Tim van Bremen

I am a Research Fellow in the School of Computing at the National University of Singapore, hosted by Kuldeep S. Meel.

I am on the academic job market for positions starting in 2024.
CV | Research Statement | Teaching Statement
Please feel free to get in touch with me to discuss potential opportunities.

My current research interests lie in statistical-symbolic artificial intelligence and data management under uncertainty. More specifically, together with a number of collaborators, I have worked in the following areas:

  • Probabilistic Inference in Statistical-Symbolic Models
    I have worked on developing scalable algorithms [UAI-21, AAAI-21, IJCAI-21, IJCAI-20] for exact and approximate inference in statistical-symbolic models, such as Markov logic networks and probabilistic logic programs in languages like ProbLog. I have also contributed to mapping the boundary between tractable and intractable models for inference [KR-21]. In addition, I have investigated applications of efficient inference algorithms in combinatorics and probabilistic networks [ILP-21, PROBPROG-21], and have further worked on formalizing and studying notions of efficient sampling in statistical-symbolic models [AAAI-22].
  • Query Evaluation on Probabilistic Databases
    More recently, I have been doing research in tuple-independent probabilistic databases, which extend classic relational databases to incorporate uncertainty. I have sought to understand when queries on such databases can and cannot be approximately evaluated (with rigorous (ε,δ)-style guarantees) in polynomial time [ICDT-24, PODS-23]. I have also worked on developing practical approaches for ontology-mediated querying of probabilistic data [CIKM-19].

Before coming to Singapore, I graduated with a PhD in Computer Science in June 2022 at KU Leuven, supervised by Luc De Raedt and Ondřej Kuželka. Prior to that, I received an MS in Computer Science from National Taiwan University in 2017, and even before that I got a BEng in Mathematics and Computer Science from Imperial College London in 2015.

Together with a few colleagues, I helped to organise the AlgoTheory seminar (2022/23) at NUS.

News

Publications (see also dblp)

Conferences:

  • Conjunctive Queries on Probabilistic Graphs: The Limits of Approximability (pdf, arXiv, doi, slides)
    Antoine Amarilli*, Timothy van Bremen*, and Kuldeep S. Meel*
    International Conference on Database Theory (ICDT) 2024
    (Invited to a special issue of the journal Logical Methods in Computer Science)
  • Probabilistic Query Evaluation: The Combined FPRAS Landscape (pdf, doi)
    Timothy van Bremen* and Kuldeep S. Meel*
    ACM Symposium on Principles of Database Systems (PODS) 2023
  • Domain-Lifted Sampling for Universal Two-Variable Logic and Extensions (doi)
    Yuanhong Wang, Timothy van Bremen, Yuyi Wang, and Ondřej Kuželka
    AAAI Conference on Artificial Intelligence (AAAI) 2022
  • Automatic Conjecturing of P-Recursions Using Lifted Inference (pdf, doi)
    Jáchym Barvínek, Timothy van Bremen, Yuyi Wang, Filip Železný, and Ondřej Kuželka
    International Conference on Inductive Logic Programming (ILP) 2021
  • Lifted Inference with Tree Axioms (doi)
    Timothy van Bremen and Ondřej Kuželka
    International Conference on Principles of Knowledge Representation and Reasoning (KR) 2021
    (Marco Cadoli Best Student Paper Award Runner-up)
  • Faster Lifting for Two-Variable Logic Using Cell Graphs (url)
    Timothy van Bremen and Ondřej Kuželka
    Conference on Uncertainty in Artificial Intelligence (UAI) 2021
  • Symmetric Component Caching for Model Counting on Combinatorial Instances (doi)
    Timothy van Bremen*, Vincent Derkinderen*, Shubham Sharma*, Subhajit Roy, and Kuldeep S. Meel
    AAAI Conference on Artificial Intelligence (AAAI) 2021
  • Fast Algorithms for Relational Marginal Polytopes (doi)
    Yuanhong Wang, Timothy van Bremen, Yuyi Wang, Juhua Pu, and Ondřej Kuželka
    International Joint Conference on Artificial Intelligence (IJCAI) 2021
  • From Probabilistic NetKAT to ProbLog: New Algorithms for Inference and Learning in Probabilistic Networks (pdf)
    Birthe van den Berg*, Timothy van Bremen*, Vincent Derkinderen*, Angelika Kimmig, Tom Schrijvers, and Luc De Raedt
    International Conference on Probabilistic Programming (PROBPROG) 2021
  • Approximate Weighted First-Order Model Counting: Exploiting Fast Approximate Model Counters and Symmetry (doi)
    Timothy van Bremen and Ondřej Kuželka
    International Joint Conference on Artificial Intelligence (IJCAI) 2020
  • Ontology-mediated Queries over Probabilistic Data via Probabilistic Logic Programming (pdf, doi)
    Timothy van Bremen*, Anton Dries*, and Jean Christoph Jung*
    ACM International Conference on Information and Knowledge Management (CIKM) 2019

Journals:

  • Lifted Inference with Tree Axioms (pdf, doi)
    Timothy van Bremen and Ondřej Kuželka
    Artificial Intelligence (AIJ) 2023
    (Journal version of the KR 2021 paper)
  • onto2problog: A Probabilistic Ontology-mediated Querying System using Probabilistic Logic Programming (doi)
    Timothy van Bremen*, Anton Dries*, and Jean Christoph Jung*
    KI - Künstliche Intelligenz (German Journal of Artificial Intelligence) 2020
    (This is a "systems description" version of the CIKM 2019 paper)

Peer-reviewed workshops and domestic conferences: (show)

  • Approximate Weighted First-Order Model Counting: Exploiting Fast Approximate Model Counters and Symmetry (arXiv)
    Timothy van Bremen and Ondřej Kuželka
    International Workshop on Statistical Relational AI (StarAI) at AAAI 2020
    (Preliminary version of a paper later published at IJCAI 2020)
  • Efficient Cardinality Constraints in ProbLog
    Timothy van Bremen, Wannes Meert, and Luc De Raedt
    Benelux Conference on Artificial Intelligence (BNAIC) 2018

(* = alphabetical order or equal contribution)

Where possible, I try to keep the PDF versions of papers linked above up-to-date with corrections to any errors appearing in the published version.

Software

  • FastWFOMC
    A tool for computing the weighted first-order model count of a two-variable sentence in a domain-lifted way.
  • onto2problog
    A tool for ontology-mediated query answering over probabilistic data for ontologies formulated in OWL 2 EL.
  • SymGANAK
    A probabilistic exact model counter with support for symmetric component caching.