Future Data Systems

ColBERT

ColBERT: state-of-the-art neural search (SIGIR'20, TACL'21, NeurIPS'21, NAACL'22, CIKM'22, ACL'23, EMNLP'23)

macrobase

MacroBase: A Search Engine for Fast Data

ARES

FrugalGPT

FrugalGPT: better quality and lower cost for LLM applications

FAST

End-to-end earthquake detection pipeline via efficient time series similarity search

gavel

Code for "Heterogenity-Aware Cluster Scheduling Policies for Deep Learning Workloads", which appeared at OSDI 2020

stk

sinkhorn-label-allocation

Sinkhorn Label Allocation is a label assignment method for semi-supervised self-training algorithms. The SLA algorithm is described in full in this ICML 2021 paper: https://arxiv.org/abs/2102.08622.

Willump

Willump Is a Low-Latency Useful Machine learning Platform.

Baleen

Baleen: Robust Multi-Hop Reasoning at Scale via Condensed Retrieval (NeurIPS'21)

Uniserve

A runtime implementation of data-parallel actors.

blazeit

Its BlazeIt because it's blazing fast

Megatron-LM

Ongoing research training transformer models at scale

ACORN

state-of-the-art search over vector embeddings and structured data (SIGMOD '24)

POP

Code for "Solving Large-Scale Granular Resource Allocation Problems Efficiently with POP", which appeared at SOSP 2021

omg

loa

Public code for LOA

tasti

Semantic Indexes for Machine Learning-based Queries over Unstructured Data (SIGMOD 2022)

cs245-as1

Student files for CS245 Programming Assignment 1: In-memory data layout

cs245-as2-public

InQuest

Accelerating Aggregation Queries on Unstructured Streams of Data

SparseJointShift

Model Performance Estimation and Explanation When Labels and A Few Features Shifts

sketchstore

Algorithms for compressing and merging large collections of sketches

smol

supg

parallel-lb-simulator

abae

Accelerating Approximate Aggregation Queries with Expensive Predicates (VLDB 21)

ezmode

An iterative algorithm for selecting rare events in large, unlabeled datasets

pop-ncflow

Code for POP (SOSP 2021) and NCFlow (NSDI 2021)

teavar