Zobrazeno 1 - 10
of 942
pro vyhledávání: '"Smith, Virginia"'
Autor:
Tobaben, Marlon, Souibgui, Mohamed Ali, Tito, Rubèn, Nguyen, Khanh, Kerkouche, Raouf, Jung, Kangsoo, Jälkö, Joonas, Kang, Lei, Barsky, Andrey, d'Andecy, Vincent Poulain, Joseph, Aurélie, Muhamed, Aashiq, Kuo, Kevin, Smith, Virginia, Yamasaki, Yusuke, Fukami, Takumi, Niwa, Kenta, Tyou, Iifan, Ishii, Hiro, Yokota, Rio, N, Ragul, Kutum, Rintu, Llados, Josep, Valveny, Ernest, Honkela, Antti, Fritz, Mario, Karatzas, Dimosthenis
The Privacy Preserving Federated Learning Document VQA (PFL-DocVQA) competition challenged the community to develop provably private and communication-efficient solutions in a federated setting for a real-life use case: invoice processing. The compet
Externí odkaz:
http://arxiv.org/abs/2411.03730
Understanding and mitigating the potential risks associated with foundation models (FMs) hinges on developing effective interpretability methods. Sparse Autoencoders (SAEs) have emerged as a promising tool for disentangling FM representations, but th
Externí odkaz:
http://arxiv.org/abs/2411.00743
Autor:
Thaker, Pratiksha, Hu, Shengyuan, Kale, Neil, Maurya, Yash, Wu, Zhiwei Steven, Smith, Virginia
Unlearning methods have the potential to improve the privacy and safety of large language models (LLMs) by removing sensitive or harmful information post hoc. The LLM unlearning research community has increasingly turned toward empirical benchmarks t
Externí odkaz:
http://arxiv.org/abs/2410.02879
A common approach to make machine learning inference more efficient is to use example-specific adaptive schemes, which route or select models for each example at inference time. In this work we study a simple scheme for adaptive inference. We build a
Externí odkaz:
http://arxiv.org/abs/2407.02348
Large language model (LLM) training and finetuning are often bottlenecked by limited GPU memory. While existing projection-based optimization methods address this by projecting gradients into a lower-dimensional subspace to reduce optimizer state mem
Externí odkaz:
http://arxiv.org/abs/2406.17660
Training on model-generated synthetic data is a promising approach for finetuning LLMs, but it remains unclear when it helps or hurts. In this paper, we investigate this question for math reasoning via an empirical study, followed by building a conce
Externí odkaz:
http://arxiv.org/abs/2406.14532
Machine unlearning is a promising approach to mitigate undesirable memorization of training data in LLMs. However, in this work we show that existing approaches for unlearning in LLMs are surprisingly susceptible to a simple set of targeted relearnin
Externí odkaz:
http://arxiv.org/abs/2406.13356
Low-rank adaptation (LoRA) is a natural method for finetuning in communication-constrained machine learning settings such as cross-device federated learning. Prior work that has studied LoRA in the context of federated learning has focused on improvi
Externí odkaz:
http://arxiv.org/abs/2406.05233
Data-dependent privacy accounting frameworks such as per-instance differential privacy (pDP) and Fisher information loss (FIL) confer fine-grained privacy guarantees for individuals in a fixed training dataset. These guarantees can be desirable compa
Externí odkaz:
http://arxiv.org/abs/2403.05598
Optimizing the performance of many objectives (instantiated by tasks or clients) jointly with a few Pareto stationary solutions (models) is critical in machine learning. However, previous multi-objective optimization methods often focus on a few numb
Externí odkaz:
http://arxiv.org/abs/2403.04099