Search papers, labs, and topics across Lattice.
Korea Advanced Institute of Science and Technology
3
0
6
3
TELLME boosts language model performance by over 23% while slashing the need for extensive domain-specific datasets through innovative quiz-based training.
Decomposing text prompts into semantic units and using VQA for fine-grained self-reflection dramatically improves image generation quality, especially for complex compositions.
VLMs still struggle to combine visual and textual information for multi-hop reasoning, but a new automatically generated dataset, CRIT, can help them learn.
