Spatially Invariant Unsupervised Object Detection with Convolutional Neural Networks
Authors: Eric Crawford, Joelle Pineau3412-3420
AAAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In this section we empirically demonstrate the advantages of SPAIR on a number of different tasks. Through a series of experiments, we demonstrate a number of features of our architecture: that, unlike AIR, it is able to discover and detect objects in large, many-object scenes; that it has a significant ability to generalize to images that are larger and contain more objects than images encountered during training; and that it is able to discover and detect objects with enough accuracy to facilitate non-trivial downstream processing. |
| Researcher Affiliation | Collaboration | Eric Crawford Mila, Mc Gill University Montreal, QC Joelle Pineau Facebook AI Research, Mila, Mc Gill University Montreal, QC |
| Pseudocode | No | The paper describes the model components and training process textually and mathematically, but does not include any explicit pseudocode or algorithm blocks. |
| Open Source Code | Yes | Further details are provided in the Supplementary Material, and source code can be found at github.com/e2crawfo/auto yolo. |
| Open Datasets | Yes | We trained both AIR and SPAIR on 48 48 images each containing scattered MNIST digits of size 14 14 rendered in white on a black background. |
| Dataset Splits | Yes | Digit instances (i.e. particular MNIST images) are never shared between training, validation and testing datasets. The training dataset contained 128,000 examples, while validation and test datasets contained 500 examples each. |
| Hardware Specification | No | The paper does not provide specific details about the hardware used for experiments, such as GPU or CPU models, memory specifications, or cloud computing instance types. |
| Software Dependencies | No | The paper does not specify particular software dependencies or their version numbers, such as programming languages, libraries, or frameworks with explicit version information. |
| Experiment Setup | No | The paper mentions that 'Further details are provided in the Supplementary Material' regarding the experimental setup, but it does not include specific hyperparameter values (e.g., learning rate, batch size, number of epochs, optimizer settings) or training configurations within the main body of the paper. |