Research Session 6 Query Processing with Best paper award talk

0:01:18 [Paper: I/O Efficient Approximate Nearest Neighbour Search based on Learned Functions]
0:02:33 Motivation
0:03:51 Problem Definition
0:04:21 Outline
0:04:53 Query Processing
0:06:29 Linear Model and Its Objective Function
0:10:20 Non-Linear Model: Neural Network
0:11:47 Non-Linear Model: Objective Function
0:12:29 Experiments
0:14:48 Conclusion
0:15:31 Q&A
0:19:35 Approach
0:20:16 [Paper: Efficient Query Processing with Optimistically Compressed Hash TabStrings in the USSR]
0:21:13 Shrinking Hash Tables
0:22:06 Overview
0:22:36 Compression
0:24:03 Optimistic Splitting
0:25:47 Strings in the USSR
0:29:11 Faster HashJoin Probe
0:29:59 Faster GroupBy on strings
0:31:57 Smaller Hash Tables in TPC-H
0:32:58 Faster Real-World Workloads (Public BI")
0:34:41 Summary
0:38:38 [Paper: UniKV: Toward High-Performance and Scalable KV Storage in Mixed Workloadsvia Unified Indexing]
0:39:18 Background
0:41:07 Related Work
0:42:18 Motivation
0:43:51 Main Idea
0:44:52 UniKV Design
0:50:21 Putting It All Together: Overall Architecture
0:51:35 Exp 2: Performance under Mixed Workloads
0:52:01 Exp 3: YCSB Performance
0:53:37 Thanks for your attention!
0:55:58 [Paper: Improved Correlated Sampling for Join Size Estimation]
0:56:08 Problem: Join Size Estimation
0:57:31 Correlated Sampling
0:58:47 Framework for Correlated Sampling
1:01:47 Limitations of Existing Approaches
1:03:36 Addressing the Limitations
1:04:13 Our approach: CSDL(Pvqv)
1:05:51 Our Approach: Case 3 and Case 4
1:07:31 Best Performing Variants
1:08:15 Experimental Results
1:10:31 Thank You
1:13:59 [MESSI: In-memory Data Series Indexing]
1:14:18 Data series
1:14:49 What do we want to do with them?
1:16:30 iSAX Summaries and Tree Index
1:16:55 Our Contributions
1:17:57 MESSI: In-memory data series index
1:19:23 MESSI Index creation - Stage 1: Load data to Index
1:20:28 MESSI Index creation - Stage 2: Grow Index
1:21:07 MESSI Query answering - Stage 3
1:23:41 Experimental Setup
1:24:51 Experiments
1:26:18 Conclusions