Spherical Text Embedding
Authors: Yu Meng, Jiaxin Huang, Guangyuan Wang, Chao Zhang, Honglei Zhuang, Lance Kaplan, Jiawei Han
NeurIPS 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In this section, we empirically evaluate the quality of spherical text embeddings for three common text embedding application tasks, i.e., word similarity, document clustering and document classification. The Spearman s rank correlation is reported in Table 1, which reflects the consistency between word similarity rankings given by cosine similarity of word embeddings and human raters. The results are reported in Table 2, with mean and standard deviation computed over 10 runs. |
| Researcher Affiliation | Collaboration | 1 Department of Computer Science, University of Illinois at Urbana-Champaign 2 College of Computing, Georgia Institute of Technology 3 U.S. Army Research Laboratory |
| Pseudocode | No | The paper does not contain structured pseudocode or algorithm blocks (clearly labeled algorithm sections or code-like formatted procedures). |
| Open Source Code | No | The paper does not include an unambiguous statement that the authors are releasing the source code for the work described in this paper, nor does it provide a direct link to a code repository. |
| Open Datasets | Yes | The training corpus for word similarity is the latest Wikipedia dump3 containing 2.4 billion tokens. Words appearing less than 100 times are discarded, leaving 239, 672 unique tokens. 3https://dumps.wikimedia.org/enwiki/latest/enwiki-latest-pages-articles.xml.bz2. The training corpus is the 20 Newsgroups dataset4, and we treat each document as a paragraph in all compared models. 4http://qwone.com/~jason/20Newsgroups/. Besides the 20 Newsgroup dataset used in Section 5.2 which is a topic classification dataset, we evaluate different document/paragraph embedding methods also on a binary sentiment classification dataset consisting of 1, 000 positive and 1, 000 negative movie reviews5. 5http://www.cs.cornell.edu/people/pabo/movie-review-data/ |
| Dataset Splits | No | For the 20 Newsgroup dataset, we follow the original train/test sets split; for the movie review dataset, we randomly select 80% of the data as training and 20% as testing. (No explicit mention of validation split details) |
| Hardware Specification | Yes | All the models except BERT are run on a machine with 20 cores of Intel(R) Xeon(R) CPU E5-2680 v2 @ 2.80 GHz; BERT is trained on 8 NVIDIA Ge Force GTX 1080 GPUs. |
| Software Dependencies | No | The paper does not list specific version numbers for key software components or libraries used in the experiments. |
| Experiment Setup | Yes | The models are trained for 10 iterations on the corpus; the local context window size is 10; the embedding dimension is 100. In our Jo SE model, we set the margin in Equation (3) to be 0.15, the number of negative samples to be 2, the initial learning rate to be 0.04 with linear decay. |