Lune

Profiles2019-2021

Sijia Liu

IBM Research

11Papers
0Last author
23Unique co-authors
1First author
0Solo

First and last author come from byline position, which some venues assign alphabetically.

Recent papers11

  1. RT3D: Achieving Real-Time Execution of 3D Convolutional Neural Networks on Mobile DevicesAuthor 8 of 11AAAI 2021 · 14 citations
  2. Self-Progressing Robust TrainingAuthor 3 of 6AAAI 2021 · 10 citations
  3. The Lottery Ticket Hypothesis for Pre-trained BERT NetworksAuthor 4 of 7NeurIPS 2020 · 428 citations
  4. Sign-OPT: A Query-Efficient Hard-label Adversarial AttackAuthor 5 of 6ICLR 2020 · 256 citations
  5. Is There a Trade-Off Between Fairness and Accuracy? A Perspective Using Mismatched Hypothesis TestingAuthor 5 of 6ICML 2020 · 171 citations
  6. Min-Max Optimization without Gradients: Convergence and Applications to Black-Box Evasion and Poisoning AttacksAuthor 1 of 8ICML 2020 · 68 citations
  7. Training Stronger Baselines for Learning to OptimizeAuthor 5 of 7NeurIPS 2020 · 61 citations
  8. Fast Learning of Graph Neural Networks with Guaranteed Generalizability: One-hidden-layer CaseAuthor 3 of 5ICML 2020 · 36 citations
  9. Towards Certificated Model Robustness Against Weight PerturbationsAuthor 3 of 6AAAI 2020 · 33 citations
  10. Adversarial Robustness: From Self-Supervised Pre-Training to Fine-TuningAuthor 2 of 6CVPR 2020
  11. On the Design of Black-Box Adversarial Examples by Leveraging Gradient-Free Optimization and Operator Splitting MethodAuthor 2 of 7ICCV 2019 · 61 citations
Dusk over the sea between two cliffs drawn in fine vertical lines