@misc{indiciae96946e353d88,
  title = {Large Stepsize Gradient Descent for Logistic Loss: Non-Monotonicity of the Loss Improves Optimization Efficiency},
  author = {Jingfeng Wu and Peter L. Bartlett and Matus Telgarsky and Bin Yu},
  year = {2024},
  url = {https://arxiv.org/abs/2402.15926},
  note = {Source identifier: 2402.15926}
}