Selecting Representative Examples for Program Synthesis
Authors: Yewen Pu, Zachery Miranda, Armando Solar-Lezama, Leslie Kaelbling
ICML 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We empirically evaluate the representativeness of the subsets constructed by our method, and demonstrate such subsets can significantly improve synthesis time and stability. and 6. Experiments Our approach is evaluated against two criteria: First, the representativeness of our selected subset is explicitly measured; Then, the time/stability improvement of using such a subset is measured against several strong baselines. |
| Researcher Affiliation | Academia | Yewen Pu 1 Zachery Miranda 1 Armando Solar-Lezama 1 Leslie Pack Kaelbling 1 Program synthesis and 1Massachusetts Institute of Technology. Correspondence to: Yewen Pu <yewenpu@mit.edu>. |
| Pseudocode | Yes | Algorithm 1 greedy selection with count oracle c, Algorithm 2 CEGIS, Algorithm 3 Synthesis with Representative Examples |
| Open Source Code | Yes | The supplementary material and the code can be found at https://github.com/evanthebouncy/icml2018_ selecting_representative_examples |
| Open Datasets | No | The paper describes datasets that were generated for the experiments (e.g., "1000 strings of variable length", "250 randomly sampled 32x32 pixel renderings") or abstract problem descriptions ("dataset of pair-wise ordering relations"), but does not provide concrete access information (links, DOIs, formal citations) for publicly available or open datasets used for training. |
| Dataset Splits | No | The paper does not provide specific dataset split information (exact percentages, sample counts, citations to predefined splits, or detailed splitting methodology) for training, validation, or testing. |
| Hardware Specification | No | The paper does not provide specific hardware details (exact GPU/CPU models, processor types, memory amounts, or detailed computer specifications) used for running its experiments. |
| Software Dependencies | No | The paper mentions using "a neural network" and "Z3 (de Moura & Bjørner, 2008)" but does not provide specific version numbers for these or any other software dependencies, which are required for reproducible descriptions. |
| Experiment Setup | No | The paper mentions general network architectures (e.g., "simple feed-forward neural network", "convolutional neural network") and high-level concepts like neighborhood functions, but it lacks specific hyperparameter values (e.g., learning rate, batch size, number of epochs, optimizer settings) or detailed system-level training configurations in the main text. It refers to "supplementary for parameters details", indicating these are not in the main paper. |