A Survey of Multimodal Learning: Methods, Applications, and Future

ACM Computing Surveys, Год журнала: 2025, Номер unknown

Опубликована: Янв. 18, 2025

The multimodal interplay of the five fundamental senses—Sight, Hearing, Smell, Taste, and Touch—provides humans with superior environmental perception learning skills. Adapted from human perceptual system, machine tries to incorporate different forms input, such as image, audio, text, determine their connections through joint modeling. As one future development artificial intelligence, it is necessary summarize progress learning. In this paper, we start form a combination provide comprehensive survey emerging subject learning, covering representative research approaches, most recent advancements, applications. Specifically, paper analyzes relationship between modalities in detail sorts out key issues application scenarios. Besides, thoroughly reviewed state-of-the-art methods datasets covered research. We then identify substantial challenges potential developing directions field. Finally, given nature survey, both modality-specific task-specific researchers can benefit advance

Язык: Английский

---

GenSumNet: A genre-specific and context-aware video summarization framework with adaptive thresholding and hybrid temporal modelling

Expert Systems with Applications, Год журнала: 2025, Номер unknown, С. 127360 - 127360

Опубликована: Март 1, 2025

Язык: Английский

---

Enhancing Video-Language Alignment Via Mining Entity Knowledge

Опубликована: Янв. 1, 2024

Язык: Английский

---

Knowledge Enhancement and Disentanglement Learning for Video Captioning

Опубликована: Янв. 1, 2024

Язык: Английский

---