Zobrazeno 1 - 10
of 474
pro vyhledávání: '"Kannan Ramchandran"'
Autor:
Amirali Aghazadeh, Hunter Nisonoff, Orhan Ocal, David H. Brookes, Yijie Huang, O. Ozan Koyluoglu, Jennifer Listgarten, Kannan Ramchandran
Publikováno v:
Nature Communications, Vol 12, Iss 1, Pp 1-10 (2021)
Finding a biologically-relevant inductive bias for training DNNs on large fitness landscapes is challenging. Here, the authors propose a method called Epistatic Net that improves DNN prediction accuracy and interpretation speed by integrating the kno
Externí odkaz:
https://doaj.org/article/aaa0481b36da43839b3bcab8d8fbcc3a
Publikováno v:
IEEE Transactions on Signal Processing. 71:587-600
Publikováno v:
Machine Learning and Knowledge Discovery in Databases ISBN: 9783031264115
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::47d00bd20121bbc6759b6d2d1c940b73
https://doi.org/10.1007/978-3-031-26412-2_19
https://doi.org/10.1007/978-3-031-26412-2_19
Publikováno v:
IEEE Transactions on Information Theory. 68:1851-1885
Publikováno v:
IEEE Journal on Selected Areas in Information Theory. 2:942-953
We develop a communication-efficient distributed learning algorithm that is robust against Byzantine worker machines. We propose and analyze a distributed gradient-descent algorithm that performs a simple thresholding based on gradient norms to mitig
Autor:
Yijie Huang, Kannan Ramchandran, Hunter M. Nisonoff, Orhan Ocal, Amirali Aghazadeh, David H. Brookes, O. Ozan Koyluoglu, Jennifer Listgarten
Publikováno v:
Nature Communications, Vol 12, Iss 1, Pp 1-10 (2021)
Nature Communications
Nature Communications
Despite recent advances in high-throughput combinatorial mutagenesis assays, the number of labeled sequences available to predict molecular functions has remained small for the vastness of the sequence space combined with the ruggedness of many fitne
Autor:
Kannan Ramchandran, Zihan Liu
Publikováno v:
SIAM Journal on Optimization. 31:1971-1998
The adaptive Douglas--Rachford splitting algorithm iteratively applies the operator $T=\kappa_{n}Q_{A}Q_{B}+(1-\kappa_{n}){Id}$ to solve the inclusion problem $\text{zer}(A+B)$. By taking a Yosida ...
Publikováno v:
IEEE Transactions on Information Theory. 68:5850-5851
Autor:
Zhengming Zhang, Yaoqing Yang, Zhewei Yao, Yujun Yan, Joseph E. Gonzalez, Kannan Ramchandran, Michael W. Mahoney
Publikováno v:
2021 IEEE International Conference on Big Data (Big Data).
Autor:
Michael W. Mahoney, Arun Kejariwal, Dhruv Choudhary, Vipul Gupta, Xing Wang, Yuzhen Huang, Ping Tak Peter Tang, Xiaohan Wei, Kannan Ramchandran
Publikováno v:
KDD
In this paper, we consider hybrid parallelism -- a paradigm that employs both Data Parallelism (DP) and Model Parallelism (MP) -- to scale distributed training of large recommendation models. We propose a compression framework called Dynamic Communic