Answer: Decrease the learning rate hyperparameter.

Oscillating loss during neural network training is a classic symptom of a learning rate that is too high. When the learning rate is excessive, the optimizer overshoots the minimum of the loss function, causing the loss to bounce back and forth rather than converging smoothly. Decreasing the learning rate (Option B) allows for more precise updates to the model parameters, enabling gradual convergence toward the optimal values. The community discussion strongly supports this, with 89% selecting B and multiple comments explaining that high learning rates cause oscillation. Other options are less suitable: decreasing batch size (A) can increase noise in gradient estimates; increasing learning rate (C) would worsen oscillation; increasing batch size (D) may smooth updates but doesn't directly address the core issue of an overly aggressive learning rate.

Author: LeetQuiz Editorial Team