Integrating Specialized Classifiers Based on Continuous Time Markov Chain
Authors: Zhizhong Li, Dahua Lin
IJCAI 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | On large public datasets, the proposed method obtains considerable improvement compared to mainstream ensemble methods, especially when the classifier coverage is highly unbalanced. ... We tested the proposed methods on large datasets. The results show that the presented methods can lead to further performance gains over state-of-the-art deep networks. |
| Researcher Affiliation | Academia | Zhizhong Li and Dahua Lin The Chinese University of Hong Kong {lz015, dhlin}@ie.cuhk.edu.hk |
| Pseudocode | No | The paper describes the methodology using prose and mathematical equations but does not include any clearly labeled pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not provide any explicit statement about releasing source code or a link to a code repository. |
| Open Datasets | Yes | We tested the proposed method in comparison with mainstream baselines on two commonly used benchmarks, ILSVRC 2012 [Russakovsky et al., 2015], and CIFAR-100 [Krizhevsky and Hinton, 2009]. The ILSVRC benchmark is constructed based on Image Net, which contains 1.2M training images in 1000 classes, and 50K validation images for testing. The CIFAR-100 dataset consists of 60K images in 100 classes, of which 50K are for training and the other 10K are for testing. |
| Dataset Splits | No | The ILSVRC benchmark is constructed based on Image Net, which contains 1.2M training images in 1000 classes, and 50K validation images for testing. The CIFAR-100 dataset consists of 60K images in 100 classes, of which 50K are for training and the other 10K are for testing. While a 'validation' set is mentioned for ILSVRC, its usage 'for testing' implies it might serve as the test set, and no explicit separate validation split for hyperparameter tuning is provided for either dataset. |
| Hardware Specification | No | The paper mentions 'computational budget' and discusses network architectures suitable for image size, but does not specify any particular hardware details such as GPU models, CPU types, or memory used for experiments. |
| Software Dependencies | No | The paper mentions deep learning architectures (Res Net-56, pre-activation Res Net-101, Inception-Res Net-v2) and optimizers (RMSProp, SGD), but does not specify any software dependencies with version numbers (e.g., Python, TensorFlow, PyTorch versions). |
| Experiment Setup | Yes | For ILSVRC, 'We followed the standard practice to train the networks, using random-crop augmentation, RMSProp [Tieleman and Hinton, 2012] with decay 0.9 and ϵ = 1.0, weight decay 0.0005, and batch size 512. The learning rate was initialized to 0.45 and scaled down by a factor 0.1 per 120K iterations. All experiments were terminated at iteration 250K to avoid overfitting.' For CIFAR-100, 'We followed the common practice to train the networks: data augmentation with random crop, SGD with momentum 0.9 and batch size 128, learning rates initialized to 0.1 and scaled down by 0.1 per 45K iterations, and weight decay 0.0005.' |