Learned Data Systems - 分类 - Adbean's Blog

Paper Reading: CatSQL: Towards Real World Natural Language to SQL Applications [VLDB 23]

adbeanx@outlook.com (Adbean) — Sun, 14 Apr 2024 14:19:15 -0400

CatSQL: Towards Real World Natural Language to SQL Applications

NL2SQL, text to SQL 是很有趣的方向。有 DL 方法也有现在的 LLM 微调。

Paper Reading: The Case for a Learned Sorting Algorithm [SIGMOD 2020]

adbeanx@outlook.com (Adbean) — Fri, 12 Apr 2024 20:27:34 -0400

The Case for a Learned Sorting Algorithm

除了 Query Optimization, Index, Tunning, ML 还可以用在 Database 其他方面，比如排序？

Paper Reading: An Inquiry into Machine Learning-based Automatic Configuration Tuning Services on Real-World Database Management Systems

adbeanx@outlook.com (Adbean) — Sun, 07 Apr 2024 14:22:23 -0400

An Inquiry into Machine Learning-based Automatic Configuration Tuning Services on Real-World Database Management Systems

CMU 的对于 DBMS 自动调优的论文，采用了 ML 机器学习方法，是 Ottertune 的论文。

Paper Reading: MB2: Decomposed Behavior Modeling for Self-Driving Database Management Systems

adbeanx@outlook.com (Adbean) — Fri, 05 Apr 2024 09:50:55 -0400

MB2: Decomposed Behavior Modeling for Self-Driving Database Management Systems

self-driving database management systems

Paper Reading: Bao: Making Learned Query Optimization Practical [SIGMOD 21]

adbeanx@outlook.com (Adbean) — Sun, 17 Mar 2024 00:14:24 -0400

Bao: Making Learned Query Optimization Practical

MLDB + query optimization

Paper Reading: Zero-Shot Cost Models for Out-of-the-box Learned Cost Prediction [VLDB 2022]

adbeanx@outlook.com (Adbean) — Sat, 16 Mar 2024 09:58:30 -0400

Zero-Shot Cost Models for Out-of-the-box Learned Cost Prediction [VLDB 2022]

Abstract

本文介绍了 zero-shot cost model，该模型可以使学习的成本估算能够 generalizes to unseen databases。与最 state-of-the-art 的工作负载驱动的方法相反，这些方法不必在每个新数据基础上执行大量 training queries ，zero-shot cost models thus allow to instantiate a learned cost model out-of-the-box without expensive training data collection。为了 zero-shot cost models，本文提出 a new learning paradigm based on pre-trained cost models。As core contributions to support the transfer of such a pre-trained cost model to unseen databases, we introduce a new model architecture and representation technique for encoding query workloads as input to those models. As we will show in our evaluation, zero-shot cost estimation can provide more accurate cost estimates than state-of-the-art models for a wide range of (real-world) databases without requiring any query executions on unseen databases. Furthermore, we show that zero-shot cost models can be used in a few-shot mode that further improves their quality by retraining them just with a small number of additional training queries on the unseen database.

Paper Reading: Spade Synthesizing Assertions for Large Language Model Pipelines

adbeanx@outlook.com (Adbean) — Fri, 01 Mar 2024 20:26:39 -0500

SPADE: Synthesizing Assertions for Large Language Model Pipelines

Synthesizing Assertions

Paper Reading: SEED Domain-Specific Data Curation With Large Language Models

adbeanx@outlook.com (Adbean) — Fri, 01 Mar 2024 10:42:53 -0500

SEED: Domain-Specific Data Curation With Large Language Models

使用大型语言模型的领域特定数据管理

Paper Reading: DB-BERT: a Database Tuning Tool that “Reads the Manual”

adbeanx@outlook.com (Adbean) — Sun, 04 Feb 2024 12:38:04 -0500

DB-BERT: a databse tuning tool that reads the manual

DB-BERT，一个读了手册的数据库调优工具