Learning Perceptual Inference by Contrasting
Authors: Chi Zhang, Baoxiong Jia, Feng Gao, Yixin Zhu, HongJing Lu, Song-Chun Zhu
NeurIPS 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We conduct comprehensive experiments on two major datasets: the RAVEN dataset [12] and the PGM dataset [14]. Empirical studies show that our model achieves human-level performance on RAVEN and a new record on PGM, setting new state-of-the-art for permutation-invariant models on the two datasets. |
| Researcher Affiliation | Academia | 1 Department of Computer Science, University of California, Los Angeles 2 Department of Psychology, University of California, Los Angeles 3 Department of Statistics, University of California, Los Angeles 4 International Center for AI and Robot Autonomy (CARA) |
| Pseudocode | No | The paper describes the architecture and mathematical formulations but does not include any explicit pseudocode or algorithm blocks. |
| Open Source Code | No | The paper mentions using a public implementation for the WReN model (a baseline) and states they implemented their models in PyTorch, but there is no explicit statement or link indicating that the source code for their own methodology (CoPINet) is publicly available. |
| Open Datasets | Yes | We verify the effectiveness of our models on two major RPM datasets: RAVEN [12] and PGM [14]. |
| Dataset Splits | Yes | Across all experiments, we train models on the training set, tune hyper-parameters on the validation set, and report the final results on the test set. [...] There are 70, 000 problems in the RAVEN dataset [12], equally distributed in 7 figure configurations. In each configuration, the dataset is randomly split into 6 folds for training, 2 folds for validation, and 2 folds for testing. [...] This split of the dataset has in total 1.42 million samples, with 1.2 million for training, 2, 000 for validation, and 200, 000 for testing. |
| Hardware Specification | Yes | Models are trained on servers with four Nvidia RTX Titans. |
| Software Dependencies | No | The paper states models are implemented in PyTorch [74] and optimized using ADAM [75], but specific version numbers for these software components are not provided. |
| Experiment Setup | No | The paper states that hyperparameters were tuned on the validation set and that early stopping based on validation loss was used, but it does not specify concrete values for hyperparameters such as learning rate, batch size, or number of epochs. |