FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2565684518> ?p ?o ?g. }

• W2565684518 endingPage "99" @default.
• W2565684518 startingPage "86" @default.
• W2565684518 abstract "Predictive modeling using machine learning is an effective method for building compiler heuristics, but there is a shortage of benchmarks. Typical machine learning experiments outside of the compilation field train over thousands or millions of examples. In machine learning for compilers, however, there are typically only a few dozen common benchmarks available. This limits the quality of learned models, as they have very sparse training data for what are often high-dimensional feature spaces. What is needed is a way to generate an unbounded number of training programs that finely cover the feature space. At the same time the generated programs must be similar to the types of programs that human developers actually write, otherwise the learning will target the wrong parts of the feature space. We mine open source repositories for program fragments and apply deep learning techniques to automatically construct models for how humans write programs. We sample these models to generate an unbounded number of runnable training programs. The quality of the programs is such that even human developers struggle to distinguish our generated programs from hand-written code. We use our generator for OpenCL programs, CLgen, to automatically synthesize thousands of programs and show that learning over these improves the performance of a state of the art predictive model by 1.27×. In addition, the fine covering of the feature space automatically exposes weaknesses in the feature design which are invisible with the sparse training examples from existing benchmark suites. Correcting these weaknesses further increases performance by 4.30×." @default.
• W2565684518 created "2017-01-06" @default.
• W2565684518 creator A5034975204 @default.
• W2565684518 creator A5052267876 @default.
• W2565684518 creator A5053757097 @default.
• W2565684518 creator A5074058311 @default.
• W2565684518 date "2017-02-04" @default.
• W2565684518 modified "2023-09-24" @default.
• W2565684518 title "Synthesizing benchmarks for predictive modeling" @default.
• W2565684518 cites W109452506 @default.
• W2565684518 cites W1488163396 @default.
• W2565684518 cites W1512154564 @default.
• W2565684518 cites W1552624537 @default.
• W2565684518 cites W179875071 @default.
• W2565684518 cites W1810943226 @default.
• W2565684518 cites W1895577753 @default.
• W2565684518 cites W1966021031 @default.
• W2565684518 cites W1982020565 @default.
• W2565684518 cites W1987564528 @default.
• W2565684518 cites W1989562524 @default.
• W2565684518 cites W1997857295 @default.
• W2565684518 cites W2016357834 @default.
• W2565684518 cites W2018389835 @default.
• W2565684518 cites W2023925487 @default.
• W2565684518 cites W2033139628 @default.
• W2565684518 cites W2043223258 @default.
• W2565684518 cites W2053619615 @default.
• W2565684518 cites W2079735306 @default.
• W2565684518 cites W2080592089 @default.
• W2565684518 cites W2098456636 @default.
• W2565684518 cites W2101807486 @default.
• W2565684518 cites W2111334029 @default.
• W2565684518 cites W2112121929 @default.
• W2565684518 cites W2121717408 @default.
• W2565684518 cites W2128120785 @default.
• W2565684518 cites W2130942839 @default.
• W2565684518 cites W2132525863 @default.
• W2565684518 cites W2140609933 @default.
• W2565684518 cites W2140952846 @default.
• W2565684518 cites W2142958724 @default.
• W2565684518 cites W2143861926 @default.
• W2565684518 cites W2145004951 @default.
• W2565684518 cites W2148190602 @default.
• W2565684518 cites W2149234156 @default.
• W2565684518 cites W2156981320 @default.
• W2565684518 cites W2160945247 @default.
• W2565684518 cites W2166536280 @default.
• W2565684518 cites W2257123346 @default.
• W2565684518 cites W2273440736 @default.
• W2565684518 cites W2292573290 @default.
• W2565684518 cites W2293015622 @default.
• W2565684518 cites W2301631282 @default.
• W2565684518 cites W2326925005 @default.
• W2565684518 cites W2402268235 @default.
• W2565684518 cites W2402619042 @default.
• W2565684518 cites W2409856616 @default.
• W2565684518 cites W2413606589 @default.
• W2565684518 cites W2414287720 @default.
• W2565684518 cites W2417132441 @default.
• W2565684518 cites W2594987874 @default.
• W2565684518 cites W2919115771 @default.
• W2565684518 cites W2963773578 @default.
• W2565684518 cites W2964279072 @default.
• W2565684518 doi "https://doi.org/10.5555/3049832.3049843" @default.
• W2565684518 hasPublicationYear "2017" @default.
• W2565684518 type Work @default.
• W2565684518 sameAs 2565684518 @default.
• W2565684518 citedByCount "33" @default.
• W2565684518 countsByYear W25656845182017 @default.
• W2565684518 countsByYear W25656845182018 @default.
• W2565684518 countsByYear W25656845182019 @default.
• W2565684518 countsByYear W25656845182020 @default.
• W2565684518 countsByYear W25656845182021 @default.
• W2565684518 countsByYear W25656845182022 @default.
• W2565684518 crossrefType "proceedings-article" @default.
• W2565684518 hasAuthorship W2565684518A5034975204 @default.
• W2565684518 hasAuthorship W2565684518A5052267876 @default.
• W2565684518 hasAuthorship W2565684518A5053757097 @default.
• W2565684518 hasAuthorship W2565684518A5074058311 @default.
• W2565684518 hasConcept C108583219 @default.
• W2565684518 hasConcept C111472728 @default.
• W2565684518 hasConcept C111919701 @default.
• W2565684518 hasConcept C119857082 @default.
• W2565684518 hasConcept C121332964 @default.
• W2565684518 hasConcept C127705205 @default.
• W2565684518 hasConcept C13280743 @default.
• W2565684518 hasConcept C138885662 @default.
• W2565684518 hasConcept C154945302 @default.
• W2565684518 hasConcept C163258240 @default.
• W2565684518 hasConcept C169590947 @default.
• W2565684518 hasConcept C177264268 @default.
• W2565684518 hasConcept C185798385 @default.
• W2565684518 hasConcept C199360897 @default.
• W2565684518 hasConcept C202444582 @default.
• W2565684518 hasConcept C205649164 @default.
• W2565684518 hasConcept C2776401178 @default.
• W2565684518 hasConcept C2776760102 @default.
• W2565684518 hasConcept C2777062904 @default.

• next