LOCAT: Low-Overhead Online Configuration Auto-Tuning Of Spark SQL Applications
How are companies using online scheduling systems to handle complicated internal scheduling? For example, many algorithms train using large batches consisting of thousands of training examples and if trained instance-by-instance, their performance degrades (Hayes et al., 2020). Batch learning is undesirable for embedded applications since… ادامه »LOCAT: Low-Overhead Online Configuration Auto-Tuning Of Spark SQL Applications