השוואת שיטות

	עיבוד שפה טבעית רב-מודאלי	מנגנון קשב	BERT Embeddings
תחום	כריית טקסט	למידה עמוקה	כריית טקסט
משפחה	Process / pipeline	Machine learning	Process / pipeline
שנת המקור	2021 (modern era, CLIP onward)	2015	2019
הוגה השיטה	Radford et al. (OpenAI) — CLIP, 2021; Li et al. — BLIP-2, 2023	Bahdanau, D.; Luong, M.T.	Devlin, Chang, Lee & Toutanova (Google AI)
סוג	Cross-modal understanding and generation pipeline	Neural attention layer (encoder-decoder)	Contextual transformer text-representation method
מקור מכונן	Radford, A., Kim, J.W., Hallacy, C., Ramesh, A., Goh, G., Agarwal, S., Sastry, G., Askell, A., Mishkin, P., Clark, J., Krueger, G., & Sutskever, I. (2021). Learning Transferable Visual Models From Natural Language Supervision. Proceedings of the 38th International Conference on Machine Learning (ICML), 8748–8763.	Bahdanau, D., Cho, K. & Bengio, Y. (2015). Neural Machine Translation by Jointly Learning to Align and Translate. ICLR.	Devlin, J., Chang, M.-W., Lee, K. & Toutanova, K. (2019). BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. NAACL-HLT, 4171-4186.
כינויים	Çok Kipli NLP (Multimodal NLP), vision-language models, multimodal learning	Dikkat Mekanizması (Bahdanau / Luong Attention), dikkat mekanizmasi, neural attention, additive attention	contextual embeddings, transformer embeddings, BERT Tabanlı Metin Gömülmeleri
קשורות	4	5	4
תקציר	Multimodal NLP is a family of natural-language-processing pipelines that combine text with one or more additional data modalities — most commonly images, but also audio and video — to perform understanding and generation tasks such as visual question answering, image captioning, and multimodal sentiment recognition. The field gained its modern form with CLIP (Radford et al., 2021) and has since advanced through architectures such as BLIP-2 (Li et al., 2023) that bridge frozen image encoders and large language models.	The attention mechanism, introduced by Bahdanau, Cho and Bengio in 2015 and refined by Luong, Pham and Manning the same year, lets a sequence decoder dynamically learn which of the encoder's outputs to focus on at each step. Before the Transformer, it substantially improved machine-translation quality by freeing models from compressing an entire input into a single fixed vector.	BERT-based text embeddings, introduced by Devlin and colleagues at Google AI in 2019, turn text into context-sensitive dense vectors using a bidirectional Transformer encoder. Because the meaning of a word shifts with its context, BERT produces richer representations than static methods such as Word2Vec or topic models like LDA.