The Hungarian Machine Learning Days provide an opportunity for Hungarian machine learning researchers working in foreign institutions to meet each other and connect with those working in Hungary, including the younger generation and Ph.D. students.
During the 3-day informal summer event, internationally recognized researchers will present tutorials, young researchers will give short presentations, posters, and ideas, and there will be many informal joint programs.
Participation, lunch, and coffee breaks are free of charge but require registration.
The Sponsor of the Event:
All presentations are in English. For posters, please apply during registration in time. Late poster applications risk no space remaining available.
8:30 Arrival, coffee
9:00 András Kemler (Bosch): Welcome
9:15 Dániel Baráth (ETH Zürich): Cross-Modal Localization
9:45 András György (Google Deepmind): On the out-of-distribution generalization ability of transformers in reasoning
10:15 Miranda Christ (UC Berkeley)
10:45 Csaba Beleznai (Austrian Institute of Technology): 3D spatial perception and object manipulation
11:15 Róbert Busa-Fekete (Google Research): TBDFiltering: Sample-Efficient Tree-Based Data Filtering
Lunch
13:30 Poster Booster
15:00-16:15 Coffee and Poster session
8:30 Arrival, coffee
9:00 Péter Horváth (state Secretary for Research and Szeged Biological Research): Welcome and Hungarian AI Strategy
9:15 Gergely Flamich (Imperial College London): Data Compression with Diffusion Models
9:45 Spencer Hill (Queen's University): Rejection Sampling is Optimal for Relative Entropy Coding
10:15 Noémi Éltető (MPI Tübingen, Google Deepmind): ATLAS: Active Theory Learning for Automated Science
10:45 Dániel Barabási (FutureHouse San Francisco): Building an AI Scientist: Applications to Neuroscience
11:15 Ádám Zsolt Wagner (DeepMind): Exploring constructions in mathematics using AI tools
Lunch
13:30 Péter Horváth (state Secretary for Research and Szeged
Biological Research): Machine vision in microbiology (TBC)
14:00 András Balogh (Szeged): Verification of the Implicit World Model in a Generative Model via Adversarial Sequences
14:30 Balázs Meszéna (Wigner): TAVAE: A VAE with Adaptable Priors Explains Contextual Modulation in the Visual Cortex
15:00 Ambrus Tamás (SZTAKI): On Rate-Optimal Partitioning Classification from Observable and from Privatised Data
15:30 - 16:15 Coffee and Poster session
8:30 Arrival, coffee
9:00 Gergely Neu (UPF Barcelona): Generative modeling by value-driven transport
9:30 Fanghui Liu (Shanghai Jiao Tong University): From SLT4AI to AI4SLT: Generalization under scaling, and formalized empirical processes in Lean 4 (infrastructure)
10:00 Tamás Stenczel (Cambridge): Nested Sampling
10:30 Mihály Petreczky (CNRS Lille): PAC-Bayesian Bounds for Learning Partially Observed Stochastic Linear Time-Invariant State-Space Systems with Inputs and Sub-Gaussian Noise.
11:00 Dániel Rácz (SZTAKI): On the PAC Learnability of Stable Selective State-Space Models from a Single Time Series
11:30 Szilvia Újváry (Cambridge): What Should Small Language
Models Learn, and When Should They Call for Help?
Lunch
13:30 Domokos Kelen (Mastercard): Regression Uncertainty in Neural Networks
14:00 Miklós Máté Badó (Eötvös University Budapest): Clustered Influence Functions
14:30 Gergely Csáji (Eötvös University Budapest), Clustering via Hedonic Games: New Concepts and Algorithms
15:00 - 16:15 Coffee and Poster session
Registration: personalized invitation is sent to those on MILAB events list
Our location is Bosch Budapest Innovation Campus (Campus II),
Budapest, Robert Bosch utca 14, 1103
- right at the Kőbánya-Kispest Metro terminal.
Offline Website Creator