Website: https://www.euromlsys.eu/Slack: #workshop-euromlsys
Organizers: Eiko Yoneki (University of Cambridge), Paul Patras (University of Edinburgh)
Schedule: https://www.euromlsys.eu/#schedule (all times are in British Summer Time)Note: Please submit your questions to EuroSys’21 Slack Channel #workshop-euromlsys!
15:00 - 15:10 Introduction15:10 - 15:50 Keynotes1: Zhihao Jia (CMU & Facebook)- Automated Discovery of Machine Learning Optimizations
15:50 - 16:55 Session 1: Systems, Compiler and PPL- DISC: A Dynamic Shape Compiler for Machine Learning Workloads
- High-Dimensional Bayesian Optimization with Multi-Task Learning for RocksDB
- Vate: Runtime Adaptable Probabilistic Programming for Java
- DistIR: An Intermediate Representation for Optimizing Distributed Neural Network
16:55 - 17:05 Break
17:05 - 18:10 Session 2: Model Optimisation and NAS- Optimizing Inference Performance of Transformers on CPUs
- Learned Low Precision Graph Neural Networks
- uNAS: Constrained Neural Architecture Search for Microcontrollers
- Towards Mitigating Device Heterogeneity in Federated Learning via Adaptive Model Quantization
18:10 - 18:20 Break
18:20 - 19:00 Keynotes 2: Anna Goldie (Google Brain & Stanford University)- Deep Reinforcement Learning for Graph Placement: Model Parallelism and Chip Floorplanning
19:00 - 19:05 Break
19:05 - 20:10 Session 3: Scheduling, Training and Prediction- Are we there yet? Estimating Training Time for Recommendation Systems
- Interference-Aware Scheduling for Inference Serving
- Developing a Siamese Network for Intrusion Detection Systems
- Predicting CPU Usage for Proactive Autoscaling
20:10 - 20:15 Break
20:15 - 20:50 Poster Session- Queen Jane Approximately: Enabling Efficient Neural Network Inference with Context-Adaptivity
- DPD-InfoGAN: Differentially Private Distributed InfoGAN
- Towards Optimal Configuration of Microservices
- AutoAblation: Automated Parallel Ablation Studies for Deep Learning
- Towards a General Framework for ML-based Self-tuning Databases
- Fast Optimisation of Convolutional Neural Network Inference using System Performance Models
20:50 - 21:00 Wrap-up
