Zobrazeno 1 - 10
of 35 983
pro vyhledávání: '"NICHOLAS, D."'
The Densest $k$-Subgraph (D$k$S) problem aims to find a subgraph comprising $k$ vertices with the maximum number of edges between them. A continuous reformulation of the binary quadratic D$k$S problem is considered, which incorporates a diagonal load
Externí odkaz:
http://arxiv.org/abs/2410.07388
Autor:
Iacob, Alex, Sani, Lorenzo, Kurmanji, Meghdad, Shen, William F., Qiu, Xinchi, Cai, Dongqi, Gao, Yan, Lane, Nicholas D.
Language model pre-training benefits from diverse data to enhance performance across domains and languages. However, training on such heterogeneous corpora requires extensive and costly efforts. Since these data sources vary lexically, syntactically,
Externí odkaz:
http://arxiv.org/abs/2410.05021
Autor:
Lu, Zhenyan, Li, Xiang, Cai, Dongqi, Yi, Rongjie, Liu, Fangming, Zhang, Xiwen, Lane, Nicholas D., Xu, Mengwei
Small language models (SLMs), despite their widespread adoption in modern smart devices, have received significantly less academic attention compared to their large language model (LLM) counterparts, which are predominantly deployed in data centers a
Externí odkaz:
http://arxiv.org/abs/2409.15790
The relaxation of atomic positions to their optimal structural arrangement is crucial for understanding the emergence of new physical behavior in long scale superstructures in twisted bilayers of two-dimensional materials. The amount of deviation fro
Externí odkaz:
http://arxiv.org/abs/2406.19462
Autor:
Qiu, Xinchi, Shen, William F., Chen, Yihong, Cancedda, Nicola, Stenetorp, Pontus, Lane, Nicholas D.
Recently, machine unlearning, which seeks to erase specific data stored in the pre-trained or fine-tuned models, has emerged as a crucial protective measure for LLMs. However, unlearning approaches for LLMs that have been considered thus far have foc
Externí odkaz:
http://arxiv.org/abs/2406.16810
Autor:
Marino, Bill, Chaudhary, Yaqub, Pi, Yulu, Yew, Rui-Jie, Aleksandrov, Preslav, Rahman, Carwyn, Shen, William F., Robinson, Isaac, Lane, Nicholas D.
As the AI supply chain grows more complex, AI systems and models are increasingly likely to incorporate multiple internally- or externally-sourced components such as datasets and (pre-trained) models. In such cases, determining whether or not the agg
Externí odkaz:
http://arxiv.org/abs/2406.14758
Graph hypernetworks (GHNs), constructed by combining graph neural networks (GNNs) with hypernetworks (HNs), leverage relational data across various domains such as neural architecture search, molecular property prediction and federated learning. Desp
Externí odkaz:
http://arxiv.org/abs/2405.20882
Autor:
Lee, Royson, Fernandez-Marques, Javier, Hu, Shell Xu, Li, Da, Laskaridis, Stefanos, Dudziak, Łukasz, Hospedales, Timothy, Huszár, Ferenc, Lane, Nicholas D.
Federated learning (FL) has enabled distributed learning of a model across multiple clients in a privacy-preserving manner. One of the main challenges of FL is to accommodate clients with varying hardware capacities; clients have differing compute an
Externí odkaz:
http://arxiv.org/abs/2405.14791
Autor:
Roth, Holger R., Beutel, Daniel J., Cheng, Yan, Marques, Javier Fernandez, Pan, Heng, Chen, Chester, Zhang, Zhihong, Wen, Yuhong, Yang, Sean, Isaac, Yang, Hsieh, Yuan-Ting, Xu, Ziyue, Xu, Daguang, Lane, Nicholas D., Feng, Andrew
Several open-source systems, such as Flower and NVIDIA FLARE, have been developed in recent years while focusing on different aspects of federated learning (FL). Flower is dedicated to implementing a cohesive approach to FL, analytics, and evaluation
Externí odkaz:
http://arxiv.org/abs/2407.00031
Autor:
Sani, Lorenzo, Iacob, Alex, Cao, Zeyu, Marino, Bill, Gao, Yan, Paulik, Tomas, Zhao, Wanru, Shen, William F., Aleksandrov, Preslav, Qiu, Xinchi, Lane, Nicholas D.
Generative pre-trained large language models (LLMs) have demonstrated impressive performance over a wide range of tasks, thanks to the unprecedented amount of data they have been trained on. As established scaling laws indicate, LLMs' future performa
Externí odkaz:
http://arxiv.org/abs/2405.10853