Tag: adversarial machine learning
Browse our exclusive articles!
Mitigating Safety Risks in Large Reasoning Models with Adaptive Steering
Explore safety risks in large reasoning AI models and how adaptive multi-principle steering reduces harmful outputs while maintaining accuracy.
Optimizing Attention in Large Vision-Language Models
Discover how inefficiencies in attention mechanisms affect large vision-language models and explore new strategies for improved AI performance.
Locality-Aware Private Class ID for Domain Adaptation
Discover a new locality-aware method for private class identification in domain adaptation, improving accuracy under extreme label shifts with ReOT.
BitCal-TTS: Boost Quantized Reasoning Model Accuracy
Discover BitCal-TTS, a novel method improving confidence calibration and stability in quantized reasoning models for better AI inference.
LANTERN: Efficient Neurosymbolic Transfer with LLMs
Discover LANTERN, a novel LLM-augmented neurosymbolic transfer framework boosting reinforcement learning efficiency by 40-60% with adaptive reasoning.