4 papers
DICE: Decoupling Capability from Intervention Necessity in LLM Tutoring
Sayantan Pal, Kaiyi Ji, Rohini K. Srihari
Fluent guidance is not the same as useful intervention. LLM tutors are typically trained to generate the next teacher utterance, implicitly assuming that every student turn warrant…
Beyond the Blood Draw: Explainable Machine Learning for Non-Invasive Dysglycemia Risk Screening
Black Sun, Chenyi Zhang, Kaiyi Ji +1
Dysglycemia, encompassing both prediabetes and diabetes, affects huge numbers of adults worldwide, yet many of them remain undiagnosed. We developed and validated machine-learning…
Turning Back Without Forgetting: Selective Backward Refinement for Parameter-Efficient Continual Learning
Anushka Tiwari, Kaiyi Ji
While prompt-based parameter-efficient continual learning mitigates catastrophic forgetting by isolating task-specific prompts, this isolation also limits later tasks from improvin…
GRID: Scaling Task-Agnostic Inference in Continual Prompt Tuning
Anushka Tiwari, Sayantan Pal, Rohini K. Srihari +1
Prompt-based continual learning (CL) offers a parameter-efficient way to adapt large language models (LLMs) across task sequences. However, existing methods often rely on task-awar…