Computer aided proofs

Discrete geometry

Solving a Conjecture of Erdős

What proportion of the plane can be colored so that no two colored points are exactly unit distance apart? This geometric question was posed by Leo Moser in the early 1960s. Paul Erdős conjectured that this proportion cannot reach ¼. This conjecture has recently been proven by a collaboration between mathematicians Gergely Ambrus, Máté Matolcsi and members of our AI research group Adrián Csiszárik, Dániel Varga, and Pál Zsámboki.
The graph providing the upper bound

Rácz Dániel (SZTAKI)

Convergence and Generalization

Event
April 3, 2024
events
I will introduce some theoretical results related to the convergence and generalization capabilities of neural networks, in light of articles published at the NeurIPS 2023 conference in December. The first part of the presentation will summarize some important, earlier results, and then it will cover numerous articles based on these, which were presented at NeurIPS. The aim of the presentation is to provide a comprehensive overview of the current state of the field and the currently popular research directions.
April 3, 2024

Juhász András (Oxford)

Knot theory and AI

Event
March 20, 2024
events
I will overview some applications of supervised and reinforcement learning methods to knot theory that might be useful in other areas of mathematics.
March 20, 2024

András Kalapos (BME TMIT)

Önfelügyelt reprezentációtanulás komplex adatokon (in hungarian)

Event
December 6, 2023
events
A reprezentációtanulás, vagyis adatokból magasabb szintű ismeret, „tudás” kinyerése a mesterséges intelligencia kutatásának egy kiemelt kérdése. Új, tudományos és társadalmi hasznosulás szempontjából is jelentős terület az önfelügyelt reprezentáció tanulás, amely a felügyelt tanuláshoz képest jóval nagyobb léptékű adatbázisokon, széles körben alkalmazható nagy modellek tanítását teszi lehetővé. Nagy nyelvi modellek tanításának mára alapeleme az önfelügyelt tanítás, és gépi látásban is több sikeres megközelítést publikáltak. Kevesebb kutatás irányul azonban arra, hogy a vizuális önfelügyelt előtanítással kialakított hálók összetett gépi látási feladatokon (pl.
December 6, 2023

Márton Muntág (Rényi)

Mode Combinability: Exploring Convex Combinations of Permutation Aligned Models (in hungarian)

Event
November 22, 2023
events
Mode Combinability: Exploring Convex Combinations of Permutation Aligned Models Adrián Csiszárik, Melinda F. Kiss, Péter Kőrösi-Szabó, Márton Muntag, Gergely Papp, Dániel Varga As recently discovered (Ainsworth-Hayase-Srinivasa 2022 and others), two wide neural networks with identical network topology and trained on similar data can be permutation-aligned. That is, we can shuffle their neurons (channels) so that linearly interpolating between the two networks in parameter space becomes a meaningful operation (linear mode connectivity).
November 22, 2023

András Horváth (PPKE)

Targeted Adversarial Attacks on Generalizable Neural Radiance Fields

Event
November 15, 2023
events
Contemporary robotics relies heavily on addressing key challenges like odometry, localization, depth perception, semantic segmentation, the creation of new viewpoints, and navigation with precision and efficiency. Implicit neural representation techniques, notably Neural Radiance Fields (NeRFs) and Generalizable NeRFs (GeNeRFs), are increasingly employed to tackle these issues. This talk focuses on exposing certain critical, but subtle flaws inherent in GeNeRFs. Adversarial attacks, while not new to various machine learning frameworks, present a significant threat.
November 15, 2023
The Team
The Rényi AI team has a careful balance of theorists and machine learning practitioners. The practitioners in the team are well-versed and up-to-date in the extremely fast-paced world of deep learning, but at the same time, they can to contribute to foundational research. The theorists of the team are world-class experts in the highly abstract theoretical machinery, but at the same time, they do not shy away from running simulations. This balance creates an optimal environment for a free flow of ideas between theory and practice, thus, underpins pursuing the general goal to bridge the gap between mathematical theory and deep learning practice.
Balázs Szegedy
Mathematical Theory
Dávid Terjék
Generalization, Mathematical Theory
Adrián Csiszárik
Representation Learning, Foundations
Domonkos Czifra
NLP, Foundations
Diego González Sánchez
Generalization, Mathematical Theory
Botond Forrai
Modeling
Péter Kőrösi-Szabó
Modeling
Gábor Kovács
NLP
Ákos Matszangosz
Topology, Foundations
Márton Muntag
Time Series, NLP, Modeling
Melinda F. Kiss
Representation Learning
Gergely Papp
Modeling
Alex Olár
Foundations
Jakab Buda
NLP
Pál Zsámboki
Reinforcement Learning, Geometric Deep Learning
Zsolt Zombori
Formal Reasoning
Dániel Varga
Foundations, Computer aided proofs
Open Positions
The Rényi AI group is actively recruiting both theorists and practitioners.
Announcement: December 1, 2023
Deadline: rolling
applications
Rényi Institute is seeking Machine Learning Engineers to join our AI Research & Development team. Preferred Qualifications: • MLOps experience (especially in cloud environments) • Industry experience working on ML solutions
Announcement: December 1, 2023
Deadline: rolling
theory
The Artificial Intelligence group of Alfréd Rényi Institute of Mathematics (Budapest, Hungary) offers a postdoc position, fixed for one year with possible extension for a second year. The position is aimed at researchers who have obtained their PhD and are at earlier stages of their research career. The primary task associated with this position is to perform fundamental research. The broad research topics are as follows: 1. Neurosymbolic Reasoning, 2. Machine Learning and Automated Theorem Proving, and 3. Embedding Formal Mathematics into Vector Space
Announcement: December 1, 2023
Deadline: rolling
theory, applications
Rényi Institute is seeking Research Scientists to join our AI Research & Development team. You will have the privilege to work at a renowned academic institute and do what you love: do research and publish in the field of machine learning / deep learning.
Rényi AI - Building bridges between mathematics and artificial intelligence.