Talk to IDEAS NCBR scientists and see what pioneering research in the field of AI looks like!
Next Tuesday, December 5, visit IDEAS NCBR to see posters from conferences in which our researchers have participated this year or will soon participate:
- NeurIPS 2023 – Neural Information Processing Systems
- ENLSP 2023 – Efficient Natural Language and Speech Processing (warsztaty w ramach NeurIPS)
- ICCV 2023 – International Conference on Computer Vision
- WACV 2024 – Winter Conference on Applications of Computer Vision
Date: December 5, 2023, 12 am – 2 pm
Place: IDEAS NCBR, Warsaw, 69 Chmielna Street
How to join: send us an email at rsvp@ideas-ncbr.pl
Number of places is limited.
Posters come from both the main track and workshops of the mentioned conferences. Below is a list of publications (only the IDEAS NCBR affiliation was noted).
NeurIPS 2023:
Bucks for Buckets (B4B): Active Defenses Against Stealing Encoders
https://arxiv.org/abs/2310.08571
Jan Dubiński (IDEAS NCBR), Stanisław Pawlak, Franziska Boenisch, Tomasz Trzciński (IDEAS NCBR), Adam Dziedzic
The Tunnel Effect: Building Data Representations in Deep Neural Networks
https://arxiv.org/abs/2305.19753
Wojciech Masarczyk, Mateusz Ostaszewski, Ehsan Imani, Razvan Pascanu, Piotr Miłoś (IDEAS NCBR), Tomasz Trzciński (IDEAS NCBR)
Focused Transformer: Contrastive Training for Context Scaling
https://arxiv.org/abs/2307.03170
Szymon Tworkowski (IDEAS NCBR), Konrad Staniszewski (IDEAS NCBR), Mikołaj Pacek (IDEAS NCBR), Yuhuai Wu, Henryk Michalewski, Piotr Miłoś (IDEAS NCBR)
Trust Your ∇: Gradient-based Intervention Targeting for Causal Discovery
https://arxiv.org/abs/2211.13715
Mateusz Olko (IDEAS NCBR), Michał Zając, Aleksandra Nowak, Nino Scherrer, Yashas Annadani, Stefan Bauer, Łukasz Kuciński (IDEAS NCBR), Piotr Miłoś (IDEAS NCBR)
Fantastic Weights and How to Find Them: Where to Prune in Dynamic Sparse Training
https://arxiv.org/abs/2306.12230
Aleksandra Nowak (w IDEAS NCBR podczas prowadzenia badań), Bram Grooten, Decebal Constantin Mocanu, Jacek Tabor
NeurIPS 2023 workshops:
Revisiting Supervision for Continual Representation Learning
https://arxiv.org/abs/2311.13321
Daniel Marczak, Sebastian Cygert, Tomasz Trzciński, Bartłomiej Twardowski (wszyscy z IDEAS NCBR)
Augmentation-aware Self-supervised Learning with Guided Projector
https://arxiv.org/abs/2306.06082
Marcin Przewięźlikowski (IDEAS NCBR), Mateusz Pyla (IDEAS NCBR), Bartosz Zieliński (IDEAS NCBR), Bartłomiej Twardowski (IDEAS NCBR), Jacek Tabor, Marek Śmieja
ENLSP 2023:
Exploiting Transformer Activation Sparsity with Dynamic Inference
https://arxiv.org/abs/2310.04361
Mikołaj Piórczyński, Filip Szatkowski (IDEAS NCBR), Klaudia Bałazy, Bartosz Wójcik (IDEAS NCBR)
ICCV 2023:
ICICLE: Interpretable Class Incremental Continual Learning
https://arxiv.org/abs/2303.07811
Dawid Rymarczyk, Joost van de Weijer, Bartosz Zielinski (IDEAS NCBR), Bartłomiej Twardowski (IDEAS NCBR)
ICCV workshops:
AR-TTA: A Simple Method for Real-World Continual Test-Time Adaptation
https://arxiv.org/abs/2309.10109
Damian Sójka, Sebastian Cygert, Bartłomiej Twardowski (all IDEAS NCBR)
Adapt Your Teacher: Improving Knowledge Distillation for Exemplar-free Continual Learning
https://arxiv.org/abs/2308.09544
Filip Szatkowski, Mateusz Pyla, Marcin Przewięźlikowski, Sebastian Cygert, Bartłomiej Twardowski, Tomasz Trzciński (all IDEAS NCBR)
Looking Through the Past: Better Knowledge Retention for Generative Replay in Continual Learning
https://arxiv.org/abs/2309.10012
Valeriya Khan, Sebastian Cygert, Bartlomiej Twardowski, Tomasz Trzciński (all IDEAS NCBR)
WACV 2024:
Towards More Realistic Membership Inference Attacks on Large Diffusion Models
https://arxiv.org/abs/2306.12983
Jan Dubiński (IDEAS NCBR), Antoni Kowalczuk, Stanisław Pawlak,Przemysław Rokita, Tomasz Trzciński (IDEAS NCBR), Paweł Morawiecki
Face Identity-Aware Disentanglement in StyleGAN
https://arxiv.org/abs/2309.12033
Adrian Suwała, Bartosz Wójcik (IDEAS NCBR), Magdalena Proszewska, Jacek Tabor, Przemysław Spurek, Marek Śmieja
CLIP-DIY: CLIP Dense Inference Yields Open-Vocabulary Semantic Segmentation For-Free
https://arxiv.org/abs/2309.14289
Monika Wysoczanska, Michael Ramamonjisoa, Tomasz Trzcinski (IDEAS NCBR), Oriane Simeoni
