Notice: This website is continuously updated as new conference proceedings and journal volumes are published. Consequently, the figures presented here may differ from those reported in our publications.
A brief overview of our methodology is provided on the Methods page, along with references to publications documenting our approach and links to the code and data used in each study.
The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates.
Lie-Access Neural Turing Machines
Authors: Greg Yang, Alexander Rush
ICLR 2017 | Venue PDF | LLM Run Details | Input Tokens: 18,442 Total number of tokens sent to the LLM as input for this paper's analysis. | Output Tokens: 4,744 Total number of tokens produced by the LLM (including reasoning/thinking tokens) for this paper's analysis.
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | To experiment with this approach, we implement a simplified Lie-access neural Turing machine (LANTM) with different Lie groups. We find that this approach is able to perform well on a range of algorithmic tasks. 5 EXPERIMENTS |
| Researcher Affiliation | Academia | Greg Yang and Alexander M. Rush {gyang@college,srush@seas}.harvard.edu Harvard University Cambridge, MA 02138, USA |
| Pseudocode | No | The paper describes procedures and models using text and mathematical equations, but does not include any explicitly labeled "Pseudocode" or "Algorithm" blocks. |
| Open Source Code | Yes | 1Our implementations are available at https://github.com/harvardnlp/lie-access-memory |
| Open Datasets | No | The paper uses custom-designed algorithmic tasks with randomly generated examples, and does not provide a link, DOI, repository, or formal citation to a publicly available or open dataset. |
| Dataset Splits | No | The paper describes how training and test data are generated based on sequence length, but does not explicitly mention a validation set or provide specific train/validation/test dataset splits (e.g., percentages, sample counts, or citations to predefined splits). |
| Hardware Specification | No | The paper does not provide specific hardware details such as GPU/CPU models, processor types, or detailed computer specifications used for running experiments. |
| Software Dependencies | No | The paper mentions 'LSTM' and 'torch' (implied by initialization details), but does not provide specific version numbers for software dependencies like Python, PyTorch, or CUDA. |
| Experiment Setup | Yes | Setup. Our experiments utilize an LSTM controller in a version of the encoder-decoder setup... Model Setup. For all tasks, the LSTM baseline has 1 to 4 layers, each with 256 cells. Each of the other models has a single-layer, 50-cell LSTM controller, with memory width (i.e. the size of each memory vector) 20. Other parameters such as learning rate, decay, and intialization are found through grid search. Further hyperparameter details are give in the appendix. Table A.1: Parameter grid for grid search. |