Software as a service, Management Cloud and AI: The new recipe for a technological revolution
Health fund. Well, angela, is an ai professional specialized in ethics. Explainability diversity and inclusion in ai and a recipient of top 10 analytic leaders 2020 from the institute of analytics professional australia. She also sits on the founding editorial board of uh springers, new and timely ai and ethics journal. Thank you so much angela for now joining us at the world. Ai show its a pleasure having you with this. I introduce you to this stage and screen. Thank you, angela for joining us today. Lovely. Thank you always pleasure. Thank you thanks for having me hi, everyone. Well, welcome to um, you know, trisco, you know ai world ai show and its such a pleasure for me to join today to share about you know ai and cloud. So let me share the screen with you, so you can see that theres um, you know people um, you know generally, when people ask about ai and cloud relationship. Most often people you know can get the idea why cloudy um cloud computing is unrelated to ai, but today id like to um. Take you through a few key points that you know really plays in a fundamental foundation block and also the application of ai and how thats related to cloud and cloud computing. So let me start with um key cloud characteristics. The main goal of cloud computing is to allow iot resources to actually delivered and consumed as a service. So, as you can see that theres scalability and electricity, massive scalability and rapid electricity includes compute storage and network capacity, and also pooling so pulling over shared resources, networks and servers and storage and applications requires multi, tenancy model and measure the service measure service or pay.
As you go only for what is needed on demand and service service, so increase or decrease resources as needed provision without requiring human interaction with the service provider. The second last point is network capabilities assessed over the network and capex to your packs. Capital expenditure is often converted to operational expenditure. So what does that mean to us so id like to um? Take on to the history of ai and cloud computing, so references to the phrase cloud computing appeared as early as 1996, so with the first no mention in compact internal document, so the term cloud was introduced even before 1993, but modern cloud computing was created and popularized By amazon in 2006, with the elastic compute cloud service, the first cloud delivery model was the infrastructure as a service to provide prepackaged iot resources to users and software service were introduced. All these delivery models were used for different workloads, including ai workloads, and you can see that you know when you um, you know talk about cloud deployment models. Um there are public, private and hybrid, so these deployment models are also important to locate ai workload based on functional and known functional requirements and constraints. So, according to the world economic forums, the future of jobs report cloud computing is the highest priority of business leaders and until 2025 cloud computing will be adopted. The most by companies, though artificial intelligence, studied much earlier than cloud computing cloud computing and its technologies have improved ai very much cloud.
Computing has been an effective catalyst so to speak, and you can see on your right hand, side where i am looking at um. You can see that you know at a glance how data you know: processing capability models and algorithms and talent, skills and artificial intelligence. It is an ecosystem for cloud and computing. So lets have a look at you know, dedicated perspectives on the impact of ai, as you can see that 2010 apple acquires siri, the ai voice assistant and microsoft launches um connect for xbox, 360. 2021. Ibm watson wins at the jail party, defeating two former champions and 2013, and an neil never ending image. Learner is a computer program created to learn information about online images and com, common sense relationships and 2015. You can see that googles, open source, deep learning framework tensorflow is released in 2016. Google demands alphago bit world champion. Lee settle at computer programming, go 4 out of 5 times and 2017. Two chatbots are trained by facebook, ai research lab to learn how to negotiate and speak 2018 wemo. The company owned by alphabet inc that is attempting to create a fully autonomous vehicle reaches 10 million self driven miles 2019. The fake detection challenge begins with the help of aws bbc news, cbc news, facebook, first draft microsoft, the new york times witness mx prize. So as you can see that its not surprising that, actually you can see that you know how ai and cloud computing there um.
They are really um, you cant, separate them and because they they go hand in hand. So when you talk about cloud, computing services has morphed from platforms such as google, app engine and azure to infrastructure, which involves the provision of machines for computing and storage. In addition to this, cloud providers also offer data platform services, which is span the different available databases, this chain of development points in the direction of the growth of artificial intelligence and cloud computing. What is artificial intelligence, then? Ai popularly referred to on artificial intelligence, which is referred as ai, refers to the simulated intelligence machines. The term refers to the end result of endowing machines with a intellectual process peculiar to humans, the ability to reason, learn and from the past discover meaning or generalize. So when you talk about you know cloud computing, you there are existing types of clouding application development services. I have lightly touched on previous slide, but this one really gives you a bit of detail about. You know infrastructure as a service and platform as a service and software as a service, which is very fundamental on the backbone when you talk about in a cloud computing application. So infrastructure service is a cloud app development service which is mostly employed by users. It allows you to pay based on the users of the service provided and flexible plan the platform as a service. This was designed to make web creation and mobile app design easier to have in built infrastructure of servers, networks, databases and storage that eliminates the need to constantly update them or manage them.
Software as a service is cloud provider. Not the user is tasked with management and maintenance, or all the user has to do to gain. Access is connected to the application of the internet with web browser or phone tablet or pc, so type of cloud deployment, public cloud, so public clouds, like microsoft, azure the cloud provider owns and manage all hardware, software and other supporting instructors and responsible for delivering computing resources. Service storage of the internet, as a user, you can gain access to these services and manage your account through the web browser. So when you talk about the private cloud, just as the name implies, private cloud service and infrastructure are maintained on a private network, either by the providing company or hired a third party service provider. Hybrid hybrid cloud, as the name indicates its both the public and the private cloud services and how the model is possible. It is made available by the integration of the personalized data and application shared by the both platforms, clients, looking for more flexible cloud, app development solution and a wide range of deployment options advised to embrace this technology. So when you talk about all these um, you know infrastructure for computer cloud computing. This is really to do with machine learning, so machine learning models when at large scale the data is up, data set is applied, it applies to certain algorithms and it becomes important to leverage the cloud for this. The models are able to learn from the different patterns which are gleaned from the available data.
So, if i have to, i can show from the example how it can be um, you know generated from the diagram ill show you um from this slide. So, as you can see that when i talked about how the model generated and when you ingest the data from data ingestion stage, by the way, this example is um, the one we um, you know deploy that work um in a database environment. And you can see that how the cloud computing plays really people to role, because you can see that you know data lake enables and it really comes together, um how its um, you know. Data can be um ingested and um. You know and the in the in the production environment how data can be scored and in the adva on the analytics factory environment. Actually, the machine learning is actually um run and score almost real time, and this can be um cyclical, and this can be done all completely end to end seamlessly. So, as you can see that the power of um cloud and computing actually enables the machine learning model and which is a crucial part, when you talk about ai and ai and cloud computing, and also we have to talk about benefits of leveraging, ai and cloud computing Costs effectiveness so by being accessible through the internet, the cloud application development eliminates the need for expenses on site hardware and software and purchases and setup. It also eliminates the need on site data centers and the expenses that comes with it.
It experts manage the center servers and run the clock, electricity to power and cool the servers. So, as you can see that ai and cloud computing is um, you know really empowered by and providing all these you know benefits which are increased productivity. Unlike a hard, develop driver, logical local storage device, which requires a lot of item management tools, hardware, setup software – you know patching racking and stacking cloud. Compute is all internet based and as such, has no need for this. This gives room for the iit team to focus on achieving other business goals, reliability with hard drive or physically accessible infrastructures. The risk of damage is heightened, so one phase is the risk of the crash lost files, lack of failure and so much more. However, cloud computing solutions ensure business continuity, faster and easier disaster recovery and easier data, backup, availability of advanced infrastructure, ai applications are generally high performance when on servers with multiple and very fast graphics processing units gpus. These systems are, however, extremely expensive and unaffordable for many organizations. Ai as a service in cloud application development becomes accessible to these organizations at a more affordable price conclusion. We strongly believe that the fusion of cloud computing services and ai technology will bring a significant change in the technology industry. Public cloud providers keep on investing in the growth of ai, and this will continue to attract the right set of clients to this technology. Even though the technology is still in its early stage.
The evolution to come is inevitable and we can expect phenomenal advancements in the future id like to show um, you know kind of infrastructures of um, the typical on the business environment. How, for example, you know when you have all the data sources and data platform, and you know how your you know. This is hybrid hybrid model, as you can see that you know data lake thats where all this some cloud computing is happening and it can consume unstructured, streamed and structured data in different formats, and it can ingest all these internal data sources, external sources and real time. Data and supplementary sources as well in a consumption layer, it enables for smooth and seamless, reporting and dashboard, and that really enables um the capability of analytics programs which can um create benefit and cost um efficiencies for the business and the customers as well. You can see that there are some the user cases ive mentioned id like to highlight on all these different data points, and you know how the organization grow. They collect more data, and nowadays you know the accumulation of data is unimaginable. You know you even um bring in all this image data and the processing on the component becomes really important and thats. Why its almost um indispensable that you need to really adopt this new way of um the using the cloudy environment for especially for in a model building and machine learning model um deployment? So, as you can see that data exploration and collection can be happening in um, the legacy in the on on premise, environment and even data profiling and quality volume can be all happening there and from the moment where data transformation and feature engineering and model classification and Model tuning all these process has been done in typically in the um cloud im computing um.
You know environment because the real time and continuous im scoring the machine learning model through the algorithm and the fine tuning and all these simplest um. The process has p. That has to be done in a very efficient manner and thats. Why again, all this typical uh for this um typical example was um the azure environment, but in data lake and data ingestion, and also that of factory construction and the all this model scoring is happening in real time analyst factory. So this is all um. You know end to end journey of um ai model deployment, and this is um happening in cloud environment. Thank you so much for the opportunity for me to share um this um um. You know insights and about you know, whats happening in ai and cloudy environment and i hope um you enjoyed my talk if theres any further questions, please reach out to me to linkedin angela kim, please thank you.