Maswali yanayounganishwa na mwongozo · Kati Kiwango

Maswali ya Kupachika Sentensi-BERT

Jaribu jinsi upachikaji wa sentensi unavyoauni utafutaji wa kisemantiki unaoweza kupanuka, ikijumuisha uwakilishi wa Matryoshka na uwekaji upya wa nafasi ya kisimbaji.

Njia za mwongozo zinazohusianaSentensi ya Bert Embeddings
Swali 1 ya 8

Sentensi-BERT inasuluhisha shida gani ya msingi na BERT wazi?

Hakiki kila swali8 maswali

Majibu na maelezo huonekana unapojibu maswali.

  1. Sentensi-BERT inasuluhisha shida gani ya msingi na BERT wazi?
    • Kulinganisha sentensi nyingi kwa jozi na BERT ni polepole sana
    • BERT haiwezi kusoma zaidi ya sentensi moja
    • BERT hufanya kazi kwa Kiingereza pekee
    • BERT haiwezi kusasishwa vizuri
  2. Sentence-BERT hutumia usanifu gani wa mtandao?
    • Avkodare moja autoregressive
    • Mtandao wa siamese wenye minara miwili ya kugawana uzani ya BERT
    • Mtandao wa picha wa kushawishi
    • Wakala wa mafunzo ya kuimarisha
  3. Which pooling strategy did the original Sentence-BERT paper evaluate for creating sentence embeddings?
    • Max akikusanyika juu ya vichwa vya umakini
    • Kwa kutumia herufi ya kwanza tu
    • Maana ya kuunganisha juu ya upachikaji wa ishara
    • Uchaguzi nasibu wa tokeni moja
  4. Sentensi zinapopachikwa, mfanano wao hupimwaje kwa kawaida?
    • Kuendesha BERT tena kwa zote mbili
    • Kuhesabu herufi zilizoshirikiwa
    • Kulinganisha ukubwa wa faili
    • Kufanana kwa cosine (bidhaa ya dot) kati ya vekta
  5. What training data did the original Sentence-BERT paper use for NLI fine-tuning?
    • Natural Language Inference examples labeled for entailment, contradiction and neutral relations
    • Only satellite images with land-cover labels
    • Only audio recordings with speaker labels
    • Only source code and bug reports
  6. In a possible semantic-retrieval workflow, what can a Sentence-BERT bi-encoder do?
    • Cross-attend to every query-candidate pair before any encoding
    • Encode a query and candidate texts separately so vector similarity can rank candidates
    • Guarantee retrieval quality for every corpus
    • Replace the need to evaluate relevance
  7. What does Matryoshka Representation Learning enable for an embedding?
    • It guarantees identical quality at every vector size
    • It converts embeddings into labeled text tokens
    • A single representation can be truncated to different dimensions to trade storage or compute for task performance
    • It removes the need for similarity comparisons
  8. In a common hybrid retrieval pipeline, what happens after a bi-encoder retrieves candidate passages?
    • The candidate text is converted into audio
    • The bi-encoder is discarded and no ranking occurs
    • The system averages all documents into one sentence
    • A cross-encoder can score query-passage pairs to rerank the candidates