I am a second year PhD student at KAUST AI Initiative, working with Jürgen Schmidhuber. My work focuses on creating scalable methods that improve generalization and sample efficiency in problem solving by utilizing compositionality.
I believe that low sample-efficiency and weak systematic generalization are the main contributing factors to (still relatively) low adaptation of AI. The holy grail of my research is to create a method that efficiently decomposes problems into subproblems similarly to humans.
In practice, it means that I am mostly interested in analyzing and improving transformer architecture, LLMs and (goal-conditioned) reinforcement learning algorithms.
Before starting my PhD I worked for over 3 years as a ML/Research Engineer. I studied simultaneously Mathematics (BSc+MSc) and Computer Science (BSc) at the University of Warsaw. I wrote my Master's thesis on improving BERT's mathematical skills under the supervision of Mateusz Malinowski and Henryk Michalewski
ACL 2021 (oral)
Piotr Piękos, Henryk Michalewski, Mateusz Malinowski
ICLR 2023 (Notable top 5%)
Michał Zawalski, Michał Tyrolski, Konrad Czechowski, Tomasz Odrzygóźdź. Damian Stachura. Piotr Piękos, Yuhuai Wu, Łukasz Kuciński, Piotr Miłoś
Mingchen Zhuge, Haozhe Liu, Francesco Faccio, Dylan R Ashley, Róbert Csordás, Anand Gopalakrishnan, Abdullah Hamdi, Hasan Abed Al Kader Hammoud, Vincent Herrmann, Kazuki Irie, Louis Kirsch, Bing Li, Guohao Li, Shuming Liu, Jinjie Mai, Piotr Piękos, Aditya Ramesh, Imanol Schlag, Weimin Shi, Aleksandar Stanić, Wenyi Wang, Yuhui Wang, Mengmeng Xu, Deng-Ping Fan, Bernard Ghanem, Jürgen Schmidhuber
June 2023 - Aug 2023
Working on sample efficient Goal-Conditioned RL and improving the speed of the Transformer.
Sept 2022 - Present
Supervised by Prof. Jürgen Schmidhuber. I am working on generalization, sample efficiency in context of language processing and reinforcement learning.
Jan 2022 - Jul 2022
Working on AdaSubS. Implementing experiments on the sokoban environment
Oct 2021 - Jan 2022
Researching neural search methods for improving search recommendations on the Allegro.pl ("Polish Amazon") platform
May 2020 - Sept 2021
After promising initial results I decided to focus fully on my Master's thesis, the effect of that is the Measuring and Improving BERT's Mathematical Abilities by Predicting the Order of Reasoning paper.
July 2017 - June 2020
In ITmagination I created DL/ML solutions to help other companies. Example projects are: Object detection system for evaluating influencers, recommender system or a system to help with estimating staff needs in physical shops.
piotrpiekos@gmail.com