Search papers, labs, and topics across Lattice.
York University, Ontario, Canada, Vector Institute, Ontario, Canada, Connected Minds, Ontario, Canada
4
0
6
Human-aligned evaluation can transform therapeutic response generation, leading to substantial improvements in mental health support quality.
LLMs may inadvertently reinforce user dependence by prioritizing directive advice over fostering independent reflection in emotionally sensitive conversations.
Training with OpenMedReason boosts VQA accuracy by 20%, showcasing a leap in reasoning capabilities for medical vision-language models.
Fine-grained preference optimization can dramatically enhance the clinical accuracy of LVLMs by correcting only the clinically erroneous spans while preserving linguistic style.
