Követés
Newsha Ardalani
Newsha Ardalani
Research Scientist, Meta AI Research (FAIR)
E-mail megerősítve itt: cs.wisc.edu
Cím
Hivatkozott rá
Hivatkozott rá
Év
Deep learning scaling is predictable, empirically
J Hestness, S Narang, N Ardalani, G Diamos, H Jun, H Kianinejad, ...
arXiv preprint arXiv:1712.00409, 2017
5802017
Sustainable ai: Environmental implications, challenges and opportunities
CJ Wu, R Raghavendra, U Gupta, B Acun, N Ardalani, K Maeng, G Chang, ...
Proceedings of Machine Learning and Systems 4, 795-813, 2022
2392022
Stream-dataflow acceleration
T Nowatzki, V Gangadhar, N Ardalani, K Sankaralingam
Proceedings of the 44th Annual International Symposium on Computer …, 2017
1862017
Cross-architecture performance prediction (XAPP) using CPU code to predict GPU performance
N Ardalani, C Lestourgeon, K Sankaralingam, X Zhu
Proceedings of the 48th International Symposium on Microarchitecture, 725-737, 2015
1392015
Dataperf: Benchmarks for data-centric ai development
M Mazumder, C Banbury, X Yao, B Karlaš, W Gaviria Rojas, S Diamos, ...
Advances in Neural Information Processing Systems 36, 2024
712024
Beyond human-level accuracy: Computational challenges in deep learning
J Hestness, N Ardalani, G Diamos
Proceedings of the 24th symposium on principles and practice of parallel …, 2019
692019
Hybrid Optimization/Heuristic Instruction Scheduling for Programmable Accelerator Codesign
T Nowatzki, N Ardalani, K Sankaralingam, J Weng
Proceedings of the 27th International Conference on Parallel Architectures …, 2018
512018
Deep learning scaling is predictable
J Hestness, S Narang, N Ardalani, G Diamos, H Jun, H Kianinejad, ...
Empirically. arXiv 1712, 2, 2017
472017
Stream-dataflow acceleration. In 2017 ACM/IEEE 44th Annual International Symposium on Computer Architecture (ISCA)
T Nowatzki, V Gangadhar, N Ardalani, K Sankaralingam
IEEE Proc. ISCA, Toronto, ON, Canada, 24th-28th Jun, 2017
232017
Systems and methods for stream-dataflow acceleration wherein a delay is implemented so as to equalize arrival times of data packets at a destination functional unit
K Sankaralingam, A Nowatzki, V Gangadhar, P Shah, N Ardalani
US Patent 11,048,661, 2021
212021
Deep learning scaling is predictable, empirically, arXiv
J Hestness, S Narang, N Ardalani, G Diamos, H Jun, H Kianinejad, ...
arXiv preprint arXiv:1712.00409, 2017
212017
A Static Analysis-based Cross-architecture Performance Prediction Using Machine Learning
N Ardalani, U Thakker, A Albarghouthi, K Sankaralingam
arXiv preprint arXiv:1906.07840, 2019
20*2019
Time and the Value of Data
E Valavi, J Hestness, N Ardalani, M Iansiti
arXiv preprint arXiv:2203.09118, 2022
162022
Deep Learning Scaling is Predictable, Empirically
J Hestness, S Narang, N Ardalani, G Diamos, H Jun, H Kianinejad, ...
arXiv preprint arXiv:1712.00409, 2017
132017
Understanding scaling laws for recommendation models
N Ardalani, CJ Wu, Z Chen, B Bhushanam, A Aziz
arXiv preprint arXiv:2208.08489, 2022
82022
Decoding data quality via synthetic corruptions: Embedding-guided pruning of code data
Y Yang, AK Singh, M Elhoushi, A Mahmoud, K Tirumala, F Gloeckle, ...
arXiv preprint arXiv:2312.02418, 2023
52023
Mp-rec: Hardware-software co-design to enable multi-path recommendation
S Hsia, U Gupta, B Acun, N Ardalani, P Zhong, GY Wei, D Brooks, CJ Wu
Proceedings of the 28th ACM International Conference on Architectural …, 2023
52023
Empirically Characterizing Overparameterization Impact on Convergence
N Ardalani, J Hestness, G Diamos
42018
Sieve: Multimodal dataset pruning using image captioning models
A Mahmoud, M Elhoushi, A Abbas, Y Yang, N Ardalani, H Leather, ...
arXiv preprint arXiv:2310.02110, 2023
32023
Towards MoE Deployment: Mitigating Inefficiencies in Mixture-of-Expert (MoE) Inference
H Huang, N Ardalani, A Sun, L Ke, HHS Lee, A Sridhar, S Bhosale, CJ Wu, ...
arXiv preprint arXiv:2303.06182, 2023
22023
A rendszer jelenleg nem tudja elvégezni a műveletet. Próbálkozzon újra később.
Cikkek 1–20