AskIDCCloudBDA

#AskIDC Cloud & BDA Tweetchat
Accelerating Analytics and Machine Learning with Cloud
IDC Asia/Pacific
Let’s jump straight into #machinelearning & all related #tech themes in this context. Why is #machinelearning a hot topic now?
CK Chua
Actually the concept of #machinelearning has been around for over 60 years
CK Chua
It is hot now because of the triple combo of data, compute and software innovations that are accelerating machine learning.
CK Chua
#Machinelearning is the ability to automatically learn & improve from experience.
CK Chua
Using algorithms that repeatedly learn from #data, #machinelearning allows computers to find hidden patterns & answers without being explicitly programmed where to look.
Chris Zhang
Repetitive aspect of #machinelearning is important coz when systems are exposed to new #data, they need 2be able to adapt independently.
Chris Zhang
Learning from previous computations 2 produce reliable & sustainable decisions/results. It’s a science that’s not new – but 1 that’s gaining fresh momentum!
IDC Asia/Pacific
What is this supposed momentum & where is it coming from?
jessiecai
#Machinelearning is a form of #artificialintelligence (#AI) that uses algorithms to identify patterns from #data. #Machinelearning systems build models out of identified patterns and make predictions
jessiecai
these predictions can be as simple as providing a recommendation to a shopper on a #retail website or as complex as determining if what viable disease/condition can be deduced from a set of patient symptoms
jessiecai
understandably, because you need to be able to feed #data into these #machinelearning systems, the overhead has traditionally been huge - prohibitive, even
Chris Zhang
#MachineLearning requires A LOT of processing power that not everyone will have access 2.
Chris Zhang
The momentum of #machinelearning is the application of #cloud technology. Using #cloud-based #data drives $ down while making access to machine learning a lot more democratic.
IDC Asia/Pacific
So #cloud is becoming the more viable model to storing #data sets from which computers programmed for #machinelearning learn from. Could you explain why, @Daphne_Chung?
Daphne
The benefit of #cloud to #machinelearning is not just in storage, but also processing power.
Daphne
Orgs who wish to use #machinelearning for predictive #analytics had to shell out significant investments for hardware & software. The #cloud is changing all of that.
Daphne
Another thing is the build-out of #machinelearning libraries that can be consumed by developers. This enables the creation of new innovative services made available As-a Service in the #cloud.
IDC Asia/Pacific
(Note to our viewers: the livestream view on the right will give you a more updated view of the responses as they come in!)
IDC Asia/Pacific
(You can also ask us questions by replying to this thread or sending us a DM on Twitter.)
Ron Stuart
In my specific context w/c is #government, I think #machinelearning & #AI thru #cloud will have a huge impact on the delivery of services, resolving citizen issues through identification of the key attributes of the citizen
IDC Asia/Pacific
From an end-user’s perspective, @ronstuart54, have you seen how much easier it is to provide #machinelearning systems w/ access to #data sets through #cloud?
Ron Stuart
This will also aid in the rapid response to topical issues & the development of policy initiatives
IDC Asia/Pacific
Will there be specific industries that will prefer the traditional delivery model of #data over #cloud?
CK Chua
For sure, industries such as the healthcare sector that need to address patient privacy requirements.
CK Chua
Or highly localized operations such as airport operations and logistics where fast streaming machine data are processed in real-time.
jessiecai
#MachineLearning & #AI applications need to stay close to the #data asset to minimize latency n provide close-to-real-time inference
jessiecai
so #cloud or not, its delivery model is more tied up with use cases, rather than industries
jessiecai
note also there are 2 organic parts of #machinelearning: training & inference.
jessiecai
training is more computation-intensive & less latency-sensitive, making more sense for this to happen in the #cloud
Chris Zhang
not forgetting a lot of digital natives R adopting a #cloud-first approach. #machinelearning and #analytics where #data resides will make sense #datagravity
Daphne
Agree with @jessiecaidanqin. Use cases rather than industries but if I had to pick, consumer focused industries will be at the forefront.
Ron Stuart
I think the question is more about timing rather than choice– in my view cloud will become ubiquitous over time but speed of uptake will vary by industry, relative to investments already made by those specific industry sectors.
IDC Asia/Pacific
What are some of the relatable/tangible applications of #machinelearning that you immediately see happening because #cloud has made it more accessible?
IDC Asia/Pacific
Who (or what industries) are the key beneficiaries for this?
jessiecai
#Microfinancing & fraud detection in #banking for example need to combine consumer #data from multiple sources, & #cloud would be the key component.
jessiecai
similarly #cloud would facilitate collaborations & further #AI development in many sectors
Ron Stuart
I have seen great advances in #machinelearning & #AI in #transportation.
CK Chua
Smart devices such as mobile phones and smart speakers are consumer facing applications that have greatly benefited from #cloud #machinelearning.
Ron Stuart
The ability to harmonize information in ways that makes #transportation decisions much easier – this will have great impact on cities by increasing productivity, efficient decisions, & use of transport.
IDC Asia/Pacific
Where does #APeJ stand in terms of #machinelearning adoption & do you see any particular countries/industries/companies benefiting from #machinelearning through #cloud?
Daphne
In healthcare, diagnosing scan results help doctors see more patients more efficiently.
Chris Zhang
Australia & Singapore R definitely some of them, but we do see a lot of traction coming from the emerging geos such as #Indonesia 2
jessiecai
.@IDC recently ran an adoption survey on #AI. Industries including #retail, wholesale & BFSI are more inclined to consume AI through #cloud
jessiecai
for countries, orgs in #China & #Thailand were more for cloud – but overall it is still too early to tell. many responded that they were not decided yet.
Ron Stuart
Without being parochial I think #NewZealand is a great example, being recognized globally as 1 of 3 most digitally-advanced nations alongside #Singapore & #HongKong.
CK Chua
almost every next-gen #machinelearning marketing service or tool is running on #cloud
Ron Stuart
This is further supported by #NewZealand being recognized in the recent @IDC #SmartCity #SCAPA2017 awards w/ 4 won categories.

#Singapore received 3 awards & #HongKong, 2, in #SCAPA2017.
Ron Stuart
Many of the projects singled out in #SCAPA2017, while not necessarily focused on #machinelearning, merge the utilization of #data & #cloud technology for better service delivery.
Chris Zhang
We R refreshing the next iteration of our #bigdataanalytics #idctracker reports. This yr, we also cover #cognitive as a dedicated section
IDC Asia/Pacific
There’s no doubt that #data is growing. But organizations are starting to realize that this data can be put to good use. Can you tell us more about that?
CK Chua
By 2020, the #digital universe will reach 44ZB of data or 44 trillion gigabytes, a tenfold increase over 2013.
CK Chua
If we look back on 2013, 22% of the information in the #digital universe was usable for analysis, but less than 5% of this information was actually analyzed.
CK Chua
In present day looking towards the future, we see companies racing to address this gap. Companies are now focused on harnessing the ability to unlock value from information. For good reason!
jessiecai
.@IDC research suggests that information-based orgs are more competitive w/in their industry — more than 2x as likely to outperform their peers, substantially
jessiecai
so developing this capability is not just for competitive advantage. Ultimately that’s where everyone is heading n where most growth is anticipated. best to get started or risk being left behind.
Chris Zhang
This in a nutshell, this movement 2wards utilization & monetization of #data, is the cornerstone of info mastery, 1 of the tenets of @IDC’s #DigitalTransformation framework. #DX
Chris Zhang
In a #digitaleconomy, success is measured by the ability to generate revenue from info-based products, experiences, & services.
Chris Zhang
Capabilities will have to be built around achieving the goals of info #DX goals w/ the help of #cloud, #bigdata & #analytics.
Daphne
These days we are witnessing an explosion of information. Orgs & even individuals are navigating a maze of ever-increasing #data that can be difficult to manoeuvre through, let alone control & dissect.
Daphne
This surge in the volume of #data presents both challenges & opportunities for the #cloud. How to manipulate & extract the essence of the data more than just storing & retrieving it.
Daphne
As mentioned by my research colleagues, businesses can get increased value from #data insights gained through #bigdata #analytics supported by #cloud infrastructure.
Daphne
This explosion in both structured & unstructured data means the ways to harness the benefits of #cloud & #bigdata are more important than ever!
Daphne
#Cloudcomputing & #bigdata, while still in constant evolution, are proving to be the ideal combination. Together they provide cost-effective, scalable infrastructure to support #bigdata & business/operational #analytics.
IDC Asia/Pacific
Over to you, @ronstuart54! When you see orgs racing to develop their information mastery as part of their #DX agenda, what are some of the common gaps you’ve observed?
Ron Stuart
From my perspective, current information practices focus too much on #analytics alone rather than the use of analytics to gain actionable insights.
Ron Stuart
I see the need to grow capability to answer the big ‘what-if’ questions rather than #datascience that primarily focuses on what the #data tells us about today.
IDC Asia/Pacific
Great insights from all! This ability for machines to determine “what-ifs” is a great lead-in to the core of our discussion today, on how #cloud is helping ramp up advancements in #analytics & #machinelearning.