Two new research papers introduce novel approaches to enhance the robustness of Random Vector Functional Link (RVFL) networks, which are known for their efficiency but susceptibility to noisy data. The first paper, RoBell-RVFL, proposes a quality-aware, sample-level weighting mechanism that preserves minority class information while adaptively down-weighting majority class samples influenced by noise or outliers. The second paper, KRPRVFL, integrates a kernel risk-sensitive mean p-power criterion with a collaborative learning mechanism to mitigate the impact of corrupted samples and improve stability. Both methods were evaluated on benchmark datasets and demonstrated superior performance over existing RVFL variants, particularly in scenarios with label noise and class imbalance. AI
IMPACT These new RVFL network variants offer improved performance on noisy and imbalanced datasets, potentially expanding their applicability in real-world classification tasks.
RANK_REASON Two academic papers published on arXiv introducing new model architectures.
- alphaXiv
- CatalyzeX
- DagsHub
- Gotit.pub
- Hugging Face
- IArxiv
- Influence Flower
- Keel
- Kernel Risk-Sensitive Mean p-Power
- KRPRVFL
- ScienceCast
- University of California, Irvine
- IArxiv Recommender
- RoBell-RVFL
- SMOTE
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →
