S$^3$: Sign-Sparse-Shift Reparametrization for Effective Training of Low-bit Shift Networks Jan 1, 2021· X. Li , B. Liu , Y. Yu , W. Liu , C. Xu , V. Nia · 0 min read PDF Cite Type Conference paper Publication Advances in Neural Information Processing Systems (NeurIPS) Last updated on Jan 1, 2021 Conference ← Quantifying and Improving Transferability in Domain Generalization Jan 1, 2021 The Art of Abstention: Selective Prediction and Error Regularization for Natural Language Processing Jan 1, 2021 →