Skip to main content

Milos Nikolic

Personal photo - Milos Nikolic

Milos Nikolic

Departmental Lecturer

Leaving date: 30th August 2018

Interests

  • databases and large-scale data management systems
  • stream processing
  • incremental computation
  • query compilation

Projects

  • Incremental Maintenance for Analytics (2016-present, Oxford)
    • unified framework for maintaining a wide range of analytics
    • factorization for queries, updates, and result representation
  • Deep Analytics over Data Streams (2015-present, Oxford, Microsoft Research)
    • deep integration of relational and digital signal processing (DSP)
    • extensible framework for defining custom DSP operators
    • incremental computation in both offline and online analysis
  • DBToaster (2011-2016, EPFL)
    • compilation of database queries into streaming engines
    • 3-6 orders of magnitude faster than commercial systems
    • single-threaded (C++, Scala) and distributed (Spark) execution
  • LINVIEW (2013-2014, EPFL)
    • incremental computation of linear algebra using matrix factorization
    • local (Octave, MATLAB) and distributed (Spark) execution
  • Youtopia (2010-2011, EPFL, Cornell)
    • declarative data-driven coordination in social applications

Publications

  • Counting Triangles under Updates in Worst-Case Optimal Time [pdf]
    Ahmet Kara, Hung Ngo, Milos Nikolic, Dan Olteanu, and Haozhe Zhang, ICDT 2019.
  • Counting Triangles under Updates (short paper) [pdf]
    Ahmet Kara, Hung Ngo, Milos Nikolic, Dan Olteanu, and Haozhe Zhang,
    Alberto Mendelzon Workshop (AMW), 2018 (extended version arXiv 1804.02780).
  • Incremental View Maintenance with Triple Lock Factorization Benefits [pdf]
    Milos Nikolic, Dan Olteanu, SIGMOD 2018
  • Distributed Incremental View Maintenance
    Milos Nikolic, Encyclopedia of Big Data Technologies, Springer, 2018
  • Enabling Signal Processing over Data Streams [pdf, slides]
    Milos Nikolic, Badrish Chandramouli, and Jonathan Goldstein, SIGMOD 2017
  • Enabling Deep Analytics in Stream Processing Systems [pdf]
    Milos Nikolic, Badrish Chandramouli, and Jonathan Goldstein, BICOD 2017
  • Efficient Incremental Data Analysis [pdf]
    Milos Nikolic, PhD thesis, 2016
  • How to Win a Hot Dog Eating Contest: Distributed Incremental View Maintenance with Batch Updates [pdfslides]
    Milos Nikolic, Mohammad Dashti, and Christoph Koch, SIGMOD 2016
  • LINVIEW: Incremental View Maintenance for Complex Analytical Queries [pdfslidesvideo]
    Milos Nikolic, Mohammed ElSeidy, Christoph Koch, SIGMOD 2014
  • DBToaster: Higher-order Delta Processing for Dynamic, Frequently Fresh Views [pdf]
    Christoph Koch, Yanif Ahmad, Oliver Kennedy, Milos Nikolic, Andres Nötzli, Daniel Lupei, and Amir Shaikhha, VLDBJ 2014 (extended version)
  • DBToaster: Higher-order Delta Processing for Dynamic, Frequently Fresh Views [pdfslides]
    Yanif Ahmad, Oliver Kennedy, Christoph Koch, and Milos Nikolic, VLDB 2012
    Invited to the "Best of VLDB 2012" issue of the VLDB Journal.
  • Entangled transactions [pdfslides]
    Nitin Gupta, Milos Nikolic, Sudip Roy, Gabriel Bender, Lucja Kot, Johannes Gehrke, Christoph Koch, VLDB 2011

Teaching

     Courses

     Classes

     Practicals

Students

  • Haozhe Zhang (DPhil, Sept 2017 - present, co-advised with Prof. Dan Olteanu)
  • Theo Dickson (MSc, April - Sept 2018)
  • Canberk Koparal (3rd year CS, Oct 2017 - May 2018)
  • Alexandru Valeanu (3rd year CS, Oct 2017 - May 2018, co-advised with Prof. Dan Olteanu)
  • Xu Sun (MSc, April - Sept 2017)

Biography

Milos Nikolic is a departmental lecturer in the Department of Computer Science at the University of Oxford. Milos' research focuses on the design and implementation of data-intensive systems. His work studies the incremental computation of complex analytical queries, such as database queries and machine learning models, in local and distributed streaming environments using novel approaches to query optimization and query compilation. He received a Ph.D. in Computer Science from EPFL.

Activities