Description Based Text Classification with Reinforcement Learning
Authors: Duo Chai, Wei Wu, Qinghong Han, Fei Wu, Jiwei Li
ICML 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | 5. Experiments 5.1. Benchmarks 5.2. Baselines 5.3. Results and Discussion 6. Ablation Studies and Analysis Table 2. Test error rates on the AGNews, 20news, DBPedia, Yahoo, Yelp P and IMDB datasets for single-label classification. |
| Researcher Affiliation | Collaboration | 1Shannon.AI 2Zhejiang University. |
| Pseudocode | No | No pseudocode or algorithm blocks were found in the paper. |
| Open Source Code | No | The paper does not explicitly state that source code for the methodology is released or provide a link to a code repository. |
| Open Datasets | Yes | AGNews: Topic classification over four categories of Internet news articles (Del Corso et al., 2005). 20newsgroups2: The 20 Newsgroups data set is a collection of documents over 20 different newsgroups. 2http://qwone.com/ jason/20Newsgroups/ Reuters3: A multi-label benchmark dataset for document classification with 90 classes. 3https://martin-thoma.com/nlp-reuters/ |
| Dataset Splits | No | Hyper-parameters for baselines are tuned on the development sets to enforce apple-to-apple comparison. This mentions development sets but does not specify their size, composition, or how they are split from the main dataset (e.g., train/val/test percentages). |
| Hardware Specification | No | The paper does not provide specific details about the hardware (e.g., CPU, GPU models, memory) used to run the experiments. |
| Software Dependencies | No | The paper mentions using BERT as the backbone model but does not specify version numbers for any software dependencies, libraries, or frameworks used for implementation. |
| Experiment Setup | No | The paper states that 'Hyper-parameters for baselines are tuned on the development sets' but does not explicitly list the specific hyperparameter values (e.g., learning rate, batch size, epochs) or other detailed training settings for the models. |