Pointer Network
An attention-based seq2seq model that directly points to input elements as outputs via attention weights, with an output space that varies dynamically with input size, making it suitable for combinatorial optimization problems such as the TSP and path planning.
Timeline (last 90 days)
When a trained Pointer Network encounters a different number of customers, the solution quality drops noticeably.
A pointer network is a class of attention-based sequence-to-sequence model that directly points to elements in the input as outputs via attention weights, so the output space changes dynamically with the input size.
All Facts (2)
When a trained Pointer Network encounters a different number of customers, the solution quality drops noticeably.
50%UnverifiedA pointer network is a class of attention-based sequence-to-sequence model that directly points to elements in the input as outputs via attention weights, so the output space changes dynamically with the input size.
50%