the website of
Gergely Neu

Gergely Neu

I am a machine learning researcher mainly interested in theoretical aspects of sequential decision making. I mainly work on online optimization, bandit problems, and reinforcement learning theory. I like to think about algorithms that come with performance guarantees both in terms of computational and statistical complexity, and are actually implementable on a computer.
Research Assistant Professor
AI group, DTIC
Universitat Pompeu Fabra


C/ Roc Boronat 138
08018 Barcelona

cell: (+34) 658 583 101

You might want to check out my
CV to get a better picture of what I've done professionally so far.

If you're not sure about how to pronounce my first name, please check out this page. The short story is that you basically have to forget about pronouncing the "l" (i.e., "ly" pronounces as "y"), and both G's are hard (as in the English word "get").
"The bandits be my weaknesses
His fault shall be my knife"
- Nico (1974)


  • I am hiring a PhD student to start in October 2023! Please find the details here. 2022. 11. 04.
  • Together with Lorenzo Rosasco, I will serve as program chair of COLT'23! 2022. 07. 17.
  • New paper about deriving generalization bounds using convex analysis to appear at COLT'22! 2022. 06. 30.
  • New paper about the generalization properties of SGD accepted to COLT'21! 2021. 07. 16.
  • I am hiring PhD students to start in October 2021! Please find the details here. I am also proud to have been recently appointed as an ELLIS Scholar, making me an official member of this amazing organization that I have been in awe of since its inception. 2020. 12. 03.
  • I have been awarded an ERC Starting Grant! Watch this space for job announcements! 2020. 09. 04.
  • New papers published: my PhD students Joan and Julia got their papers accepted to L4DC'20 and COLT'20! Plus another COLT paper with the inimitable Nikita Zhivotovskiy. Head over to my publications page to take a glimpse! 2020. 07. 06.
  • 2020 is a weird year. From my own (professional) perspective, I can describe it in two ways:
    • The year of cancelled events. While ALT'20 went down fantastically well, the other events I was involved in organizing this year ended up being not so lucky: DALI'20, MSL'20 and ITW'20 all got rescheduled to later dates, for obvious reasons. I am honestly very proud of the program I helped put together at these events, so I remain hopeful that they'll eventually take place in one form or another.
    • The year of reinforcement learning theory. I am involved in a few initiatives around RL theory this year. Perhaps the biggest news is that I'll be spending the Fall semester at the Simons Institute in Berkeley for the program on reinforcement learning theory (hoping that it won't end up being cancelled). In other very exciting news, together with Ciara and Csaba, I am organizing a virtual seminar series on RL theorya humble attempt to make up for the lack of physical meetings on this exciting topic this year. Another related (virtual) event that I'm looking forward to is the "Theory of RL" workshop at ICML'20, where I'm giving an invited talk. 2020. 04. 21.
  • I am the recipient of the first Bosch AI Young Researcher Award! 2019. 10. 31.
  • Heading to NeurIPS 2019! I have two papers in the main program with brilliant coauthors, plus I'll be serving as one of the two Social Chairs this year along with Maria Skoularidou. In completely unrelated news, you can now find me on Twitter! 2019. 09. 12.
  • Together with Aryeh Kontorovich, I will be serving as one of the program chairs of ALT'20! The conference is set to take place in San Diego, USA, February 8-11, co-located with ITA'20. Please consider contributing to this rapidly growing conference by submitting your work before the deadline on September 20, 2019! 2019. 06. 22.
  • Our paper with Wojciech on Bandit PCA is accepted to COLT'19! 2019. 04. 19.
  • I received a Google Faculty Research Award! 2019. 03. 14.
  • It's been a while since I've posted updates here, although a lot has been going on:
    • My PhD student Julia got her first paper accepted at ALT'19!
    • I went to NIPS'18 NeurIPS'18 where, besides enjoying the usual opportunity of being able to hang out with great minds, I had the pleasure of appearing on stage at the closing banquet, playing drums behind some amazing musicians. This was incredibly fun, even despite having to fight a massively adversarial drumkit :)
    • I gave a tutorial on reinforcement learning at the yearly Greek Stochastics meeting in Athens, to a truly fantastic audience. I even managed to get the award for the best talk of the workshop, which was very surprising to me given the high quality of the other talks I've attended! 2019. 01. 14.
  • Back at UPF after a busy summer and giving an invited talk at EWRL'18 in good old Lille! Next up is AWRL'18 in Beijing, China. In other news, I'll be serving on the program committee for COLT'19. 2018. 10. 08.
  • Busy summer: I'll be visiting Google Brain Zürich for 3 months starting in July! During the same period, I'll also be serving as an area chair at NIPS'18, so not much downtime for me this summer. 2018. 06. 05.
  • Paper on iterate averaging accepted at COLT'18! You can also catch me talking about this work at this workshop at UPF/BGSE and at ISMP'18. 2018. 05. 02.
  • Just back from an amazing learning thery workshop in Montreal! All talks including mine are now online. 2018. 04. 28.
  • I'm a recipient of the “la Caixa” Junior Leader fellowship. In related news, I'm looking to hire a postdoc, so please get in touch if interested in doing some (reinforcement) learning theory in Barcelona! 2018. 03. 22.
  • Together with Vicenç, I am co-organizing ALT'18 on Lanzarote (Canary Islands), collocated with AISTATS'18. It's a wonderful location and, as always, the conference has a fantastic PC, led by the chairs Mehryar Mohri and Karthik Sridharan. Please consider submitting a paper and attending! 2017. 10. 18.
  • One paper accepted at NIPS'17, thanks to my amazing coauthors: Nicolň, Claudio, and Gábor! 2017. 09. 16.
  • Just back from COLT'17 where I received an award for my superbly shambolic talk! (Update: video now online!!) I also gave another talk at the Learning Theory workshop at FOCM'17. 2017. 07. 17.
  • Two new papers out! One at COLT'17, another at ICML'17. I'm also set to give talks at a Deep Learning Workshop at Google Zürich and an ICML workshop on Picky Learners (update: I'm not going to ICML after all, Michal will replace me at the workshop!). 2017. 05. 24.
  • I just came back from a workshop on the "Theoretical Foundations for Learning from Easy Data" organized at the Lorentz Center of Leiden (Netherlands). I gave a tutorial on all the known ways of going beyond worst-case performance guarantees in bandit problems. Check out the slides here! 2016. 11. 13.
  • We are organizing EWRL 2016 on the weekend preceding NIPS 2016 in Barcelona with Vicenç and Csaba—please submit your recent work on Reinforcement Learning! 2016. 07. 17.
  • All these years of reviewing finally paying off! I have received an "outstanding reviewer award" from ICML'16 and a "blue ribbon" for outstanding reviewing from IJCAI'16. 2016. 06. 28.
  • New year, new papers! One at AISTATS 2016, another in JMLR. In other news, I'm attending the DALI'16 meeting again, this time invited to speak at the Reinforcement Learning workshop organized by Csaba and Yasin. 2016. 01. 10.
  • Going to NIPS 2015! I have one paper in the main program and am also invited to give a talk at the Learning Faster from Easy Data II workshop organized by Tim van Erven and Wouter Koolen.2015. 09. 18.
  • This Fall, I will be joining the AI group at the Universitat Pompeu Fabra (Barcelona) as a postdoc supported by the UPFellows grant. 2015. 05. 29.
  • Two papers accepted! One at COLT 2015, another at the IEEE Transactions on Information Theory. Preprints here. 2015. 05. 03.
  • I will be serving on the program committee for ALT 2015. It's a very good conference at a beautiful location, so please consider submitting your work—I'm looking forward to reviewing some good papers! 2015. 02. 07.
  • I got invited to the DALI'15 workshop on Learning Theory, organized by Nicolň Cesa-Bianchi and Ulrike von Luxburg. 2015. 01. 14.
  • I'm spending a few days at the University of Alberta with my former advisors Csaba and Andris before going to NIPS. Also giving a talk about one of my NIPS papers! 2014. 12. 03.
  • I gave a talk at the AI group at the Universitat Pompeu Fabra, Barcelona. Fun times, thanks for inviting me, Vicenç! 2014. 11. 14.
  • Talk at the LSOLDM 2014 workshop! I saw some cameras around so there's probably gonna be a video released soon. 2014. 09. 12.
  • Three papers accepted to NIPS 2014! Head over to my publications page to get a glimpse! 2014. 09. 09.
  • New website online! 2014. 06. 25.