Program
The workshop will take place on the 9th of December 2022. The schedule of the day is provided below:
| Local Time | Description |
|---|---|
| 10.00 - 10.15 | Opening remarks (video) (Presented by Mario Almeida) |
| 10.15 - 11.15 | Keynote #1 Programmable Networks for Distributed Deep Learning: Advances and Perspectives (video) Marco Canini (KAUST) |
| 11.15 - 11.45 | Session #1: Transient resources (co-ordinated by Mario Almeida) - Exploring Learning Rate Scaling Rules for Distributed ML Training on Transient Resources (video) Joel Andre, Foteini Strati, Ana Klimovic (ETH Zurich) - SmartNICs at Edge for Transient Compute Elasticity (video) Diman Zad Tootaghaj, Anu Mercian (Hewlett Packard Labs), Vivek Adarsh (University of California, Santa Barbara), Mehrnaz Sharifian (University of California Davis), Puneet Sharma (Hewlett Packard Labs) |
| 11.45 - 12.00 | Break |
| 12.00 - 12.30 | Session #2: Federated Learning (co-ordinated by Stefanos Laskaridis) - Rethinking Normalization Methods in Federated Learning (video) Zhixu Du, Jingwei Sun, Ang Li (Duke University), Pin-Yu Chen (IBM Research AI), Jianyi Zhang, Hai “Helen” Li, Yiran Chen (Duke University) - Towards Predicting Client Benefit and Contribution in Federated Learning From Data Imbalance (video) Christoph Düsing, Philipp Cimiano (CITEC(Cluster of Excellence Cognitive Interaction Technology), University Bielefeld) |
| 12.30 - 14.00 | Lunch Break |
| 14.00 - 15.00 | Keynote #2: Scalable Input Data Processing for Resource-Efficient Machine Learning (video) Ana Klimovic (ETH Zurich) |
| 15.00 - 15.30 | Session #3: Efficient & robust ML deployment (co-ordinated by Mario Almeida & Stefanos Laskaridis) - Activation Sparsity and Dynamic Pruning for Split Computing in Edge AI (video) Janek Haberer (Kiel University, Germany), Olaf Landsiedel (Kiel University, Germany and Chalmers University of Technology, Sweden) - Pelta: Shielding Transformers to Mitigate Evasion Attacks in Federated Learning (video) Simon Queyrut (University of Neuchatel), Yérom-David BROMBERG (Univ. Rennes - Inria - IRISA), Valerio Schiavoni (University of Neuchatel) |
| 15.30 - 15.45 | Break |
| 15.40 - 17.00 | Panel session (video) (co-ordinated by Stefanos Laskaridis) Dimitrios Dimitriadis (Amazon), Christos Louizos (Qualcomm AI), Ana Klimovic (ETH Zurich), Marco Canini (KAUST) |