Some mismatches between the code and the paper and some error in the code / paper.

[bad-meets-joke]

Hello,

Really thank you for releasing the code. However, I find several obvious mismatches between the code and the paper, and I also find some error.

• https://github.com/shiming-chen/MSDN/blob/ec3598bc6639732dfc46205de0e4cb4958774a1e/core/MSDN.py#L316 the softmax operation is along dim=-1, which represents the size of visual feature set, not like Eq. 1 in the paper
• https://github.com/shiming-chen/MSDN/blob/ec3598bc6639732dfc46205de0e4cb4958774a1e/core/MSDN.py#L376 the softmax operation is along dim=1, which represents the number of attributes , not like Eq. 4 in the paper
• https://github.com/shiming-chen/MSDN/blob/ec3598bc6639732dfc46205de0e4cb4958774a1e/core/MSDN.py#L380 and https://github.com/shiming-chen/MSDN/blob/ec3598bc6639732dfc46205de0e4cb4958774a1e/core/MSDN.py#L393 are totally different to Eq. 6 and subsequent PHI(x_i).
• https://github.com/shiming-chen/MSDN/blob/ec3598bc6639732dfc46205de0e4cb4958774a1e/core/MSDN.py#L202 the self.is_conservative is set True, which means the denominator of the first term in Eq. 7 is summed on C not C^s
• https://github.com/shiming-chen/MSDN/blob/ec3598bc6639732dfc46205de0e4cb4958774a1e/core/MSDN.py#L177 The sum on C^u of the second term in Eq. 7 should be placed after the symbol log
• https://github.com/shiming-chen/MSDN/blob/ec3598bc6639732dfc46205de0e4cb4958774a1e/core/MSDN.py#L237 line 238-242 do not calculate the l2 distance in the right way. Line 245-249 change the predictions to probabilities by F.softmax(), and the JSD eqution in the paper is wrong.

Could you please give some explanation about these mismatches, and correct the wrong code and equations?

Thanks.