23
https://pubmed.ncbi.nlm.nih.gov/38114881
This study proposes a new computational model that effectively balances language and vision by aligning textual embeddings with visual information, improving the quality of embeddings for both concrete and abstract words.