Probing Brain Activation Patterns by Dissociating Semantics and Syntax in Sentences
Authors: Shaonan Wang, Jiajun Zhang, Nan Lin, Chengqing Zong9201-9208
AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments Experimental setup We randomly sample 500,000 paraphrase pairs from Para NMT-50M (Wieting and Gimpel 2018) as our training set. ... Table 1: Semantic and syntactic evaluation results. ... Figures 5 and 6 show the averaging results over five subjects on two f MRI experiments respectively. |
| Researcher Affiliation | Academia | 1National Laboratory of Pattern Recognition, Institute of Automation, CAS 2School of Artificial Intelligence, University of Chinese Academy of Sciences 3CAS Key Laboratory of Behavioural Science, Institute of Psychology 4Department of Psychology, University of Chinese Academy of Sciences 5CAS Center for Excellence in Brain Science and Intelligence Technology |
| Pseudocode | No | The paper describes the model architecture and methods using diagrams and textual descriptions, but does not include formal pseudocode or algorithm blocks. |
| Open Source Code | No | The paper mentions the code for a baseline model (VGVAE) is available: 'For a fair comparison, we use the same training dataset to retrain the VGVAE model in which data and code are from https://github.com/mingdachen/disentangle-semantics-syntax.' However, it does not state that the code for their proposed DFRM method is open-source or provide a link to it. |
| Open Datasets | Yes | We randomly sample 500,000 paraphrase pairs from Para NMT-50M (Wieting and Gimpel 2018) as our training set. For the STS task, we the STS benchmark dataset and a dataset containing the concatenation of STS tasks from 2012 to 2016 which are from http://ixa2.si.ehu.es/stswiki/index.php/STSbenchmark. Brain activation data Our experiments are conducted on the dataset from Pereira et al. (2018) which is publicly available at https://osf.io/crwz7/. |
| Dataset Splits | No | The paper mentions a 'training set' and 'test set' for evaluating the model, but it does not explicitly specify the use of a 'validation set' or the proportions of the dataset splits (e.g., 80/10/10 split or specific sample counts for each split). |
| Hardware Specification | No | The paper does not provide specific details regarding the hardware (e.g., GPU models, CPU types, or memory) used to train or run their disentangled feature representation model (DFRM). |
| Software Dependencies | No | The paper states 'Models are implemented with Pytorch' but does not specify the version number of PyTorch or any other software dependencies required to replicate the experiments. |
| Experiment Setup | Yes | Models are implemented with Pytorch and parameters are trained for 20 epochs, with each epoch consisting of multiple batches optimized with Adam. Same with the baseline VGVAE model, the dimensions of word embeddings, MLP, and LSTM hidden layers of DFRM are all set to 100. |