4 papers · 1 filter
One-Step Curvature Probes Miss the Fitting Operator: Retained Capacity and Terminal Null-Space Correction for Continual Learning
Abu Sa-Adat Mohamed Moon-Im Al Ahsan, Ibne Farabi Shihab, Md Najmus Swaqeeb
A one-step curvature probe evaluates an initial direction, whereas continual learners are judged after reaching comparable new-task fit. In an overparameterized linearization, proj…
When Is Language-Model Advice Worth Buying? A Matched-Budget Study of Predictive Gating in Reinforcement Learning
Ibne Farabi Shihab, Md Najmus Swaqeeb, Abu Sa-Adat Mohamed Moon-Im Al Ahsan +1
Language-model advice can accelerate reinforcement learning, but every call has a price, and a returned suggestion may be unparsable, unsafe to execute, or no better than the agent…
Spectral-Guided Diffusion: Accelerating Inference via Static Spectral Layer Scheduling
Ibne Farabi Shihab, Abu Sa-Adat Mohamed Moon-Im Al Ahsan, Anuj Sharma
Diffusion inference repeatedly evaluates the same large network. We ask whether pretrained weights alone can identify residual branches that need not be recomputed throughout the t…
Spectral Tail Interventions in Decoder-Only Language Models: Reasoning-Sensitive Weight Structure from Controlled Surgery
Ibne Farabi Shihab, Sanjida Akhter, Md Najmus Swaqeeb +2
Weight-space structure often correlates with language-model behavior, but correlation alone does not establish computational involvement. We study concentrated upper spectral tails…