Improving the Universality and Learnability of Neural Programmer-Interpreters with Combinator Abstraction
Authors: Da Xiao, Jo-Yu Liao, Xingyuan Yuan
ICLR 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | While the previous section analyzes the universality of CNPI, this section shows results on the empirical evaluation of its learnability via both SL and RL experiments. |
| Researcher Affiliation | Collaboration | Da Xiao1,2, Jo-Yu Liao2, Xingyuan Yuan2 1School of Cyberspace Security, Beijing University of Posts and Telecommunications, China 2Colorful Clouds Technology Co., Ltd, Beijing, China |
| Pseudocode | Yes | Figure 2: Pseudo-code for the set of combinators. and Algorithm 1 Combinatory neural programming inference and Algorithm 2 Convert a recursive program set to a combinatory one. |
| Open Source Code | No | The paper does not provide any explicit statement or link for open-source code. |
| Open Datasets | No | The paper mentions using 'synthetic abstract traces' and refers to tasks like 'grade school addition and bubble sort' but does not provide concrete access information (link, DOI, formal citation) for any publicly available or open dataset. |
| Dataset Splits | No | The paper does not provide specific dataset split information (exact percentages, sample counts, or citations to predefined splits) needed to reproduce the data partitioning. It mentions accuracy improvement, implying validation, but without explicit split details. |
| Hardware Specification | No | The paper does not provide specific hardware details (exact GPU/CPU models, processor types, or memory amounts) used for running its experiments. |
| Software Dependencies | No | The paper does not provide specific ancillary software details (e.g., library or solver names with version numbers) needed to replicate the experiment. |
| Experiment Setup | Yes | We trained the CNPI using plain SGD with batch size 1, and learning rate of 0.5 and 0.1 for SL and RL experiments respectively. For the SL experiments, the learning rate was decayed by a factor of 0.1 if prediction accuracy did not improve for 10 epochs. and MAX NSTEP = K n, where K is the number of callable arguments for combinators, n is the complexity of the problem to be solved. |