9.4 PIU: A 248GOPS/W Stream-Based Processor for Irregular Probabilistic Inference Networks Using Precision-Scalable Posit Arithmetic in 28nm.
Nimish ShahLaura Isabel Galindez OlascoagaShirui ZhaoWannes MeertMarian VerhelstPublished in: ISSCC (2021)
Keyphrases
- probabilistic inference
- graphical models
- bayesian networks
- message passing
- conditional probabilities
- bayesian belief networks
- probabilistic reasoning
- efficient inference
- weighted model counting
- influence diagrams
- belief networks
- approximate inference
- bucket elimination
- variable elimination
- exact inference
- context specific independence
- elimination algorithm
- belief propagation
- learning algorithm
- parallel processing
- class labels