CRIMP: Compact & Reliable DNN Inference on In-Memory Processing via Crossbar-Aligned Compression and Non-ideality Adaptation.
Shuo HuaiHao KongXiangzhong LuoShiqing LiRavi SubramaniamChristian MakayaQian LinWeichen LiuPublished in: ACM Trans. Embed. Comput. Syst. (2023)
Keyphrases
- random access
- real time
- computational power
- memory management
- image processing operations
- image compression
- cost effective
- high quality
- compute intensive
- processing elements
- limited memory
- compression scheme
- compression ratio
- information processing
- graphical models
- bayesian networks
- belief networks
- associative memory
- memory requirements
- memory space
- database management systems
- inference process
- compressed data
- data processing
- memory size
- case based reasoning