Online Reasoning Calibration for Efficient LLM Test-Time Training
Discover ORCA, a test-time training framework that boosts large language model efficiency and accuracy through adaptive reasoning calibration.
AdaLoRA-QAT: Efficient Low-Rank Quantized Segmentation
Discover AdaLoRA-QAT, a framework combining adaptive low-rank and quantization-aware training for efficient, high-accuracy chest X-ray segmentation.
Brainstacks: Efficient Cross-Domain Continual Learning for LLMs
Discover Brainstacks, a modular architecture enhancing continual multi-domain learning in LLMs with frozen MoE-LoRA stacks for efficient, zero-forgetting A...
Generative Nonlinear Unmixing for Hyperspectral Pixels
Discover a novel generative approach to nonlinear hyperspectral unmixing using bi-directional GANs for accurate pixel analysis without prior mixing models.
PaperRecon: Evaluating AI Paper Presentation & Hallucination
Discover PaperRecon, a new framework to assess AI-written papers by analyzing presentation quality and hallucination accuracy in academic research.