Profiles2021-2024
Zhao Song
zsong@adobe.com. Adobe Research
15Papers
2Last author
39Unique co-authors
0First author
0Solo
First and last author come from byline position, which some venues assign alphabetically.
Recent papers15
- How to Capture Higher-order Correlations? Generalizing Matrix Softmax Attention to Kronecker ComputationAuthor 2 of 2ICLR 2024 · 53 citations
- On Computational Limits of Modern Hopfield Models: A Fine-Grained Complexity AnalysisAuthor 3 of 6ICML 2024 · 47 citations
- H2O: Heavy-Hitter Oracle for Efficient Generative Inference of Large Language ModelsAuthor 7 of 12NeurIPS 2023 · 1,003 citations
- Deja Vu: Contextual Sparsity for Efficient LLMs at Inference TimeAuthor 6 of 11ICML 2023 · 318 citations
- Fast Attention Requires Bounded EntriesAuthor 2 of 2NeurIPS 2023 · 115 citations
- Bypass Exponential Time Preprocessing: Fast Neural Network Training via Weight-Data Correlation PreprocessingAuthor 3 of 5NeurIPS 2023 · 32 citations
- Differentially Oblivious Relational Database OperatorsAuthor 4 of 6VLDB 2023 · 12 citations
- Exact Representation of Sparse Networks with Symmetric Nonnegative EmbeddingsAuthor 6 of 7NeurIPS 2023 · 5 citations
- Smoothed Online Combinatorial Optimization Using Imperfect PredictionsAuthor 2 of 4AAAI 2023 · 1 citation
- Pixelated Butterfly: Simple and Efficient Sparse training for Neural Network ModelsAuthor 5 of 7ICLR 2022 · 94 citations
- Perfectly Balanced: Improving Transfer and Robustness of Supervised Contrastive LearningAuthor 5 of 7ICML 2022 · 58 citations
- Bounding the Width of Neural Networks via Coupled Initialization A Worst Case AnalysisAuthor 3 of 4ICML 2022 · 17 citations
- Fast Graph Neural Tangent Kernel via Kronecker SketchingAuthor 3 of 5AAAI 2022 · 9 citations
- Evaluating Gradient Inversion Attacks and Defenses in Federated LearningAuthor 3 of 5NeurIPS 2021 · 419 citations
- Scatterbrain: Unifying Sparse and Low-rank AttentionAuthor 4 of 6NeurIPS 2021 · 165 citations
