Causal Feature Expansion to Enhance Class-Incremental Learning
Improve class-incremental learning with causal feature expansion and CPNS regularization to reduce forgetting and boost model robustness.
TildeOpen LLM: Boosting Multilingual AI for European Languages
Discover how TildeOpen LLM uses curriculum learning to improve AI performance across 34 European languages, ensuring equitable language representation.
CoFL: Advanced Language-Based Navigation with Flow Fields
Discover CoFL, a cutting-edge continuous flow field model for real-time, language-conditioned navigation in dynamic environments with high precision.
How Regularity Boosts Learnability in Numeral Systems
Explore how regular numeral systems enhance learnability using Reinforcement Learning, revealing key insights into language and cognition.
ReLoop: Enhancing Reliability in LLM Optimization Code
Discover how ReLoop improves accuracy and reliability in LLM-generated optimization code using structured modeling and behavioral verification.