Deep Learning
5c186016d0844767209dc36e9e61441b-Paper-Conference.pdf
DeMa's focus on sequences diminishes approximately exponentially.
Personalized Steering of Large Language Models: Versatile Steering Vectors Through Bi-directional Preference Optimization
Models (LLMs) and build personalized LLMs tailored for various applications. While fine-tuning seems to be a direct solution, it requires substantial computational resources and may significantly affect the utility of the original LLM.
Zero-Shot Tokenizer Transfer Benjamin Minixhofer
Our method comes close to the original models' performance in cross-lingual and
Initializing Variable-sized Vision Transformers from Learngene with Learnable Transformation
Moreover, the training and storage costs of such method grow linearly with the number of potential scenarios.
Revealing Distribution Discrepancy by Sampling Transfer in Unlabeled Data
The assumption that data are independently and identically distributed (IID) is staple in statistical machine learning. It suggests that a hypothesis selected by an algorithm, after observing several training samples, should perform effectively on test samples from the same unknown distribution.
Dual-Personalizing Adapter for Federated Foundation Models
Foundation models, especially the large language model (LLM) in natural language processing (NLP), have nearly exhausted public data sources for training.
Effective Sharpness A ware Minimization Requires Layerwise Perturbation Scaling
Our findings reveal that the dynamics of standard SAM effectively reduce to applying SAM solely in the last layer in wide neural networks, even with optimal hyperparameters.
Learning Action and Reasoning-Centric Image Editing from Videos and Simulations
Object, attribute or stylistic changes can be learned from visually static datasets. On the other hand, high-quality data for action and reasoning-centric edits is scarce and has to come from entirely different sources that cover e.g.