Recurrent Convolutional Neural Networks for Text Classification

Authors: Siwei Lai, Liheng Xu, Kang Liu, Jun Zhao

AAAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We conduct experiments on four commonly used datasets. The experimental results show that the proposed method outperforms the state-of-the-art methods on several datasets, particularly on document-level datasets.
Researcher Affiliation Academia Siwei Lai, Liheng Xu, Kang Liu, Jun Zhao National Laboratory of Pattern Recognition (NLPR) Institute of Automation, Chinese Academy of Sciences, China {swlai, lhxu, kliu, jzhao}@nlpr.ia.ac.cn
Pseudocode No The paper includes a network structure diagram (Figure 1), but no pseudocode or algorithm blocks.
Open Source Code No The paper does not contain any statement or link indicating that the source code for the described methodology is publicly available.
Open Datasets Yes We perform the experiments using the following four datasets: 20Newsgroups, Fudan Set, ACL Anthology Network, and Sentiment Treebank. The paper provides URLs for each dataset: 'qwone.com/ jason/20Newsgroups/', 'www.datatang.com/data/44139 and 43543', 'old-site.clsp.jhu.edu/ sbergsma/Stylo/', 'nlp.stanford.edu/sentiment/'.
Dataset Splits Yes Table 1 provides detailed information about each dataset, including Train/Dev/Test set entries (e.g., 20News: 7520/836/5563, Fudan: 8823/981/9832, ACL: 146257/28565/28157, SST: 8544/1101/2210). Additionally, it states: 'The ACL and SST datasets have a pre-defined training, development and testing separation. For the other two datasets, we split 10% of the training set into a development set and keep the remaining 90% as the real training set.'
Hardware Specification No The paper does not provide any specific details about the hardware used for running the experiments.
Software Dependencies No The paper mentions software like 'Stanford Tokenizer', 'ICTCLAS', and 'word2vec' but does not specify their version numbers, which is required for reproducibility.
Experiment Setup Yes We set the learning rate of the stochastic gradient descent α as 0.01, the hidden layer size as H = 100, the vector size of the word embedding as |e| = 50 and the size of the context vector as |c| = 50.