S2M2 Charting the Right Manifold: Manifold Mixup for Few-shot Learning

A few-shot classification algorithm: Charting the Right Manifold: Manifold Mixup for Few-shot Learning

Our code is built upon the code base of A Closer Look at Few-shot Classification and Manifold Mixup: Better Representations by Interpolating Hidden States

Running the code

Donwloading the dataset and create base/val/novel splits:

miniImageNet

• Change directory to filelists/miniImagenet/
• Run 'source ./download_miniImagenet.sh'

CUB

• Change directory to filelists/CUB/
• Run 'source ./download_CUB.sh'

CIFAR-FS

• Change directory to filelists/cifar/
• Run 'source ./download_cifar.sh'

Training

DATASETNAME: miniImagenet/cifar/CUB/tieredImagenet

METHODNAME: S2M2_R/rotation/manifold_mixup

For CIFAR-10

python train_cifar.py --method [METHODNAME] --model WideResNet28_10 --batch_size <batch_size> --stop_epoch <stop_epoch>

For miniImagenet/CUB/tieredImagenet

python train.py --dataset [DATASETNAME] --method [METHODNAME] --model WideResNet28_10 --batch_size <batch_size> --stop_epoch <stop_epoch>

• For tieredImagenet Novel/Base split taken from https://github.com/renmengye/few-shot-ssl-public

---

Example Training script to replicate our result on CUB Dataset:

• python train.py --dataset CUB --method rotation --model WideResNet28_10 --stop_epoch 200 --batch_size 64 --test_batch_size 16
• python train.py --dataset CUB --method S2M2_R --model WideResNet28_10 --stop_epoch 100 --batch_size 64 --test_batch_size 16

Fetching pretrained WideResNet_28_10 model checkpoints for evaluation

Directory path to save models should be: checkpoints/[DATASETNAME]/WideResNet28_10_[METHODNAME]/

Pre-trained mdoels can be downloadeded from https://drive.google.com/open?id=1S-t56H8YWzMn3sjemBcwMtGuuUxZnvb_. Move the tar files for each dataset into 'checkpoints' folder and untar it if required.

Few-shot evaluation

Create an empty 'features' directory inside 'S2M2'

python save_features.py --dataset [DATASETNAME] --method [METHODNAME] --model WideResNet28_10
python test.py --dataset [DATASETNAME] --method [METHODNAME] --model WideResNet28_10 --n_shot [1/5]

Features of pre-trained network can also be be directly downloaded at this link 'https://drive.google.com/open?id=1JtA7p3sDPksvBmOsJuR4EHw9zRHnKurj' for easy evaluation without the need to download datasets and models. Move the tar files for each dataset into 'features' folder and untar it.

Comparison with prior/current state-of-the-art methods on mini-ImageNet, CUB and CIFAR-FS dataset.

Note: We implemented LEO on CUB dataset. Other numbers are reported directly from the paper.

Method	mini-ImageNet		CUB		CIFAR-FS
	1-shot	5-shot	1-shot	5-shot	1-shot	5-shot
Baseline++	57.33 +- 0.10	72.99 +- 0.43	70.4 +- 0.81	82.92 +-0.78	67.5 +- 0.64	80.08 +- 0.32
LEO	61.76 +- 0.08	77.59 +- 0.12	68.22+- 0.22	78.27 +- 0.16	-	-
DCO	62.64 +- 0.61	78.63 +- 0.46	-	-	72.0 +- 0.7	84.2 +- 0.5
Manifold Mixup	57.6 +- 0.17	75.89 +- 0.13	73.47 +- 0.89	85.42 +- 0.53	69.20 +- 0.2	83.42 +- 0.15
Rotation	63.9 +- 0.18	81.03 +- 0.11	77.61 +- 0.86	89.32 +- 0.46	70.66 +- 0.2	84.15 +- 0.14
S2M2_R	64.93 +- 0.18	83.18 +- 0.11	80.68 +- 0.81	90.85 +- 0.44	74.81 +- 0.19	87.47 +- 0.13

If you use this code for your research, Please cite using

@inproceedings{mangla2020charting,
  title={Charting the right manifold: Manifold mixup for few-shot learning},
  author={Mangla, Puneet and Kumari, Nupur and Sinha, Abhishek and Singh, Mayank and Krishnamurthy, Balaji and Balasubramanian, Vineeth N},
  booktitle={The IEEE Winter Conference on Applications of Computer Vision},
  pages={2218--2227},
  year={2020}
}

References

A Closer Look at Few-shot Classification

Meta-Learning with Latent Embedding Optimization

Meta Learning with Differentiable Convex Optimization

Manifold Mixup: Better Representations by Interpolating Hidden States