Few-shot Visual Reasoning with Meta-Analogical Contrastive Learning

Authors: Youngsung Kim, Jinwoo Shin, Eunho Yang, Sung Ju Hwang

NeurIPS 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We validate our method on RAVEN dataset, on which it outperforms state-of-the-art method, with larger gains when the training data is scarce.
Researcher Affiliation Collaboration 1Samsung Advanced Institute of Technology, Samsung Electronics, 2Korea Advanced Institute of Science and Technology (KAIST)
Pseudocode No The paper does not contain structured pseudocode or algorithm blocks.
Open Source Code No The paper does not include an explicit statement about releasing source code for the methodology, nor does it provide a link to a code repository.
Open Datasets Yes We validate our method on a publicly available RPM dataset, RAVEN [2]. ... The RAVEN dataset consists of 70,000 problems, equally distributed across seven different tasks (e.g. Center , 2 2 Grid , 3 3 Grid , etc) [2, 3].
Dataset Splits Yes The RAVEN dataset... The dataset is split into ten subsets, where six of them are used as the training set, two are used for validation, and the remaining two are used for test.
Hardware Specification Yes We resize the input images to 80 80 pixels, and train all models on NVIDIA Tesla V100 GPUs using the ADAM optimizer.
Software Dependencies No The paper mentions using the ADAM optimizer but does not specify version numbers for any software dependencies like programming languages, libraries, or frameworks (e.g., Python, PyTorch, TensorFlow).
Experiment Setup Yes We set the number of epoch using the validation set provided in the dataset. ... We resize the input images to 80 80 pixels, and train all models on NVIDIA Tesla V100 GPUs using the ADAM optimizer. ... Batchsize=32. ... Batchsize=2.