DataTransform_FD funtion implement is inconsistent with the paper?

[iorange-77]

In DataTransform_FD funtion： li = np.random.randint(0, 2, size=[sample.shape[0]]) # there are two augmentations in Frequency domain li_onehot = one_hot_encoding(li) aug_1[1 - li_onehot[:, 0]] = 0 # the rows are not selected are set as zero. aug_2[1 - li_onehot[:, 1]] = 0

The above four lines of code only assign 0 to the values of rows 0 and 1 in aug_1 and aug_2, which is different from removing frequency components and adding frequency components.

[xiangzhang1015]

Hi,

Thanks for pointing that out. The four lines you mentioned are the old version of frequency augmentation, they should be commented out. The augmentations are implemented by the following:

def DataTransform_FD(sample, config):
    aug_1 = remove_frequency(sample, pertub_ratio=0.1)
    aug_2 = add_frequency(sample, pertub_ratio=0.1)
    aug_F = aug_1 + aug_2
    return aug_F

def remove_frequency(x, pertub_ratio=0.0):
    mask = torch.FloatTensor(x.shape).uniform_() > pertub_ratio  # maskout_ratio are False
    mask = mask.to(x.device)
    return x*mask

def add_frequency(x, pertub_ratio=0.0):

    mask = torch.FloatTensor(x.shape).uniform_() > (1-pertub_ratio)  # only pertub_ratio of all values are True
    mask = mask.to(x.device)

Moreover, we have updated the TFC implementation. Please check more details in the Updates on Jan 2023 section of the repo readme. In summary:

• Fixed bugs, cleaned the codes, and added comments for better understanding.
• For the contrastive encoders (in both time and frequency domains), we replaced the 3 layers of CNN blocks with 2 layers of Transformer.
• For the downstream classifier, we added a KNN classifier in parallel with the original MLP (2-layer) classifier.
• Shared more ideas that may improve the TF-C framework in follow-up works.