july aug | Page 12

Cover Story

local computing on local data .
AI for the home will apply to a
local context and corresponding
datasets , the insights of a much
larger model . CPEs are already
in the home , so it would be
sensible for them to make the
proper hardware and software
available to enable AI for the
home .”
“ In the context of global
warming , CPE-centric
computing and storage would
mean lower costs and energy
consumption related to data
transfer to and from the cloud .
This is why we believe that
CPE combined with the next
generation of sensors we are
exploring and building , will allow us to render
the home ’ s context with the best possible
granularity . This will enable a CPE-based AI
to provide insights and improvements for the
home and the people who live there .”
INFORMED . How will AI help hyperpersonalise
services to optimise content
selection and engage customers ? “ This is
already happening when it comes to the
selection of entertainment , but you can make
more informed predictions and ensure more
targeted awareness of products and services
while maintaining a focus on privacy and
security ,” notes Broadband Forum ’ s Walls .
“ Smarter Customer Premise Equipment
( SCPE ) with built-in application intelligence is
already enabling operators to offer customised
offerings as traditional revenue streams such
as fixed voice and IPTV decline . SCPE creates
an app store-like experience for customers
and offers providers the opportunity to
expand their subscriber base by offering a
“ At the moment , we tend to throw bandwidth and speed at the consumer to serve their digital life and requirements .” – Charles Cheevers , CommScope
host of different consumer and business services which can include e-health / e-learning , IoT applications , enhanced privacy and security and employee homeworking . By adding AI , operators will be able to tailor services much more efficiently by ‘ learning ’ from a user ’ s previous preferences and choices .”
According to CommScope ’ s Cheevers , companies such as NVIDIA highlight the capabilities of AI to work on data streams from the device through the network and cloud – from a Home perspective and more importantly from a consumer perspective – the ability to be able to recognise the specific consumer use of applications and services in the home ( which in itself is an AI function of analysing real-time packet streams from IP destination addresses , UDP ports , Header information , Patterns , correlation to Device types ) can then enable better per home or per services quality of experience and quality of services .
“ At the moment , we tend to throw bandwidth and speed at the consumer to serve their digital life and requirements ” he admits . “ The next phase in home connectivity is to understand what the service needs to be optimal – from its Wi-Fi QoS requirements to its jitter and latency tolerances and roundtrip latency window from source to sink .” POTENTIAL . “ As we use Device and Service identification in the home more , there can also be additional value extracted from this ( assuming the consumer opts in to any privacy agreements ) where improvements in the types of services and devices offered to the consumer
are more appropriate for things
like home size ( extenders
required on larger vs smaller
homes ), reliability wired offload
to LTE / 5G ( where there is higher
frequency of outages based on
historical information , terrain ,
weather ),” notes Gamgee ’ s
Hendricks . “ Within the home
and the devices we can also see
a lot more potential to use AI on
the information that is ‘ given ’ by
devices wishing to participate
more easily in our connected
home – devices like printers ,
bonjour capable devices , AllShare ,
TVs telling you what they support
and their specifications via
protocols are easy to interchange
with a based Wi-Fi AP .
“ While personalisation
and hyper-personalisation
can be achieved without AI ,
the increased availability of
behavioural data necessitates faster and
more accurate predictions and anticipation
of individual user needs . This becomes
particularly relevant for smart home services .
When it comes to content selection , such
as video content , the home network ’ s
contribution is minimal . Major video
distributors , like Netflix , Disney +, and Prime ,
already offer highly personalised content
selection recommendations based on viewing
behaviour , which cannot be tracked by home
network equipment or set-top boxes .”
“ Customers increasingly expect their
experience to be hyper-personalised , including
when it comes to content services ,” says
Omnisperience ’ s Cottam . “ Although current
SVoD players personalise content to the tastes
of individual consumers , the approach is
quite clunky and still requires a lot of effort
on the part of the customer . Tomorrow ’ s
AI-led approach will see the death of scrolling
by being far more accurate , proactive and
intuitive . Instead of scrolling through long
lists , AI will bring the best-fit content to
consumers based on past viewing and current
mood .”
“ Search technologies will be enriched with a hyper-personalised service .” - David Souhami , SoftAtHome
INSIGHT . “ Through the collection of
usage data by various operators , businesses
can leverage AI to gain unique insight into
preferences , behaviour , and context ,” notes
12 EUROMEDIA