ML4Sys-and-Sys4ML

Logo

View the Project on GitHub luthramanisha/ML4Sys-and-Sys4ML

Second Workshop on ML4Sys and Sys4ML

Organizers

Date and Location

Program

Time   Speaker Title
09:00 - 09:20   Matthias and Andreas Opening Remarks

Systems for ML (Matthias)

Time   Speaker Title
09:20 - 09:40 Photo Maximilian Schüle (Uni Bamberg) Blue Elephants Inspecting Pandas: Inspection and Execution of Machine Learning Pipelines in SQL
09:40 - 10:00 Photo Maximilian Böther (ETH Zurich) Modyn: Data-Centric Machine Learning Pipeline Orchestration
10:00 - 10:30 Photo Stefan Grafberger (TU Berlin) mlwhatif: Data-Centric What-If Analysis for Native Machine Learning Pipelines

Coffee Break

Applications & Benchmarks (Matthias)

Time   Speaker Title
11:00 - 11:20 Photo Photo Stefan Hagedorn, Steffen Kläbe (Actian) Experiences of Implementing In-Database TPCx-AI
11:20 - 11:50 Photo Thaleia-Dimitra Doudali (IMDEA) Keep it Simple, Sustainable! When Is ML Necessary in Cloud Resource Management?
11:50 - 12:10 Photo Photo Jan-Micha Bodensohn, Liane Vogel (TU Darmstadt & DFKI) Large Language Models for Enterprise Data Engineering
12:10 - 12:30 Photo Akanksha Vijayvergiya (Uni Passau) Time-Series Analysis for Life-Science Data

Lunch Break

ML for Systems (Andreas)

Time   Speaker Title
13:15 - 13:45 Photo Silvan Reiner (Uni Konstanz) ML4DB: Don’t Learn What You Already Know
13:45 - 14:05 Photo Johannes Wehrstein (TU Darmstadt) GRACEFUL: A Learned Cost Estimator For UDFs
14:05 - 14:35 Photo Giorgio Vinciguerra (Uni Pisa) Learned Compression of Nonlinear Time Series With Random Access
14:35 - 15:05 Photo Immanuel Trummer (Cornell) CheaPT: Using Language Models Without Breaking the Bank

Workshop Description

The rapid advances in machine learning (ML) have significantly increased its adoption across various fields including data systems, both in academia and industry. These advancements have not only enhanced existing data systems but, in some cases, have completely transformed their internal components, leading to the development of an important field of “learned data system components” in the ML for Systems area. Similarly, a well-structured systems approach has also played a crucial role in advancing current ML techniques and systems, forming the basis for the Systems for the ML area.

The workshop aims to bring together renowned researchers both by means of keynotes and invited talks to discuss and debate intriguing topics at the core of the two focus areas. The open format for submitting abstracts from anything like a position paper to technical experiences allows for lively debates and discussions as well as fosters collaborations among participants to work on common topics. The speakers are expected to also share insights into their ongoing projects and the open research challenges they are currently addressing.

A key objective of this workshop is to promote interaction between all attendees—whether they are presenting authors, keynote or invited speakers, or other participants—to foster new collaborations.

Call for Abstracts

This year in the ML4Sys and Sys4ML workshop we are calling for one-page abstract submissions that can be anything from a position paper representing a novel vision about a viewpoint or a controversial topic, a summary of lessons learned by working on a research area, an open research problem, or technical experience working on a large system (academic or industry). The requirements for the abstracts is as follows

Please use the submission instructions of the main conference, LNI style (see https://gi.de/service/publikationen/lni) for the abstract. Additionally, you can submit a 2-minute video about the position paper, e.g., by demonstration or including a presentation about their work. Instead of a video, a maximum of 3 slides explaining the idea can also be submitted.

Topics of Interest

Submission Timeline

Where to submit: Submission Form

In case of questions, contact us at mlsystemsworkshop.btw@gmail.com