Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].
Hybrid Active Learning with Uncertainty-Weighted Embeddings
Authors: Yinan He, Lile Cai, Jingyi Liao, Chuan-Sheng Foo
TMLR 2024 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We extensively evaluate the proposed hybrid active learning method on image classification, semantic segmentation and object detection tasks, and demonstrate that it achieves state-of-the-art performance. |
| Researcher Affiliation | Academia | Yinan He EMAIL Nanyang Technological University, Singapore; Lile Cai EMAIL Institute for Infocomm Research (I2R), A*STAR, Singapore |
| Pseudocode | Yes | Algorithm 1 Active Learning with Uncertainty-Weighted Embeddings |
| Open Source Code | No | The paper does not provide an explicit statement or link for the open-source code of the methodology described within this paper. |
| Open Datasets | Yes | Datasets We evaluate our method on CIFAR-10 (Krizhevsky et al., 2009) and CIFAR-100 datasets (Krizhevsky et al., 2009)... We evaluate our method on Cityscapes dataset (Cordts et al., 2016)... We evaluate our method on PASCAL VOC0712 dataset (Everingham et al., 2010)... We present the results on mini-Imagenet (Vinyals et al., 2016). |
| Dataset Splits | Yes | We evaluate our method on CIFAR-10 (Krizhevsky et al., 2009) and CIFAR-100 datasets (Krizhevsky et al., 2009), each consisting of 50,000 training images and 10,000 testing images of resolution 32 × 32 × 3. We use the training set as the initial unlabelled pool and report the model performance on the testing set. |
| Hardware Specification | No | The computational work for this article was partially performed on resources of the National Supercomputing Centre, Singapore (https://www.nscc.sg). |
| Software Dependencies | No | We conduct our experiments based on the open-source MMSegmentation (Contributors, 2020) framework. Our experiments are based on the open-source MMDetection (Chen et al., 2019) framework. |
| Experiment Setup | Yes | The hyperparameters for training on both CIFAR-10 and CIFAR-100 are set as follows: batch size = 256, total epochs = 100, initial learning rate = 2e-2 which is decayed by 0.5 after epoch 60 and 80. |