Filip Szatkowski

5 papers · 2024–2025 · 3 venues · across top CS/AI venues

Authors are matched by name, so researchers who share a name ("Yang Liu", "Wei Wang") appear as a single profile and their paper counts, venues and achievements are combined. How identity is handled.

Achievements

Jump to papers ↓

🌍 Conference Polyglot · Bronze 🗺️ Taxonomy Completionist (17) 🏆 Keyword Champion (2)

Venues

ICML (2) WACV (2) NeurIPS (1)

Top co-authors

Tomasz Trzcinski (3) Bartłomiej Twardowski (2) Marcin Przewięźlikowski (1) Yaoyue Zheng (1) Aleksandra Nowak (1) Fei Yang (1) Mateusz Pyla (1) Jacek Tabor (1) Daniel Marczak (1) Wojciech Łapacz (1)

Research topics

Machine Learning (5) Deep Learning (4) Application Areas (4) Core Methods (3) Artificial Intelligence (2) Architectures (2) Core AI (2) Learning Types · Machine Learning (2) Learning Types · Deep Learning (2) Optimization & Theory · Machine Learning (2) Optimization & Theory · Deep Learning (1)

Keywords

continual learning (3) inference cost (2) catastrophic forgetting (2) class-incremental learning (1) classification head (1) mixture of expert (1) inference efficiency (1) activation sparsity (1) dynamic expert selection (1) intermediate representation (1) representation shift (1) dynamical isometry (1) givens rotation (1) auxiliary classifier (1) teacher adaptation (1) transformer model (1) sparse initialization (1) stability gap (1) dense initialization (1) accuracy (1)

Papers

Exploring the Stability Gap in Continual Learning: The Role of the Classification Head WACV 2025 Improving Continual Learning Performance and Efficiency with Auxiliary Classifiers ICML 2025 Adapt Your Teacher: Improving Knowledge Distillation for Exemplar-Free Continual Learning WACV 2024 Exploiting Activation Sparsity with Dense to Dynamic-k Mixture-of-Experts Conversion NeurIPS 2024 Sparser, Better, Deeper, Stronger: Improving Static Sparse Training with Exact Orthogonal Initialization ICML 2024