Name: Immanuel
Type: User
Company: Cornell University
Bio: My name is Immanuel, I'm assistant professor at Cornell University. I mostly publish code that relates to my research which is about database and data analysis.
Twitter: immanueltrummer
Blog: http://www.itrummer.org
Immanuel's Projects
A list dedicated to products, demos and articles related to 🤖 OpenAI's Codex.
BALM - A framework for Batch Analysis with Language Models
CodexDB generates code for SQL query processing via OpenAI's GPT-3 Codex model.
Website of the Cornell Database Seminar
Web site summarizing our work on voice query interfaces
This project aims at predicting correlated column pairs in data tables by analyzing column names via large language models.
Resources to learn data processing with GPT and other language models
DB-BERT tunes database systems for optimal performance, using tuning hints mined from text.
This repository contains tools for finding query plans that produce the least number of join result tuples (the so-called "Cout" metric). Those tools are not suitable for query optimization at run time - instead, they can be used for offline analysis to assess the quality of query optimizers.
Personal page of Immanuel Trummer.
Implements theta-joins with free-text join predicates using large language models.
Material for my VLDB'22 and BTW'23 tutorials on application for language models in data management
Neural Relation Understanding: neural cardinality estimators for tabular data
Mine data for patterns described in natural language
Algorithms for mapping database optimization problems to quantum hardware (D-Wave).
Contains implementations of various algorithms for database query optimization. The algorithms cover multiple problem variants (e.g., traditional and multi-objective query optimization) and offer different tradeoffs between optimality guarantees and optimization time (e.g., fast randomized algorithms, approximation schemes, exhaustive algorithms, parallel algorithms, ...).
Compress database schemata to reduce cost for LLM processing
Slate is a Jekyll theme for GitHub Pages
TUPAC: Tuning with Partial Configurations
Frontend for virtual TA