Algorithm 1. Calculate similarity.

Input: Support set $S=\{s_1,s_2,\dots ,s_n\}$, Query sample xq

Output: Similarity score Score

1: $n\leftarrow$ size of S                             ⊳ Get the size of support set S

2: $W\leftarrow 0$                                     ⊳ Initialize total weight W

3: $\theta \leftarrow 0.008$                                     ⊳ Set hyperparameter θ

4: $X_{proto}\leftarrow 0$                          ⊳ Initialize prototype vector Xproto

5: if $n>1$ then              ⊳ If support set contains more than one sample

6:   for $i\leftarrow 1$ to n do             ⊳ Loop over each sample in the support set

7:     $D_i\leftarrow (s_i-x_q{)}^2$                 ⊳ Compute Euclidean distance Di

8:     $W\leftarrow W+exp(\theta \times D_i)$                  ⊳ Calculate total weight W

9:   end for

10:  for $i\leftarrow 1$ to n do         ⊳ Loop over each sample in the support set again

11:     $w_i\leftarrow exp(\theta \times D_i)/W$                      ⊳ Compute weight wi

12:    $X_{proto}\leftarrow X_{proto}+w_i\times s_i$            ⊳ Update prototype vector Xproto

13:  end for

14:   $Score\leftarrow (X_{proto}-q_i{)}^2$                   ⊳ Calculate similarity score

15: else                             ⊳ If support set contains only one sample

16:   $Score\leftarrow (s_1\times q_i)$          ⊳ Calculate similarity score for single sample

17: end if