Meta Research AI System Hardware/Software Codesign

Closing Date: 15/06/2022

Support for innovative research on topics related to AI System Hardware/Software Codesign.

Meta Research works on cutting edge research with a practical focus, and builds long-term relationships with top research institutions around the world. It also publishes papers, give talks, and collaborates broadly with the academic community. Areas include data science, Artificial Intelligence (AI), machine learning and computer vision. Research Networking aims to foster further innovation in networking and to deepen Meta’s collaboration with academia.

To explore codesign opportunities across a number of new dimensions, Meta is inviting proposals for research on AI System Hardware/Software Codesign to further innovation in this area and deepen collaboration with academia. Areas of interest include, but are not limited to:

  • Recommendation models
    • Compression, quantisation, pruning, adaptive sparsity techniques.
    • Hardware-aware novel modeling techniques.
  • Hardware/software co-design for deep learning
    • Energy-efficient hardware architectures.
    • Hardware efficiency-aware neural architecture search.
    • Mixed-precision linear algebra and tensor-based frameworks.
  • Distributed training
    • Software frameworks for efficient use of programmable hardware.
    • Scalable communication-aware and data movement-aware algorithms.
    • High-performance and/or fault-tolerant communication middleware.
    • Systems/components for enabling high performance training of large-scale models (eg checkpointing, transfer learning, data reading, model publishing).
  • Performance, programmability, and efficiency at data center scale
    • Machine learning-driven data access optimisation (eg prefetching and caching).
    • Enabling large model deployment through intelligent memory and storage.
    • Training un-/self-/semi-supervised models on large scale video data sets.
    • Meta-learning and continual learning techniques.
Funding body Meta Research
Maximum value 50,000 USD
Reference ID S23782
Category Science and Technology
Fund or call Fund