Maswali yanayounganishwa na mwongozo · Kati Kiwango

Maswali ya Graph Neural Networks

Kagua upitishaji wa ujumbe wa GNN, utofauti wa vibali, ulainishaji kupita kiasi, umakini, grafu za molekuli, na humle za uga zinazopokea.

Njia za mwongozo zinazohusianaMitandao ya Neural ya Grafu
Swali 1 ya 8

Je! ni utaratibu gani wa msingi wa hesabu wa mitandao mingi ya neva ya grafu?

Hakiki kila swali8 maswali

Majibu na maelezo huonekana unapojibu maswali.

  1. Je! ni utaratibu gani wa msingi wa hesabu wa mitandao mingi ya neva ya grafu?
    • Kupitisha ujumbe na kujumlisha kati ya nodi za jirani
    • Kutelezesha mzunguuko juu ya gridi ya pikseli isiyobadilika
    • Utabiri wa ishara unaofuata unaoendelea
    • Kutembea bila mpangilio bila kujifunza
  2. Kwa nini kazi ya kujumlisha jirani ya GNN lazima iwe isiyobadilika ya vibali?
    • Kufanya mafunzo haraka
    • Kwa sababu majirani wa grafu hawana utaratibu wa asili
    • Kwa sababu grafu huelekezwa kila wakati
    • Ili kupunguza idadi ya vigezo hadi moja
  3. Je, 'over-moothing' inarejelea nini katika GNN za kina?
    • Picha zinakuwa na ukungu
    • Grafu inapoteza kingo zake
    • Upachikaji wa nodi zinazobadilika kuwa karibu kufanana kadiri tabaka zinavyoongezeka
    • Gradients kulipuka kwa infinity
  4. How does a Graph Attention Network differ from a basic GCN layer?
    • It removes all graph edges before message passing
    • It assigns each node a unique text token
    • It guarantees deeper networks cannot over-smooth
    • It learns attention coefficients for neighbors, while the GCN uses degree-normalized propagation weights
  5. Katika GNN inayotumika kwa molekuli, nodi na kingo kawaida huwakilisha nini?
    • Atomi na vifungo vya kemikali
    • Pixels na rangi
    • Maneno na sentensi
    • Watumiaji na machapisho
  6. Je, habari inaweza kusafiri kwa umbali gani hadi kwenye nodi baada ya safu za kupitisha ujumbe wa L?
    • Daima hasa hop moja
    • Hadi L hops
    • Grafu nzima katika safu moja
    • Zero humle
  7. What makes GraphSAGE an inductive graph-learning method?
    • It stores a separate free parameter vector for every possible test node
    • It requires one fixed graph and cannot process new nodes
    • It learns neighborhood aggregation functions that can be applied to unseen nodes
    • It replaces node features with graph-level labels
  8. In the original GCN propagation rule, what does symmetric degree normalization do to adjacency messages?
    • Sorts neighboring nodes alphabetically
    • Deletes every edge with a low feature value
    • Makes the graph directed by reversing all edges
    • Scales each edge contribution using the inverse square roots of the source and target node degrees