Parameterless Transductive Feature Re-representation for Few-Shot Learning
Authors: Wentao Cui, Yuhong Guo
ICML 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We conduct experiments on three benchmark datasets by applying the framework to both representative meta-learning baselines and state-of-the-art FSL methods. Our framework consistently improves performances in all experiments and refreshes the state-of-the-art FSL results. |
| Researcher Affiliation | Academia | 1School of Computer Science, Carleton University, Canada 2Canada CIFAR AI Chair, Amii. Correspondence to: Wentao Cui <wentaocui@cmail.carleton.ca>, Yuhong Guo <yuhong.guo@carleton.ca>. |
| Pseudocode | No | The paper describes the proposed framework using mathematical equations but does not include any pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not provide concrete access information (e.g., a specific repository link or an explicit code release statement) for the source code of the methodology described. |
| Open Datasets | Yes | We conducted experiments on three FSL benchmark datasets: mini-Image Net (Ravi & Larochelle, 2016), tiered-Image Net (Ren et al., 2018) and CUB (Welinder et al., 2010). |
| Dataset Splits | Yes | We follow the train/validation/test split configuration in (Ravi & Larochelle, 2016; Ren et al., 2018; Chen et al., 2019) for the three datasets respectively and report the average test results over multiple runs. |
| Hardware Specification | No | The paper does not provide specific details about the hardware used to run the experiments, such as GPU models or CPU specifications. |
| Software Dependencies | No | The paper does not provide specific software dependencies with version numbers (e.g., Python 3.x, PyTorch 1.x) that would be needed for replication. |
| Experiment Setup | Yes | We find that 5 × 10−5 generally works best for all baseline model training. When the re-representation layer is enabled, we fix the learning rate to 10−5. The key hyperparameters in our proposed framework are α1, α2 and τ. Their values are selected using the validation split of the corresponding datasets. |