KAUST SANDS Lab Recruiting CS MS/PhD - 留學

來自義大利的老闆今天請我幫忙宣傳招生, 竟然直接點名PTT studyabroad板...

個人在 KAUST 兩年來對他的印象一直很好, 確實有實力也能給學生幫助

所以滿願意協助宣傳, 若有相關問題歡迎推文或站內信詢問

關於 KAUST 的文章板上也能找到幾篇, 請 /KAUST 搜尋即可


以下代PO

-----------------------------------------------------------------------



We are looking for two PhD students with a strong interest to work at the
intersection of distributed systems and machine learning. The successful
candidates will be expected to devise scalable and robust distributed
algorithms and systems for (deep) machine learning models.



Machine learning (ML) is the foundation of many of today's applications,
from classification systems for image and speech recognition, to guiding
self-driving cars. Due to the large data volume and ubiquitous data sensing
from edge devices, e.g., smart phones, ML is shifting from the centralized
cluster setting to distributed systems: in a setting known as federated
learning, data holders collaborate to train a global model by sharing their
parameter updates. This setting is challenging because of many real-world
factors like resource heterogeneity, varying network connectivity, and
different levels of client participation, which make it hard to learn high
quality, robust models. Besides, these models can take days or even weeks to
train!



One research direction is to design algorithms to optimize the trade-off
between computation and communication of distributed learners and for
compressing communication of model updates (see our GRACE project:
https://sands.kaust.edu.sa/project/grace ).



A second direction considers opportunities for accelerating deep learning
computations via new hardware architectures (whether on-device, cloud offload
or edge/in-network computing), including FPGAs in collaboration with Dr.
Suhaib Fahmy at the University of Warwick, UK.



Another is focused on designing fairness-aware algorithms to mitigate
system-level causes that bias the model towards certain clients over others.



Having a strong background in machine learning is essential. Being familiar
with distributed learning systems or federated learning is helpful but not
required.



Interested?

Please get in touch with [email protected] to discuss your application.

But make sure you read this first: https://mcanini.github.io/students.html



If you have a BS degree, apply to our combined MS/PhD program. If all goes
well, you will get an MS degree in 1.5 years (3 semesters), and will then
continue towards your PhD. If you have an MS degree already, apply directly
to the PhD program.



SANDS Lab

At SANDS Lab (http://sands.kaust.edu.sa/ ), we perform world-class research in
the design, implementation, deployment, and analysis of large-scale networked
systems. We are a group of systems builders and we currently focus on the
scalability, efficiency and robustness of deep learning. Much of our work is
experimental: to validate our proposed concepts, we build system prototypes
that directly improve the lives of real users.

Located on the shores of the Red Sea, KAUST is a fantastic place to study and
live at. Every student is fully funded with a generous fellowship (~$30,000
USD/year) that covers a stipend, housing and more.

Information on application requirements and on the university is available at
this page: https://www.kaust.edu.sa/en/study

--