ENTERPRISE TECHNOLOGY
Christian Putz, Director Emerging Markets
EMEA, Pure Storage.
clean, transform, label, and store larger
amounts of data. Adding additional high-
quality data points directly translates to
more accurate models and better insights.
“A fundamental reason why deep
learning has seen a surge in success is the
continued improvement of models with
larger data set sizes. In contrast, classical
machine learning algorithms, like logistic
regression, stop improving in accuracy at
smaller data set sizes. As of 2016, a rough
rule of thumb is that a supervised deep
learning algorithm will generally achieve
acceptable performance with around
5,000 labelled examples per category, and
will match or exceed human performance
when trained with a dataset containing
at least 10 million labelled examples,”
explains Christian Putz, Director Emerging
Markets EMEA, Pure Storage.
Recent research from Google has
shown the advantages of increasing the
dataset size, showing logarithmic increase
on performance in vision tasks as data
sets increase in size to 300 million images.
Even further, this research suggests
that higher capacity models require
proportionally larger datasets.
Separation of compute and storage also
allows independent scaling of each tier,
18
Nick Jheng, Sales Account Manager at Synology.
A fundamental
reason why deep
learning has seen
a surge in success
is the continued
improvement
of models with
larger data
set sizes.
avoiding many complexities of managing
both together. As the data set size grows
or new data sets are considered, a scale
out storage system must be able to expand
easily. Similarly, if more concurrent
training is required, additional GPUs
or DGX-1 servers can be added without
concern for their internal storage.
Solution selling
The demand for cloud storage, no matter
public or private, will only increase as data
grows exponentially. With on-premises
private cloud solution such as a network
attached storage, businesses can rest
assured that they enjoy higher levels of
data protection and data sovereignty.
The enterprise storage market continues
to evolve, especially with the trends and
growths of mobile devices, Internet of
Things, Big Data, and Artificial Intelligence.
“The technology for data storage
will only advance as demands grow.
The network attached storage industry
itself has gone through many stages of
innovations as well, from scale-up to scale-
out and the currently highly requested
hyper-converged infrastructure. Channel
partners need to continue cultivating their
knowledge in the application of network
attached storage,” explains Nick Jheng,
Sales Account Manager at Synology.
In addition, customers, especially
business users, need a complete solution.
As network attached storage is only
one puzzle piece in the business IT
infrastructure, channel partners need to
constantly educate themselves in other
aspects, such as networking products and
information security. This allows them
to provide integrated, comprehensive
solutions for their customers.
Issue 15
INTELLIGENT TECH CHANNELS