Posts on Sebastian Schelter

Action Editor at DMLR

Wed, 18 Oct 2023 00:00:00 +0200

I will join the newly formed Journal of Data-centric Machine Learning Research (DMLR) as an Action Editor

Keynote at BIFOLD Summer School

Wed, 30 Aug 2023 00:00:00 +0200

I will give a keynote on Directions Towards Resource-Efficient Machine Learning Systems in e-Commerce at the BIFOLD Weizenbaum Summer School on Artificial Intelligence and Ecological Sustainability in Berlin

Paper accepted at SIGMOD

Thu, 22 Jun 2023 00:00:00 +0200

I am part of the large group of contributors to Apache Flink, who won the ACM SIGMOD Systems Award 2023! Furthermore, Stefan, Shubha and me won an ACM SIGMOD Best Demo Runner Up Award for our demo on Proactively Screening Machine Learning Pipelines with ArgusEyes

Paper accepted at SIGMOD

Wed, 22 Feb 2023 00:00:00 +0100

We will present a paper on Automating and Optimizing Data-Centric What-If Analyses on Native Machine Learning Pipelines and a demo on Proactively Screening Machine Learning Pipelines with ArgusEyes at SIGMOD in Seattle

Vision paper accepted at ICDE

Thu, 02 Feb 2023 00:00:00 +0100

We will present a vision paper on the impact of data cleaning on the fairness of ML models at the new special track of ICDE, originating from our ongoing collaboration with the Center for Responsible AI at New York University

Presentation on ML-specific data debugging at CIDR

Wed, 18 Jan 2023 00:00:00 +0100

I gave a short presentation at CIDR on our ideas for using provenance to debug the data flowing through machine learning pipelines

Stefan to intern with the Gray Systems Lab

Tue, 06 Dec 2022 00:00:00 +0100

Stefan will intern with the Gray Systems Lab at Microsoft next year, working on data systems for the Azure cloud

Best paper runner up award the TRL workshop at NeurIPS

Sun, 04 Dec 2022 00:00:00 +0100

Congratulations to David for winning the best-paper runner-up award at the Table Representation Learning workshop at NeurIPS in New Orleans. NeurIPS made a recording of his presentation available

Oral presentation at the TRL workshop at NeurIPS

Thu, 10 Nov 2022 00:00:00 +0100

I am on parental leave until January 2023. I will be back for CIDR in Amsterdam

Oral presentation at the TRL workshop at NeurIPS

Mon, 07 Nov 2022 00:00:00 +0100

David is going to present our work on Parameter-Efficient Automation of Data Wrangling Tasks with Prefix-Tuning at the Table Representation Learning workshop at NeurIPS in New Orleans

kNN paper accepted at WSDM

Tue, 18 Oct 2022 00:00:00 +0200

Mozhdeh’s paper on A Personalized Neighborhood-based Model for Within-basket Recommendation in Grocery Shopping has been accepted at WSDM. Another piece of evidence that kNN models provide state-of-the-art prediction quality and achieve low latency inference for sequential recommendation tasks

Zeyu started as a PhD Student

Mon, 10 Oct 2022 00:00:00 +0200

Zeyu Zhang starts as a new PhD student, working on responsible data management and natural language processing for mental health research. He will be jointly supervised with Iacer Calixto from the Amsterdam Medical Center

Invited talk at Megagon Labs

Thu, 18 Aug 2022 00:00:00 +0200

I gave an invited talk on Data Provenance as a Foundation for AI Governance at Megagon Labs

Interview published by the Informatics Institute

Fri, 22 Jul 2022 00:00:00 +0200

The Informatics Institute of the UvA has published an interview with me on making data management responsible

Article on Responsible Data Management published in the Communications of the ACM

Sat, 21 May 2022 00:00:00 +0200

Julia Stoyanovich, Serge Abiteboul, Bill Howe, H.V. Jagadish and me have published an article on Responsible Data Management in the Communications of the ACM. We discuss perspectives on the role and responsibility of the data management research community in designing, developing, using and overseeing automated decision making systems

Invited to the Career Panel at the PhD Symposium at ICDE'22

Fri, 06 May 2022 00:00:00 +0200

I have been invited to a panel on Career Paths after Ph.D. – Perspectives from Senior and Junior Researchers at the PhD Symposium of ICDE’22

Special issue on 'Special Issue on Directions Towards GDPR-Compliant Data Systems and Applications' of the IEEE Data Engineering Bulletin

Wed, 20 Apr 2022 00:00:00 +0200

I have been the editor for a Special Issue on Directions Towards GDPR-Compliant Data Systems and Applications of the IEEE Data Engineering Bulletin

PhD opening in responsible data management and NLP

Wed, 09 Mar 2022 00:00:00 +0100

I have a fully funded PhD opening on responsible data management and NLP for mental health data, jointly supervised with Iacer Calixto. This is a unique chance to work on data integration, GDPR compliance and neural natural language processing at the University of Amsterdam

Invited as an Action Editor for the MLOSS track of the Journal of Machine Learning Research

Tue, 08 Mar 2022 00:00:00 +0100

I have been invited to serve as an Action Editor for the Open Source Software Track of the Journal for Machine Learning Research

Serenade is available as an Open Source library

Thu, 17 Feb 2022 00:00:00 +0100

We released an open source version of our recently developed session-based recommender system ‘Serenade’, which is in production usage at bol.com

Talk on Session-Based Recommendations at ECIR

Fri, 11 Feb 2022 00:00:00 +0100

I will discuss the social, economic and ecological sustainability of AI as a panelist at an upcoming online event in Berlin organised by AlgorithmWatch

Talk on Session-Based Recommendations at ECIR

Thu, 10 Feb 2022 00:00:00 +0100

Barrie will present our work on scalable session-based recommendations at bol.com as part of the industry day at the European Conference on Information Retrieval (ECIR)

Paper on mlinspect available at VLDBJ

Mon, 31 Jan 2022 00:00:00 +0100

Stefan’s paper on Data distribution debugging in machine learning pipelines is now published online at VLDB Journal, as part of the special issue on responsible data science for data management

Paper on Serenade accepted at SIGMOD

Tue, 14 Dec 2021 00:00:00 +0100

Barrie and Olivier’s paper on Serenade - Low-Latency Session-Based Recommendation in e-Commerce at Scale was accepted to SIGMOD 2022! This work details the novel recommender system that we built together with our industry partner bol.com

AI for Retail Lab on Albert Heijn's tech blog

Fri, 03 Dec 2021 00:00:00 +0100

Stefan’s paper on Data Distribution Debugging in Machine Learning Pipelines was accepted for the VLDB Journal Special Issue on Data Science for Responsible Data Management! This work is part of our ongoing collaboration with Julia Stoyanovich from NYU

Two abstracts accepted at the Dutch-Belgian Database Day

Wed, 24 Nov 2021 00:00:00 +0100

Stefan, Barrie, Olivier and Shubha have two abstracts accepted at the Dutch-Belgian Database Day. We will present our recent work on screening ML pipelines and building low-latency recommender systems

Talk on Data Provenance at the DataCentric AI workshop

Thu, 18 Nov 2021 00:00:00 +0100

My student Stefan and me gave a talk on Data Provenance as the Foundation for Data Governance at the virtual Data Centric AI workshop organised by Stanford and ETH Zuerich

AI for Retail Lab on Albert Heijn's tech blog

Wed, 17 Nov 2021 00:00:00 +0100

Our student Mozhdeh published a post about her research on ‘Understanding customer behavior in multi-channel retail’ on Albert Heijn’s tech blog

AI for Retail Lab on recommendations – podcast

Thu, 04 Nov 2021 00:00:00 +0100

Barrie Kersbergen and me were interviewed by bol.com’s techlab podcast this week. We discussed our recent work on rebuilding a production recommendation system for bol.com, as well as the collaboration between bol.com and the University of Amsterdam

Interdisciplinary PhD position

Fri, 22 Oct 2021 00:00:00 +0200

A proposal by Iacer Calixto and me for an interdisciplinary PhD position at the intersection of mental health research, responsible data management and natural language processing will be funded by the Data Science Centre of the University of Amsterdam

Abstract accepted at CIDR

Thu, 21 Oct 2021 00:00:00 +0200

Our abstract on Screening Native Machine Learning Pipelines with ArgusEyes will be presented at CIDR’22. This is joint work with Ce Zhang from ETH Zürich

Paper accepted at WSDM

Mon, 11 Oct 2021 00:00:00 +0200

Our work on Understanding and Mitigating the Effect of Outliers in Fair Ranking led by Arezoo Sarvi will be presented at WSDM’22

Presentation at the ORSUM workshop

Thu, 07 Oct 2021 00:00:00 +0200

Benjamin has been presenting our work on machine unlearning for next-basket recommendations at the workshop on Online Recommender Systems at RecSys’21

Roundtable on Systems for ML at VLDB'21

Wed, 18 Aug 2021 00:00:00 +0200

I have been a participant in the roundtable discussion on Systems for ML at VLDB’21

Paper accepted at CIKM

Mon, 09 Aug 2021 00:00:00 +0200

Our work on Understanding Multi-channel Customer Behavior in Retail led by Mozhdeh Ariannezhad will be presented at CIKM’21

Industry track chair at EDBT 2022

Wed, 30 Jun 2021 00:00:00 +0200

I have been invited to chair the industry and applications track at EDBT 2022 together with Fatma Ozcan

Benjamin Longxiang Wang wins our ML competition

Mon, 21 Jun 2021 00:00:00 +0200

Congratulations to Benjamin Longxiang Wang for winning our bol.com Unwanted Product Detection Challenge 2021 with a test set accuracy of 88.6%!

Benjamin Longxiang Wang wins our ML competition

Mon, 21 Jun 2021 00:00:00 +0200

Two papers on DuckDQ: Data Quality Assertions for Machine Learning Pipelines and Towards Efficient Machine Unlearning via Incremental View Maintenance accepted at the workshop on Challenges in Deploying and Monitoring ML Systems at ICML’21

New recommender system in production at bol.com

Mon, 21 Jun 2021 00:00:00 +0200

My students Barrie Kersbergen, Olivier Sprangers and me have built a new recommendation system for the product detail pages of bol.com. It’s written in Rust, it’s super fast, and it is in production since today!

Lunch at ICAI: Data Management for Machine Learning

Tue, 15 Jun 2021 00:00:00 +0200

We will present our lab and its ongoing research at the ICAI lunch meetup on Data Management for Machine Learning on Thursday the 17th of June

Interview on data management for machine learning

Thu, 10 Jun 2021 00:00:00 +0200

I have been interviewed by the Innovation Center for Artificial Intelligence on the role of data management for machine learning

Academic Keynote at DEEM

Tue, 01 Jun 2021 00:00:00 +0200

I will give the adademic keynote on “Connecting ML Applications to Real World Data” at the workshop on Data Management for End-to-End Machine Learning (DEEM) in conjunction with SIGMOD 2021

PGBM at KDD

Fri, 28 May 2021 00:00:00 +0200

We will present a paper on probabilistic gradient boosting machines for large-scale probabilistic regression at KDD 2021

Two papers at SIGMOD

Thu, 20 May 2021 00:00:00 +0200

We will present a paper on maintaining randomised trees for low-latency machine unlearning and a demo on a data distribution debugger for machine learning pipelines at SIGMOD 2021

Paper at ICDE

Sun, 18 Apr 2021 00:00:00 +0200

Barrie will present our work on learnings from a real world recommender system at one of Europe’s largest e-commerce platforms at ICDE 2021

Editing for the IEEE Data Engineering Bulletin

Tue, 30 Mar 2021 00:00:00 +0200

I have been the editor for the special issue on “Data validation for machine learning models and applications” of the IEEE Data Engineering Bulletin

Associate Editor for PVLDB

Mon, 01 Mar 2021 00:00:00 +0100

I will serve as an Associate Editor for the Scalable Data Science track of PVLDB Volume 15

Stefan and Barrie joining!

Mon, 01 Feb 2021 00:00:00 +0100

Stefan Grafberger and Barrie Kersbergen joined our lab as Ph.D. students to work on data validation for ML pipelines and scalable recommender systems!

mlinspect at CIDR'21

Fri, 15 Jan 2021 00:00:00 +0100

We presented our work on Lightweight Inspection of Data Preprocessing in Native Machine Learning Pipelines at CIDR. Checkout the video of our talk and our example notebook!

AIRLab at the Dutch Belgian Database Day

Fri, 11 Dec 2020 00:00:00 +0100

Stefan, Mozhdeh, Arezoo and Barrie presented some of our recent work at the Dutch-Belgian Database Day, and the Dutch-Belgian Information Retrieval Workshop

Paper on Apache Mahout accepted at JMLR

Sun, 15 Nov 2020 00:00:00 +0100

Our work on scalable data validation has been integrated into the new Amazon SageMaker Model Monitor service for concept drift detection. Here is the announcement at the AWS Reinvent conference:

Keynote at the ORSUM workshop at ACM RecSys

Fri, 25 Sep 2020 00:00:00 +0200

I gave the keynote at the workshop on Online Recommender Systems at ACM RecSys, check out the recording

FairPrep mentioned in a VLDB Keynote

Fri, 04 Sep 2020 00:00:00 +0200

Julia Stoyanovich mentioned our work on the FairPrep framework in her keynote at VLDB 2020

Talk on Deequ at CWI

Thu, 25 Jun 2020 00:00:00 +0200

I gave a talk on Unit Tests for Data with Deequ at the database group of CWI, checkout the video recording

Paper on Apache Mahout accepted at JMLR

Sat, 20 Jun 2020 00:00:00 +0200

We presented a paper on Learning to Validate the Predictions of Black Box Classifiers on Unseen Data, an industry paper on Elastic Machine Learning Algorithms in Amazon SageMaker at SIGMOD’20, as well as initial ideas on the Fairness-Aware Instrumentation of Preprocessing Pipelines for Machine Learning in the Human-in-the-Loop-Data-Analytics workshop

DEEM at SIGMOD'20

Sun, 14 Jun 2020 00:00:00 +0200

We ran the 4th edition of the Workshop on Data Management for End-to-End Machine Learning (DEEM) with more than 120 attendees via Zoom this year. Checkout the videos of the presentations and invited talks on our website

Paper on Apache Mahout accepted at JMLR

Mon, 01 Jun 2020 00:00:00 +0200

Our paper summarizing the Apache Mahout project called Apache Mahout: Machine Learning on Distributed Dataflow Systems has been accepted to the open source track of JMLR

Presentation at CIDR'20

Thu, 02 Apr 2020 00:00:00 +0200

We presented a short paper on FairPrep: Promoting Data to a First-Class Citizen in Studies on Fairness-Enhancing Interventions and a full paper on Zooming Out on an Evolving Graph at EDBT’20

Presentation at CIDR'20

Wed, 15 Jan 2020 00:00:00 +0100

We presented our work on “Amnesia” - A Selection of Machine Learning Models That Can Forget User Data Very Fast at CIDR’20

Paper on Missing Value Imputation accepted JMLR

Fri, 01 Nov 2019 00:00:00 +0100

Our paper on Datawig: Missing Value Imputation for Tables has been accepted at the open source track of the Journal of Machine Learning Research (JMLR)