Awọn nẹtiwọki Opopona ati Rekọja Awọn adanwo Awọn isopọ

Awọn idahun ati awọn alaye han bi o ṣe n mu ibeere naa.

1. A very deep network struggles with optimization. What can a skip path provide?
   • A guarantee of perfect gradients at every depth
   • A more direct signal and gradient route that can ease training
   • A proof that more layers always improve accuracy
   • A replacement for the loss function
2. Kí ni a ipilẹ péye Àkọsílẹ isiro?
   • iṣẹjade = F (x) ti o pọ nipasẹ x
   • iṣẹjade = x pin nipasẹ F(x)
   • iṣẹjade = o pọju F (x) ati x
   • iṣẹjade = F(x) pẹlu x, nibiti x jẹ ọna abuja titẹ sii
3. Kini o ṣe atilẹyin ẹrọ ẹnu-ọna ni awọn nẹtiwọọki opopona?
   • Awọn ẹnu-ọna LSTM lati awọn nẹtiwọki loorekoore
   • Convolutional pooling
   • Ipese deede
   • Jade kuro
4. Ninu nẹtiwọọki opopona kan, kini iṣakoso ẹnu-ọna T(x) iyipada?
   • Oṣuwọn ẹkọ ti oluṣeto
   • Awọn iwọn ti o wu aworan
   • Elo ti igbewọle ti yipada dipo gbigbe taara nipasẹ
   • Awọn nọmba ti fẹlẹfẹlẹ ni awọn nẹtiwọki
5. Iru faaji igbalode wo ni o da lori awọn isopọ to ku ni ayika awọn alaṣẹ rẹ?
   • K-tumo si iṣupọ
   • Igi ipinnu
   • Ipadasẹyin laini
   • Awọn transformer
6. Bawo ni awọn asopọ U-Net ṣe iranlọwọ ni pataki ipin?
   • Wọn pa kooduopo rẹ patapata
   • Wọn kọja alaye aye to dara lati koodu koodu si oluyipada
   • Wọn yi aworan pada si ọrọ
   • Wọn dinku nọmba awọn kilasi iṣelọpọ si ọkan
7. In y = F(x)T(x) + x(1-T(x)), what remains when T(x)=0?
   • Only F(x)
   • The sum 2x
   • The carried input x
   • A mandatory zero output
8. For the same coupled-gate highway form, what remains when T(x)=1?
   • The transformed value F(x)
   • Only the original x
   • F(x) plus x
   • An automatically normalized x