, and designed several experiments to show how user vectors affect the results. |
| Researcher Affiliation | Academia | Ziqian Zeng1, Yichun Yin2,1, Yangqiu Song1 and Ming Zhang2 1Department of CSE, HKUST, Hong Kong. 2School of EECS, Peking University, China. |
| Pseudocode | Yes | Algorithm 1 Socialized Word Embedding Algorithm. |
| Open Source Code | Yes | The code is available at https://github.com/ HKUST-Know Comp/Socialized Word Embeddings. |
| Open Datasets | Yes | We use the Yelp Challenge1 datasets as our evaluation sets. 1 https://www.yelp.com/dataset challenge |
| Dataset Splits | Yes | We randomly split the data to be 8:1:1 for training, developing, and testing. All the following results are based on this fixed segmentation. |
| Hardware Specification | Yes | All the experiments were conducted on a Super Micro server with E5-2640v4 CPUs. |
| Software Dependencies | No | The word embedding model was written with C language based on the original release of word2vec2. All the experiments were conducted on a Super Micro server with E5-2640v4 CPUs. This describes the language and the base model, but doesn't list specific software dependencies with version numbers (e.g., specific C compiler version, libraries with versions used for the word embedding model beyond 'word2vec'). |
| Experiment Setup | Yes | For example, the window size was set to be five and the dimension of embeddings was set to be 100. We used CBOW model for all the word embeddings. We performed grid search in the range of {2 5, . . . , 25} over both r and λ based on validation set, select the best hyperparameters, and show the final results on test set in Table 2. |