Matches in SemOpenAlex for { <https://semopenalex.org/work/W4280559323> ?p ?o ?g. }
Showing items 1 to 90 of
90
with 100 items per page.
- W4280559323 abstract "Bayesian machine learning is useful for applications that may make high-risk decisions with limited, noisy, or unlabeled data, as it provides great data efficiency and uncertainty estimation. Building on previous efforts, this work presents CoopMC, an algorithm-architecture co-optimization for developing more efficient MCMC-based Bayesian inference accelerators. CoopMC utilizes dynamic normalization (DyNorm), LUT-based exponential kernels (TableExp), and log-domain kernel fusion (LogFusion) to reduce computational precision and shrink ALU area by 7.5× without noticeable reduction in model performance. Also, a Tree-based Gibbs sampler (TreeSampler) improves hardware runtime from $mathcal{O}$(N) to $mathcal{O}$(log(N)), an 8.7× speedup, and yields 1.9× better area efficiency than the existing state-of-the-art Gibbs sampling architecture. These methods have been tested on 10 diverse workloads using 3 different types of Bayesian models, demonstrating applicability to many Bayesian algorithms. In an end-to-end case study, these optimizations achieve a 33% logic area reduction, 62% power reduction, and 1.53× speedup over previous state-of-the-art end-to-end MCMC accelerators." @default.
- W4280559323 created "2022-05-22" @default.
- W4280559323 creator A5001514720 @default.
- W4280559323 creator A5011630384 @default.
- W4280559323 creator A5026496503 @default.
- W4280559323 creator A5030969018 @default.
- W4280559323 creator A5043327132 @default.
- W4280559323 creator A5074229216 @default.
- W4280559323 date "2022-04-01" @default.
- W4280559323 modified "2023-10-01" @default.
- W4280559323 title "CoopMC: Algorithm-Architecture Co-Optimization for Markov Chain Monte Carlo Accelerators" @default.
- W4280559323 cites W1517555081 @default.
- W4280559323 cites W1593793857 @default.
- W4280559323 cites W1609010894 @default.
- W4280559323 cites W2001082470 @default.
- W4280559323 cites W2001975024 @default.
- W4280559323 cites W2107884096 @default.
- W4280559323 cites W2146807782 @default.
- W4280559323 cites W2150731624 @default.
- W4280559323 cites W2501634593 @default.
- W4280559323 cites W2928320209 @default.
- W4280559323 cites W2941433885 @default.
- W4280559323 cites W2949353087 @default.
- W4280559323 cites W2949665610 @default.
- W4280559323 cites W2988443989 @default.
- W4280559323 cites W3008515144 @default.
- W4280559323 cites W3011316788 @default.
- W4280559323 cites W3085257293 @default.
- W4280559323 cites W3088358415 @default.
- W4280559323 cites W3093720766 @default.
- W4280559323 cites W3125630704 @default.
- W4280559323 cites W3125994923 @default.
- W4280559323 cites W3133687618 @default.
- W4280559323 cites W3135697241 @default.
- W4280559323 cites W3155846605 @default.
- W4280559323 cites W4232344051 @default.
- W4280559323 cites W4254865403 @default.
- W4280559323 cites W3048404194 @default.
- W4280559323 doi "https://doi.org/10.1109/hpca53966.2022.00012" @default.
- W4280559323 hasPublicationYear "2022" @default.
- W4280559323 type Work @default.
- W4280559323 citedByCount "2" @default.
- W4280559323 countsByYear W42805593232023 @default.
- W4280559323 crossrefType "proceedings-article" @default.
- W4280559323 hasAuthorship W4280559323A5001514720 @default.
- W4280559323 hasAuthorship W4280559323A5011630384 @default.
- W4280559323 hasAuthorship W4280559323A5026496503 @default.
- W4280559323 hasAuthorship W4280559323A5030969018 @default.
- W4280559323 hasAuthorship W4280559323A5043327132 @default.
- W4280559323 hasAuthorship W4280559323A5074229216 @default.
- W4280559323 hasConcept C107673813 @default.
- W4280559323 hasConcept C111335779 @default.
- W4280559323 hasConcept C111350023 @default.
- W4280559323 hasConcept C11413529 @default.
- W4280559323 hasConcept C154945302 @default.
- W4280559323 hasConcept C158424031 @default.
- W4280559323 hasConcept C160234255 @default.
- W4280559323 hasConcept C173608175 @default.
- W4280559323 hasConcept C2524010 @default.
- W4280559323 hasConcept C33923547 @default.
- W4280559323 hasConcept C41008148 @default.
- W4280559323 hasConcept C68339613 @default.
- W4280559323 hasConceptScore W4280559323C107673813 @default.
- W4280559323 hasConceptScore W4280559323C111335779 @default.
- W4280559323 hasConceptScore W4280559323C111350023 @default.
- W4280559323 hasConceptScore W4280559323C11413529 @default.
- W4280559323 hasConceptScore W4280559323C154945302 @default.
- W4280559323 hasConceptScore W4280559323C158424031 @default.
- W4280559323 hasConceptScore W4280559323C160234255 @default.
- W4280559323 hasConceptScore W4280559323C173608175 @default.
- W4280559323 hasConceptScore W4280559323C2524010 @default.
- W4280559323 hasConceptScore W4280559323C33923547 @default.
- W4280559323 hasConceptScore W4280559323C41008148 @default.
- W4280559323 hasConceptScore W4280559323C68339613 @default.
- W4280559323 hasLocation W42805593231 @default.
- W4280559323 hasOpenAccess W4280559323 @default.
- W4280559323 hasPrimaryLocation W42805593231 @default.
- W4280559323 hasRelatedWork W2038626737 @default.
- W4280559323 hasRelatedWork W2075470739 @default.
- W4280559323 hasRelatedWork W2146501959 @default.
- W4280559323 hasRelatedWork W2172275095 @default.
- W4280559323 hasRelatedWork W2182500694 @default.
- W4280559323 hasRelatedWork W2255115219 @default.
- W4280559323 hasRelatedWork W2506941132 @default.
- W4280559323 hasRelatedWork W2950044886 @default.
- W4280559323 hasRelatedWork W3121470121 @default.
- W4280559323 hasRelatedWork W4280559323 @default.
- W4280559323 isParatext "false" @default.
- W4280559323 isRetracted "false" @default.
- W4280559323 workType "article" @default.