SLAM: A Malware Detection Method Based on Sliding Local Attention Mechanism

Algorithm 3

Construct SLAM framework.

	Input: input_vector, step_size
	Output: output
	function MAKE_SLAM (input_vector)
	Initialize length
	Initialize CNN function from CNN layer
	Initialize Softmax function from Dense layer
	Initialize concatenate as a middle layer
	length = get the length of input_vector
	for (index = 0, index < length, index += step_size) do
	temp_tensor = SPLIT_TENSOR (input_vector, index, step_size)
	local_tensor = LOCAL_ATTENTION (temp_tensor, temp_tensor, temp_tensor)
	append local_tensor into concatenate
	end for
	tensor = CNN (concatenate)
	output = Softmax (tensor)
	return output
	end function