Research map / Learning theory and algorithms / Prediction and decision losses
Snap ML: A Hierarchical Framework for Machine Learning
Celestine Dünner, Thomas Parnell, Dimitrios Sarigiannis, Nikolas Ioannou, Andreea Anghel, Gummadi Ravi, Madhusudanan Kandasamy, Haralampos Pozidis
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use Snap ML instead of scikit-learn | Logistic regression training · criteo-kaggle | training time: around 200 seconds -> around 20 seconds | Improved |
| Use Snap ML instead of TensorFlow | Logistic regression training · criteo-kaggle | training time: approximately 500 seconds -> around 20 seconds | Improved |
| Use NVLINK 2.0 for pipelined data streaming instead of the PCI-based setup | Snap ML streaming · criteo-tb first 200 million training examples | speedup: 3.5x | Improved |
| Increase inner CoCoA iterations t_2 above 1 | Hierarchical CoCoA · criteo-tb first billion examples, InfiniBand | time to target training suboptimality | No clear effect |
| Increase inner CoCoA iterations t_2 above 1 | Hierarchical CoCoA · criteo-tb first billion examples, 1Gbit Ethernet | time to target training suboptimality | Improved |
| Use Snap ML instead of the best previously reported TensorFlow result | Logistic regression · criteo-tb | training speed: 46x faster | Improved |
| Use Snap ML instead of TensorFlow on the same hardware | Logistic regression · criteo-tb | training speed: over 500× faster | Improved |
Closest papers by meaning
- ZipML: Training Linear Models with End-to-End Low Precision, and a Little Bit of Deep Learning (ICML 2017 · similarity 0.87)
- Pfeife: Automatic Pipeline Parallelism for PyTorch (ICML 2025 · similarity 0.86)
- Joker: Joint Optimization Framework for Lightweight Kernel Machines (ICML 2025 · similarity 0.85)
- MetaOptimize: A Framework for Optimizing Step Sizes and Other Meta-parameters (ICML 2025 · similarity 0.85)
- Sparse maximal update parameterization: A holistic approach to sparse training dynamics (NeurIPS 2024 · similarity 0.85)
- HyperPrism: An Adaptive Non-linear Aggregation Framework for Distributed Machine Learning over Non-IID Data and Time-varying Communication Links (NeurIPS 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
