Trans. Mach. Learn. Res. · 2023 · Journal article

Improving Continual Learning by Accurate Gradient Reconstructions of the Past

Erik Daxberger, S. Swaroop, Kazuki Osawa, Rio Yokota, Richard E. Turner, José Miguel Hernández-Lobato, Mohammad Emtiyaz Khan

Source ↗

Citations: 8

Cite

@article{daxberger2023improving,
  title = {Improving Continual Learning by Accurate Gradient Reconstructions of the Past},
  author = {Erik Daxberger and S. Swaroop and Kazuki Osawa and Rio Yokota and Richard E. Turner and José Miguel Hernández-Lobato and Mohammad Emtiyaz Khan},
  year = {2023},
  journal = {Trans. Mach. Learn. Res.},
  url = {https://www.semanticscholar.org/paper/45718eedb6398473d3805b2d0bb1163e02965949},
}