DNCs Require More Planning Steps

Authors: Yara Shamshoum, Nitzan Hodos, Yuval Sieradzki, Assaf Schuster

ICML 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We evaluate our method on Graph Shortest Path, Convex Hull, Graph Min Cut and Associative Recall, and show how the planning budget can drastically change the behavior of the learned algorithm, in terms of learned time complexity, training time, stability and generalization to inputs larger than those seen during training.
Researcher Affiliation Academia 1Department of Computer Science, Technion Israel Institute of Technology, Haifa, Israel.
Pseudocode No The paper describes algorithms and processes in text and figures, but does not include structured pseudocode or algorithm blocks.
Open Source Code No The paper does not provide an explicit statement or link to open-source code for the described methodology.
Open Datasets No The paper does not provide concrete access information (link, DOI, repository, or formal citation with author/year) for a publicly available or open dataset. It describes how the data was generated but does not offer access.
Dataset Splits No The paper describes a curriculum-based training approach and when the model is evaluated to move to the next lesson, but it does not specify exact training/validation/test split percentages or sample counts for reproduction.
Hardware Specification No The paper discusses computational time and memory, and mentions 'FLOPs' and 'memory size' but does not specify any particular hardware (GPU/CPU models, etc.) used for the experiments.
Software Dependencies No The paper mentions 'Python' but does not provide specific version numbers for Python or any other libraries/frameworks used.
Experiment Setup Yes We train our models with various planning budgets including the baseline of p(n) = 10, larger constant planning budgets and adaptive planning budgets. The specific constants tested depend on the problem, with the maximum being the memory size to guarantee that it can fit in memory during training. When choosing the specific function for the adaptive budgets, we rely on the known problem complexity as our guideline. We thus begin by comparing the constant budgets with a linear one p(n) = n, for Graph Shortest Path. Additionally, we experiment with different coefficients p(n) = kn for the Convex Hull problem, testing values such as k = 0.5, 1.5. All of the models are trained using the same constant memory size and on the same data distribution following a curriculum. We refer to Appendix A for problem descriptions and training details, and Appendix B for the curriculums used for training.