ScholarGate
Asistente

Comparar métodos

Revisa los métodos seleccionados uno junto a otro; las filas que difieren aparecen resaltadas.

Incrustaciones de oraciones débilmente supervisadas×Incrustaciones de oraciones auto-supervisadas×
CampoAprendizaje profundoAprendizaje profundo
FamiliaMachine learningMachine learning
Año de origen2016–20192019–2021
Autor originalRatner et al. (weak supervision framework); Reimers & Gurevych (sentence embeddings)Gao, T., Yao, X., & Chen, D. (SimCSE); Reimers, N. & Gurevych, I. (Sentence-BERT)
TipoRepresentation learning under weak supervisionSelf-supervised representation learning
Fuente seminalRatner, A., De Sa, C., Wu, S., Selsam, D., & Re, C. (2016). Data Programming: Creating Large Training Sets, Quickly. Advances in Neural Information Processing Systems (NeurIPS), 29. link ↗Gao, T., Yao, X., & Chen, D. (2021). SimCSE: Simple Contrastive Learning of Sentence Embeddings. Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing (EMNLP), 6894–6910. DOI ↗
AliasWS sentence embeddings, noisy-label sentence representation learning, weakly supervised sentence representation, distant-supervision sentence embeddingsself-supervised sentence representation learning, contrastive sentence embeddings, SimCSE, unsupervised sentence encoders
Relacionados65
ResumenWeakly supervised sentence embeddings train dense sentence representations using noisy, heuristic, or programmatically generated labels instead of costly human annotation. Labeling functions — rules, distant supervision signals, or lightweight classifiers — supply approximate supervision that a label model aggregates into probabilistic labels, which then guide the sentence encoder to produce task-useful representations at scale.Self-supervised sentence embeddings train a neural encoder to map sentences into a dense vector space without requiring manually labeled pairs. By constructing positive examples automatically — for instance by passing the same sentence through dropout twice — and using contrastive objectives, the model learns semantically rich representations that transfer well to similarity, retrieval, and classification tasks.
ScholarGateConjunto de datos
  1. v1
  2. 2 Fuentes
  3. PUBLISHED
  1. v1
  2. 2 Fuentes
  3. PUBLISHED

Ir a la búsqueda Descargar diapositivas

ScholarGateComparar métodos: Weakly supervised sentence embeddings · Self-supervised Sentence Embeddings. Recuperado el 2026-06-17 de https://scholargate.app/es/compare