Idanwo ti o sopọ mọ itọsọna · Alabọde Ipele

ELMo Contextual ifibọ adanwo

Review how ELMo's bidirectional language model creates context-sensitive word vectors and combines recurrent layers for downstream tasks.

Jẹmọ itọsọna awọn ọnaElmo Contextual ifibọ
Ibeere 1 ti 8

Kini iyatọ bọtini laarin ELMo ati awọn ifibọ iṣaaju bii Word2Vec?

Awotẹlẹ gbogbo ibeere8 ibeere

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

  1. Kini iyatọ bọtini laarin ELMo ati awọn ifibọ iṣaaju bii Word2Vec?
    • ELMo n fun ni ọrọ kanna ni oriṣiriṣi awọn fekito da lori ọrọ-ọrọ ọrọ rẹ
    • ELMo nlo awọn fekito kekere
    • ELMo nikan ṣiṣẹ lori awọn ọrọ ẹyọkan, kii ṣe awọn gbolohun ọrọ
    • ELMo ko nilo data ikẹkọ
  2. Kini faaji nkankikan ti ELMo lo lati ka ọrọ?
    • A Ayirapada pẹlu ara-akiyesi
    • Awoṣe ede LSTM bidirectional ti o jinlẹ
    • A convolutional image nẹtiwọki
    • Nẹtiwọọki ifunni ti o rọrun
  3. Bawo ni ELMo ṣe darapo awọn ipele inu rẹ fun iṣẹ-ṣiṣe ti isalẹ?
    • O nlo nikan ni ipari oke Layer
    • O ṣe aropin gbogbo awọn ipele ni dọgbadọgba ati ṣatunṣe abajade
    • O nlo ikẹkọ, awọn akojọpọ iwuwo iṣẹ-ṣiṣe pato ti gbogbo awọn fẹlẹfẹlẹ
    • O mu ọkan Layer ni ID kọọkan akoko
  4. Kini ELMo nlo gẹgẹbi awọn ẹya titẹ sii lati mu awọn ọrọ ti a ko ri?
    • Wiwa gbogbo-ọrọ nikan
    • Awọn ID iwe-itumọ ti o wa titi
    • Baiti-bata-awọn ọrọ-ọrọ abẹlẹ lati BERT
    • Awọn aṣoju CNN ipele ti iwa
  5. Ni aijọju nigbawo ni a ṣe afihan ELMo ati nipasẹ tani?
    • 2018, nipasẹ Ile-iṣẹ Allen fun awọn oniwadi AI
    • 2013, nipasẹ Google's Word2Vec egbe
    • Ọdun 2020, nipasẹ OpenAI
    • 2015, nipasẹ Facebook AI
  6. Which corpus was used to train the bidirectional language model described in the ELMo paper?
    • The One Billion Word Benchmark
    • The Penn Treebank only
    • A corpus of labeled images
    • The GloVe Common Crawl vectors
  7. What language-model prediction does ELMo's forward LSTM learn?
    • Predicting the previous word from future context
    • Assigning a fixed vector to every document
    • Predicting the next word from the preceding context
    • Predicting a sentence's image resolution
  8. What linguistic information do ELMo's lower and higher LSTM layers tend to capture?
    • Lower layers encode image color, while higher layers encode audio pitch
    • Every layer stores only word frequency and ignores context
    • Higher layers are always limited to punctuation prediction
    • Lower layers tend toward syntax, while higher layers capture more context-dependent semantics