Alignment, interpretability, red-teaming, bias and privacy: the research on what these systems do when they misbehave.
arXiv:2610.34623v1 Announce Type: new Abstract: Text injection trains speech recognition models on unpaired text alongside paired audio. We feed the speech encoder synthetic input derived from text s...
arXiv:2610.35163v1 Announce Type: new Abstract: Disseminating synthetic data in place of original ones is a promising way to enforce privacy while complying with open science principles, especially i...
arXiv:2610.35311v1 Announce Type: new Abstract: We develop a derivative-free temporal-difference (TD) framework for calibrating black-box simulators using one-step transitions from a target system an...
EF-LLM is a continual learning framework that combines domain knowledge, temporal data, and multimodal inputs to forecast energy demand and generation. It uses updatable Low‑Rank Adaptation and a multi‑channel architecture to continuously adapt to new data, improving accuracy especially under sparse‑data conditions. The model also uniquely detects and quantifies hallucinations through multi‑task learning and semantic similarity analysis.
arXiv:2509.06984v4 Announce Type: replace-cross Abstract: Federated Learning with LoRA fine-tuning offers an efficient and privacy-aware solution for institutions to collaboratively leverage their la...
arXiv:2510.15038v2 Announce Type: replace-cross Abstract: Flow-based Generative Models (FGMs) effectively transform noise into complex data distributions. Incorporating Optimal Transport (OT) to coup...
arXiv:2511.03190v2 Announce Type: replace-cross Abstract: Attention mechanisms are effective for multivariate time series modeling because they adaptively aggregate information across timestamps and...
arXiv:2603.02531v3 Announce Type: replace-cross Abstract: Classifier-Free Guidance (CFG) improves sample quality in diffusion models, but its dual-pass inference and null-condition training limit its...
arXiv:2603.13272v2 Announce Type: replace-cross Abstract: Electroencephalography (EEG) foundation models have shown promise for learning generalizable representations, yet they remain sensitive to ch...
arXiv:2610.32822v2 Announce Type: new Abstract: Sparse linear regression is a fundamental problem in high-dimensional statistics, and its private counterparts leave a large gap between the central an...
arXiv:2504.11952v4 Announce Type: replace-cross Abstract: An ideal detection system for machine generated content is supposed to work well on any generator as many more advanced LLMs come into existe...
arXiv:2505.01670v3 Announce Type: replace-cross Abstract: Reconstructing visual images from fMRI data is a central but challenging problem in neuroscience. Despite recent progress, current methods fa...
arXiv:2601.06961v2 Announce Type: replace Abstract: The success of deep neural networks largely depends on the statistical structure of the training data. While learning dynamics and generalization o...
Health‑SCORE is a scalable rubric‑based framework designed to evaluate and train health‑focused large language models (LLMs). It reduces the need for extensive human expertise and development cost while maintaining evaluation quality comparable to human‑created rubrics. The framework also serves as a structured reward signal for reinforcement learning and can be embedded in prompts to enhance response quality via in‑context learning.
arXiv:2601.21200v3 Announce Type: replace Abstract: Classifier-guided diffusion models generate conditional samples by augmenting the reverse-time score with the gradient of the log-probability predi...
arXiv:2604.08362v3 Announce Type: replace-cross Abstract: The emergence of Large Language Models (LLMs) has illuminated the potential for a general-purpose user simulator. However, existing benchmark...
arXiv:2604.18835v2 Announce Type: replace-cross Abstract: We propose a scalable, multifactorial experimental framework that systematically probes LLM sensitivity to subtle semantic changes in pairwis...
arXiv:2605.28036v2 Announce Type: replace Abstract: Diffusion models steer conditional generation with a tunable guidance scale to trade off prompt alignment and diversity. However, existing debiasin...
arXiv:2607.16675v2 Announce Type: replace Abstract: A predictor with small overall bias can still exhibit systematic bias conditional on its predicted value, limiting its reliability for downstream u...
arXiv:2608.25302v2 Announce Type: replace Abstract: Guided depth super-resolution (GDSR) typically extracts RGB guidance features through convolutional hierarchies, inheriting their fine-to-coarse bi...