AI in Telecommunications

Summary

In this session led by Asia Open RAN Academy, artificial intelligence (AI) and machine learning (ML) are explored as transformative forces in the telecommunications industry. Despite an unexpected speaker substitution, the talk delves into how intelligent technologies are revolutionizing everything from network optimization to customer service. AI is not just enhancing daily communication but is pivotal in futuristic innovations like open radio access networks (Open RAN). This summary highlights AI's role in automating and optimizing network performance, ensuring cybersecurity, and predicting future trends in telecom, ultimately leading to smarter, more efficient, and dynamic telecommunications systems.

Highlights

• AI's pivotal role in telecommunication is optimizing network performance and predicting congestion 🚦.
• Enhanced customer service through AI chatbots reduces the need for human assistance 🤗.
• AI's predictive maintenance detects potential network failures early, saving costs 💰.
• AI and ML are reshaping telecommunications with smarter, more autonomous networks 🌐.
• Open RAN offers flexibility and cost savings by integrating multi-vendor components 📡.

Key Takeaways

• AI is revolutionizing telecommunications by making networks smarter and faster 🚀.
• Open RAN represents a breakthrough in telecommunications, allowing integration of components from different vendors 🔗.
• AI-powered customer service is enhancing efficiency, reducing the need for human intervention 🤖.
• Predictive maintenance with AI is preventing network failures before they occur 🔧.
• AI in telecom is paving the way for futuristic applications like smart cities and autonomous vehicles 🌆.

Overview

AI and ML are at the forefront of technological advancement in telecommunications, transforming how networks are optimized and managed. These technologies enhance network flexibility and reliability, offering faster and more effective service to consumers worldwide. From self-optimizing networks to predictive maintenance and fraud detection, AI is pushing the boundaries of what telecom networks can achieve.

Open RAN architecture is a game changer, allowing telecom operators to mix components from different vendors for a more flexible and cost-effective network setup. This initiative has quickly gained traction, enabling smarter network solutions and the potential for even greater innovations in the future.

The future of telecommunications with AI is bright, promising endless possibilities such as smart cities and autonomous vehicles. As AI continues to evolve, it will drive more personalized services and enhance network security, making the global telecommunications landscape more advanced and capable than ever before.

AI in Telecommunications Transcription

• 00:00 - 00:30 [Music] so welcome so to this uh session in AI in telecommunications honestly this is the first time that I will be presenting this AI in telecommunications uh this is for for
• 00:30 - 01:00 emergency purposes like this where uh plan is not happening especially specifically there should be another speaker that will be uh uh delivering the AI in uh uh in the industry but in the unfor events uh he cannot come so I'll be uh covering uh this session for you and I hope you could still learn even though you think of me as not as
• 01:00 - 01:30 expert in AI but uh we'll we'll discuss about AI in telecommunications all right so we're about to discuss how these Smart Technologies are transforming okay uh the way uh our phones or networks and even everyday Communications work behind the scenes No so just imagine you're streaming a movie or making a video call or chatting with the virtual assistant so everything feels seamless Okay so we'll examine
• 01:30 - 02:00 today the this uh AI technology how it is making in the how is it making our networks smarter and uh faster and how it's causing massive changes uh to the entire telecommunications industry okay and in this session as well we'll start with the look of everyday use of these Technologies and move on on how they are applied specifically in telecommunications and we'll go to something really exciting what we call
• 02:00 - 02:30 the open run okay and uh the main reason how this uh series uh or why is this series all about what is this series all about okay so I prepared uh few topics for you uh for today's session so introduction to Ai and machine learning or ml so I believe you're already have a very indepth knowledge about the AI and uh machine
• 02:30 - 03:00 uh uh language and deep learning or machine learning and deep learning coming from our one of the best u co-creators in Asia open Academy that is Liber okay and then applications of AI and ml across industry so where do we see um this Ai and ml are being applied in other Industries okay and then after that we'll focus on the Ai and ml in
• 03:00 - 03:30 telecommunications and then after that we'll go particularly uh the application of AIML in open run or open radio Access Network okay so by by the time we wrap it up you'll have a uh or an an understanding of how telecom companies are using these tools to build future networks and how the future is uh already here impacting our daily lives
• 03:30 - 04:00 uh in ways that we might not even realize okay and you will also learn that the learning artificial intelligence as early as today uh will give you benefit especially if you are looking at a technology company or a telecom company that are looking as well of implementing artificial intelligence in their Network and of course we have the summary and future of AI in telecommunications okay
• 04:00 - 04:30 so we'll start with the introduction to Ai and machine learning okay so we already have an idea on this one the basics so you know already the basics of AI so AI is about making machines act smart and while ml is teaching them to learn from data and deep learning is uh where things get interesting so with systems mimicking
• 04:30 - 05:00 how the human brain works definitions so we will not go deeper here so that we will not replicate what's uh being discussed by Liber so okay so the topics that uh he already covered are I think I I believe these are already sufficient for you to understand what is AI ML and deep learning okay so we'll move directly to uh applications of
• 05:00 - 05:30 AI uh ml across Industries so we have laid the groundwork so by discussing you the basics of AI machine learning and deep learning now we will move on into something bit more um um I will say exiting because on the applications no so seeing this uh Technologies in action across uh different Industries so we've all heard about transforming sectors like uh Healthcare or for finance and
• 05:30 - 06:00 sometimes security and then we already know that Ai and ml will become something like a versatile tool like a Swiss army knife for data so they are not also limited to one job they solve problems um what else optimize processes and even predict the future across a wide range of Industries no so we'll discuss the different uh Industries so last April of this year 2004 we had the
• 06:00 - 06:30 opportunity to visit uh the headquarters of Nvidia in San Francisco California so we have the we have visited the two main buildings of uh Nvidia if you are familiar with Nvidia so they are the one uh pioneering processes or equipment or tools or the technology behind artificial intelligence deor
• 06:30 - 07:00 and build St Tre St Tre uh inspired yeah so if you will see there uh there's a lot of applications they have already exploring uh using the artificial intelligence so what they have presented to us that time was the uh was the application of artificial intelligence in the medical industry and then uh after that they Pro
• 07:00 - 07:30 they presented uh an application about uh open run and Telecommunications and then during the visit as well to different uh Laboratories you will see an Nvidia GPU installed in the baseband unit or the or the U uh centralized or distributed unit in their Laboratories okay so most of the most of the videos that I will be showing to you in the session uh are coming from Nvidia as
• 07:30 - 08:00 well so it's a great place as well to learn more uh things about what are the different things that Nvidia are doing and then uh uh yeah let's see it so I'm not sure if you will hear the sound but I'll try my best to put it on the microphone so we'll start with uh with a video okay so it's uh one thing to talk about how AI is re shaping our
• 08:00 - 08:30 world is about to see it in action [Music] [Applause] [Music] right I am a [Music] [Applause] Visionary Illuminating galaxies to witness the birth of star Illuminating galaxies to witness
• 08:30 - 09:00 the birth of [Music] [Applause] stars sharpening our understanding of extreme weather [Music] [Applause] [Music] events I am a helper guiding the blind through a crowded world I was thinking about running to
• 09:00 - 09:30 the store and giving voice to those who cannot speak do not make me love I am a [Music] Transformer harnessing gravity to store Renewable [Applause] [Music] Power and Paving the way towards unlimited clean energy for us all
• 09:30 - 10:00 [Music] I am a [Applause] trailer teaching robots to assist to watch out for [Applause] danger and help save lives I am a Healer providing a new generation of
• 10:00 - 10:30 cures and new levels of patient care that I am allergic to penicillin is it still okay to take the medications definitely these antibiotics don't contain penicillin so it's perfectly safe for you to take them I am a navigator generating virtual [Music] scenarios to let us safely explore the
• 10:30 - 11:00 real world and understand every [Applause] decision I even help write the script breathe life into the words [Music]
• 11:00 - 11:30 I am AI brought to life by Nvidia deep learning and Brilliant Minds [Music] everywhere all right okay yeah so that is uh I am AI know
• 11:30 - 12:00 from Nvidia so you will see that uh Nvidia has been dri has been a driving force in advancing AI helping bring it from concept to reality so so in that video you saw that there are some incredible examples of AI at work across Industries like uh the health care science art and Beyond so and then in this session we will talk about telecommunications so it demon demonstrates how AI isn't just a tool
• 12:00 - 12:30 it's becoming an innovation partner so helping us solve some of the biggest challenges okay so now let's uh have to let's look at this uh boring and boring slides so when we talk about artificial intelligence and machine learning okay it's uh it's easy to get caught up in in technical jargon so but the truth is AI and ml are already woven in the fabric of our daily lives
• 12:30 - 13:00 so they're also making huge impact across Industries so you will see that the different uh uh applications for example in searching and ranking so AI is uh making it easier and faster for us to find information online by ranking the most relevant results um search engines like for example Google has the sophisticated um AI algorithms to to analyze massive
• 13:00 - 13:30 amounts of data no in Split seconds so presenting us with the most accurate answers so for example if uh uh every time you search for something online so AI Works behind the scenes No to understand your query and prioritize the most helpful results so that is AI already um recommending okay so as liary mentioned a while ago if you're just browsing s Viber and and then suddenly
• 13:30 - 14:00 check Facebook or Lazada so there's a an ad that is already recommending something for you to purchase is a or okay so recommending so recommender systems are another common AI application used for almost everywhere from online shopping um streaming services uh these systems are analyzing your behavior and suggest content or
• 14:00 - 14:30 products that uh that is matching your for your preferences example Netflix Amazon okay so just just imagine when you're watching Netflix it will suggest you videos now since you are watching this movie we are recommending this kind of movie so that are recommender systems okay so don't be surprised new is uh a certain kind of
• 14:30 - 15:00 topic no because that is powered by recommender systems AI um detecting threats and fraud okay in terms of Finance so AI is also critical tool in cyber security so it can detect anomalies um in data patterns to flag you potential online attacks frauds and other malicious activities for example Banks they use use AI to
• 15:00 - 15:30 monitor transactions and detect fraudulent activities in real time so they're helping to protect your accounts from unauthorized actions okay and then we also have ai in finance from automating Trading strategies to offering personalized fin Financial advice so AI is also helping uh the financial uh Finance Industries um if you if you heard about the AI powered Robo
• 15:30 - 16:00 advisors to analyze financial data and provide recommendations for investment Port portfolios so they're also using AI um research and unraveling complexities if you have seen research okay so there are also um areas so unraveling proteins and exoplanet search Nvidia so AI is also has the
• 16:00 - 16:30 ability to process massive data sets so this is uh the one that is no um um AI is getting more powerful than human because uh it has the capability to process massive amount of data okay and then it can also has it also has ability to calculate um massive amount of permutations no um beyond the capabilities human brain yeah and in in a faster uh uh
• 16:30 - 17:00 sense medal okay medical and digital doctors Internet of Things iot smart Services um sensor AI okay so enabling machines to process sensory inputs like sound uh site blind person can navigate just like humans do BL is a human right I
• 17:00 - 17:30 mean okay um understanding words and speech this I think this one is one of the um common that we are using uh voice assistants like Alexa and Google yeah uh Google assistant and then um AI in companionship okay um movement and Mobility uh driess cars okay and uh in
• 17:30 - 18:00 some cases we don't want to use it uh actually AI in Ware no in defense yeah okay all right so this are the different applications no of AI in uh or ml across different Industries so there's a lot of applications as you already saw s video Nvidia so now let's focus on the uh applications of AI and ml let's zoom in
• 18:00 - 18:30 telecommunications yeah so we have explored applications different Industries so now we're going to uh zoom in telecommunications so in telecommunications right uh uh there's a lot of processes and there's a lot of activities and as a typical citizen we rely on Telecom networks every day whether it's for streaming a movie or
• 18:30 - 19:00 making a video call or sending a simple text message so uh what is powering the seamless connection and fast data speeds uh behind the scenes might be artificial intelligence probably in some networks they are still using um Engineers no uh but in some Advanced networks like what we saw in the US and in the in Japan they are also they're already using artic icial intelligence so
• 19:00 - 19:30 telecommunications is an industry where speed um reliability and adaptability are key and the uh artificial intelligence or machine learning are playing a critical role in making this modern Network smarter uh more efficient and more responsive to this ever growing demand for connectivity so different uh um applications no so from optimizing Network performance performance to enhancing customer service uh the eii
• 19:30 - 20:00 and the or ml are reshaping the way our Telecom networks operate so let's start with for example AI in network optimization so we'll compare it to today's uh or in some in my in my uh previous experience as a as an optimization engineer as well so traditionally uh managing the Telecom networks involved manual process so manual procedure like monitoring traffic
• 20:00 - 20:30 so you are checking the kpi then you see if you are if there's a trending traffic of or the traffic Coast down or congestion okay so that is monitoring traffic so most of the time we are use we are doing it manually you get it from the OSS and then for some Engineers they're just using Excel no for some Engineers they're using uh programs to to to make it uh faster no in in terms of analysis but uh in some networks they
• 20:30 - 21:00 are already using artificial intelligence to do this traffic monitoring and then troubleshooting issues okay what are the issues uh why are there a lot of drops happening in the network why there are congestions in this sector why the second sector is not congested okay so there's a lot of troubleshooting issues and most of the time those are being analyzed or being uh trouble troubleshooted manually and then if you want to adjust
• 21:00 - 21:30 bandwidth so you need to change it manually so you need to configure it in the OSS that to change the bandwidth from 5 MHz to 10 MHz to 20 MHz now with artificial intelligence or AI networks can adjust in real time based on demand okay so you will train the system or you will train the artificial intelligence on how to respond on a realtime basis so AI will analyze traffic patterns predicts
• 21:30 - 22:00 congestion and optimize bandwidth on the fly to ensure smooth service okay so it's something like um self-optimizing Network so we had that term previously the son or the self-organizing network so this uh AI for Network optimization is a higher version of that okay so for example inuring Peck times so from 600 p.m. to 8:00 p.m. so traffic EDS okay so AI can
• 22:00 - 22:30 automatically redirect Network traffic to prevent slowdowns ensuring that the user streaming video or gaming have seamless experience okay so traffic steering optimization next is preventive maintenance and Fa detection so one of the key AI Keys benefit in telecommunication is its ability to predict and prevent Network failures so the machine learning models um can
• 22:30 - 23:00 analyze data from Network components like for example cell towers um you can detect early signs of where and te allowing Engineers to fix uh issues before they lead to outages like for example the one that I posted so you can predict to so so those are the examples of predictive maintenance and F detection
• 23:00 - 23:30 so AI uh driven predictive maintenance uh will help Telecom providers avoid costly downtime by identifying equipment failures before they happen Okay and next is one of the no common application uh in telecommunications AI powered customer service so AI has transformed the customer service um uh in telecommunication so you will see chat that powered by AI that can handle
• 23:30 - 24:00 queries like troubleshooting or Bly issues providing quick and efficient responses so slowly transition from Human to AI uh chat Bots okay so telecom companies they use already AI chatbots to assist customers uh with routine questions reducing need for human intervention and speeding up service okay
• 24:00 - 24:30 um next is network security and fraud detection so with increasing uh cyber threats uh just like what's happening here in the Philippines uh Telecom operators are turning to AI for network security so in which AI can detect unusual um activity or anomalies in the network traffic that can or that may indicate fraud or Cyber attack so by identifying these areas or threats early so AI can help protect users from malicious actions okay example AI based
• 24:30 - 25:00 security systems that can block suspicious transactions okay and detect unauthorized access sensitive information in real [Music] time system AI has identified you as a malicious or suspicious uh transactions or system um AI in open run or in 5G no in
• 25:00 - 25:30 5G uh in 5G networks so as 5G networks continue to roll out globally uh actually when you're deploying 5G as we discussed last week um you will also realize that deployment of a new technology becomes complex because it involves a lot of frequency a lot of equipment a lot of features a lot of processes no and sometimes even an engineer even a su or a magnl or a top
• 25:30 - 26:00 one ECE uh engineer cannot really handle it alone okay you need some tools no to manage the complexity of a network okay so as the network or the 5G Network continue to roll out globally so managing this complexity will require artificial intelligence or AI That's the reality and then AI will help with the network slicing okay where the part of the networks are reserved for specific
• 26:00 - 26:30 services like iot the mmtc right URC and embb that you will ensure the right resources are allocated no where they are needed okay so those are the typical applications of AI and then um later on you will see the application of artificial intelligence in open run so one of the most exciting applications uh in AI is uh in telecommunications is within the open run system okay so open
• 26:30 - 27:00 run is new and open architecture for radio access networks that allows Telecom operators to mix and match components from different vendors so that is open run in nutshell uh this flexibility will open up new possibilities for Innovation and cost savings okay so we will discuss later on what is the Run intelligent controller the Rick and the two types of uh Rick the non-real time Rick and the near real time uh rake so as an example for the
• 27:00 - 27:30 application of AI in open run uh open run can automatically detect um Network congestion in specific area and shift the resources to prevent delays and then ensuring uh experience seamless uh on those Services no with the higher demand traffic steering or current setup is the load balancing load balancing but
• 27:30 - 28:00 in opener we call it traffic steering and qos optimization or the quality of service optimization so those are the different applications of uh Ai and ml in telecommunications and then there is a particular test uh um process in uh in telecommunications where radio network design is very uh is essential no uh
• 28:00 - 28:30 currently when we do uh radio network design so we do simulations we simulate um the coverage of a specific site uh either 4G or 5G and then and then what would be the impact of that uh simulation no in in certain area and that depends on different clutter types it could be a rural area a suburban area a dense Urban
• 28:30 - 29:00 or an urban area so we all know that the the effect or the performance of a cell site or a network depends on one of this one of the one of the factors is the Clutter area no it depends on the surrounding area and it also depends on the type of subscribers the type of the users or the type of the user equipment they're using okay so simulating this
• 29:00 - 29:30 kind of scenario sometimes requires a lot of or not sometimes most of the time requires a lot of time especially if you are using a not so powerful device so I remember when I was uh in a in a telecom operator when we do a simulation of a nationwide uh coverage now for example for the entire Philippines you will leave the server for probably 24 to 48 hours before it completes the entire
• 29:30 - 30:00 Philippines it's only simulation it's not even considering the Monte Carlo simulation if you're aware of Monte Carlo simulation that's like uh simulating the performance of the network if there will be congestions what will be the expected throughput where are the possible location of subscribers no so the Monte Carlo is just like uh it's name after a casino no Monte Carlo in Monte Carlo so you are trying to roll the the dice and then what would be the possibilities
• 30:00 - 30:30 performance possibilities of your network if you do that in a in in a typical server uh right now uh it might take you 24 or 48 hours to do it that's why um one of the um um significance of using uh GPU no uh in your one of your tools will simplify and will make your processes uh more
• 30:30 - 31:00 efficient okay so now we're we're uh I I'll be showing to you a a deployment of a highly performant uh 5G Network um model or a modeling and inter uh of a 5G radio waves no okay so we need this to propagate to to simulate the propagation of a 5G radio waves across real Dynamic World environment with hundreds of transmitters and
• 31:00 - 31:30 thousands of receivers that will be affected by physical factors no so it also it will also um simulate the accurate 3D Sim uh 3D simulations of the buildings as you will see in the in the in the video so the one that is performing this is Ericson so Ericson is the one developing a new approach to traditional 5G deployment so building a city scale physically accurate we call it digital twin so you will hear the
• 31:30 - 32:00 digital twin uh terminology and this is uh being helped by uh an Nvidia product they call it the Nvidia Omniverse okay so I'll uh I'll play the video and then hopefully you can hear the the know the the sound again oops sorry there are 15 million 5G microcells and
• 32:00 - 32:30 Towers planned for Global deployment in the next 5 years ER inidia obers to build digital twin environments to help determine how to place and configure each of their sites for the best coverage and network performance in Omniverse ER bu City scale models that are physically accurate down to the materials of the buildings vegetation and foliage then wireless network components are added including the precise location height and antenna pattern of each transmitter
• 32:30 - 33:00 Ericson built a custom Omniverse extension enabling them to integrate radi propagation data and leverage Omniverse RTX accelerated Ray tracing to quickly visualize and calculate the quality of the signal at every point in the city because Omniverse materials are physically accurate the intensity of Reflections are precisely determined antenna beam forming and Signal paths can be accurately simulated and visualized in the simulations the Lo signal IFI transmitter antenna beam forming and the straight lines are
• 33:00 - 33:30 signal paths the colors of the signal paths denote strength in Deb and data through point from Blue is the weakest to Red is the strongest visualization is a critical to theity for ER with Omniverse VR Network Engineers can virtually explore any part of the model teleporting to any location anywhere in the world at 1:1 scale as they tune the network for Optimal Performance or identify path disruptions they can literally see the effects of their adjustments in real
• 33:30 - 34:00 time things which aren't visible in real life in Omniverse ericon can perform true to reality remote simulation of entire 5G networks enabling them to design more efficient and reliable networks conduct remote field trials and speed of deployments [Music] deoy okay so that was uh the Omniverse no application uh in a 5G deployment uh of
• 34:00 - 34:30 of uh a 5G simulation deployment for Ericson so as you can see no it's not only just uh like using artificial intelligence CH no so for telecommunications and vendors and ICT vendors they're using this one for the tools so just imagine you can somewhat visualize in real time about the wireless telecommunications or spectum performance system as you can
• 34:30 - 35:00 see there although awward movement signal when you pass through one antenna or a massive myo antenna but it's a good thing that uh you can simulate those uh those uh data and the one thing that is uh being uh or good about using artificial intelligence that you can process a lot of data or massive amount of data and then you can get the results um as fast
• 35:00 - 35:30 as while you are doing the testing if you are if you are using some some technologies like that and they're using a very low uh Power processor it will take you probably weeks before you visualize those things so that's the power of artificial intelligence uh in the radio network planning now let's explore some how AI is uh revolutionizing the deployment and
• 35:30 - 36:00 management of 5G networks so how it will affect no um uh managing 5G networks and how you how in the future you can also decide of deploying 6G technology okay so the next video uh we're about to watch will highlight this uh AI powered solution developed by heavy heavy Ai No and Nvidia that could uh help Telecom operators build and manage their
• 36:00 - 36:30 networks so to keep up with this uh this kind of demand uh Telos are expected to deploy uh a big a huge amount number of around 17 million 5G microcells and Towers over the next 5 years so however the complexity uh will give us a real challenge okay and uh the real challenge actually lies in building these towers and optimizing their deployment and operations because
• 36:30 - 37:00 as I mentioned 5G or future uh s sites so with the help of AI uh telecom companies can save uh probably billions of dollars uh worldwide in a $2 a$2 trillion dollar industry yeah improving you efficiency and performance at this unprecedented scale okay so you will see in this video uh
• 37:00 - 37:30 how uh AI driven Solutions like that uh heavy AI NV has developed no and how uh Telos will probably in the future will plan and manage their 5G Network okay so I'll play the video again hopefully uh you can hear the sound Telos plant
• 37:30 - 38:00 deoy over 7 billion 5G microc cells and Towers worldwide by 2025 optimizing and managing new and existing infrastructure while maximizing customer experience is increasingly complex heavy AI built an AI accelerated application framework on Nvidia Omniverse that enables Telos to develop physically accurate interactive digital twins to plan build and operate four in 5G networks at Nationwide scales
• 38:00 - 38:30 to create a digital twin heavy AI used multi-billion Point lar scans satellite imagery environmental data and network parameter changes they paired this with open and proprietary data sets of network supply and demand to enable Telos to better understand Network reach and performance in real time for Network planning heavy ai's RF model in the digital twin lets Telo simulate optimal infrastructure deployment based on realtime radio wave propagation with
• 38:30 - 39:00 extreme accuracy unlike Legacy Solutions heavy AI combines Precision timed signal metrics with demographic and customer Behavior data giving Telos immediate insight into the effects of planning and operational decisions Beyond planning digital twins give Telos enhanced operational abilities letting them see issues that occur in real time or predict areas needing attention this reduces maintenance costs and Max maximizes customer satisfaction even at massive scale the
• 39:00 - 39:30 heavy Ai and Omniverse powered digital twins are fully interactive enabling Telos to test and validate scenarios before investing in physical changes Telos can leverage heavy AI extensible application framework as an integrated solution or develop their own applications on the platform heavy Ai and Nvidia omers are introducing a new era of real time interactive Precision digital twins enhance Network planning building and operations
• 39:30 - 40:00 operations okay now so that is heavy Ai and inidia so you can see uh how AI is uh also transforming your radio network planning and design So currently when we are doing those kind of simulation it's only 2D uh in some cases we can perform 3D but it's not like that kind of 3D view no which is very no very rich in data and then you will see that there's a lot of uh data integration happening
• 40:00 - 40:30 so satellite um user uh behavior and then CR sourcing data informations and then you will see that almost instantly it simulates and predicts performance site on one from one location to another location process so that's one of the
• 40:30 - 41:00 applications a telecomunications so I will be sharing as well to you this state of AI in telecommunications so if you want to um um learn more about the state of AI in telecommunications so Invidia has uh released this kind of report no state of AI in telecommunications in 2024 trends so it examines how AI is
• 41:00 - 41:30 being used to address some of the most pressing challenges in telecommunications uh the report explores you key trends real world use cases and future projections of how AI will continue to evolve within the industry okay so this report also highlights the key Sur uh survey insights that Define the state of AI in telecommunications in 2024 it shows you different AI adoptions especially generative AI uh with notable increase in AI
• 41:30 - 42:00 enthusiasm at the executive level so so one of the topic that is being discussed here in the state of AI in telecommunications um uh survey report they're asking which of the following is your company investing in today no so is it machine learning in 2022 deep learning generative AI high performance Computing and and digital Twins or metaverse as you will see uh there's a
• 42:00 - 42:30 increase from 2022 to 2023 about the interest of investing in deep learning um generative AI there's no comparison from 2022 um and digital Twins or meta bur so this the the survey respondents represented a mix of T cost no including Mobile Fix and cable companies and their hardware and software so suppliers so the respondents are represented or
• 42:30 - 43:00 representing a broad mix of different groups like customer service marketing and business development and the results include feedback from manager level industry uh in the in the uh in uh in the professions including senior Executives um next is uh we go to AI in open run yeah so open run so now that we have covered how Ai and uh ml are transforming telecommunications now let's focus on this uh uh specific uh
• 43:00 - 43:30 application in telecommunication with is the open run so open run or open radio access network is a new way of designing or implementing it's a new paradigm or a new philosophy of designing or implementing a telecom Network it will also allow you to use different components from different vendors and then giving them the greater flexibility and lowering the cost is essentially okay so what is open run know okay so
• 43:30 - 44:00 open run is uh yeah radio Access Network okay so if you compare it to a traditional radio Access Network or Sal site a traditional uh run architecture uh Telecom operat used to connect individual devices to the core Network so it involves primarily proprietary hardware and software from specific vendors making it a tightly integrated and often uh expensive solution now open
• 44:00 - 44:30 run is a recent Innovation so it aims to revolutionize the traditional run architecture it promotes using open and interoperable hardware and software components from multiple vendors so open run decouples the hardware and software allowing for greater flexibility and cost savings and Innovation so P opener you can do a mix and matching from one vendor to another vendor so for example example you have a radio or one Dru you
• 44:30 - 45:00 can connect it to the different CU on ndu that is the main um um objective no but in some cases you you will see or in some countries you will see that they're um implementing open run as a single vendor but the system is open run so it depends on the operator but the bottom line is it the op open should meet the operator's business model so what kind of business model will uh be effective when he implement or
• 45:00 - 45:30 when they uh decided to implement open run because you can Implement open run in numerous uh architecture ways probably you will have a centralized uh CU sitting on a cloud or a du sitting on a cloud or there could be an implementation uh like dish uh in the US where the ru the and the DU sits on the cell site premises and the CU is centralized and located in the cloud so it depends on what type of uh
• 45:30 - 46:00 architecture that will fit in your business model requirement okay so using standardized interfaces and protocols uh open run will enable a more competitive and diverse ecosystem okay so it will reduce dependency single vendor okay when we say single vendor meaning currently what's happening in the Telecommunications uh network is that uh um from the base band unit to the uh front hole to the remote radio
• 46:00 - 46:30 head or remote radio unit this is coming from a single vendor probably coming from vendor H vendor Zed vendor e or vendor n but in open run like for example in in uh in the Dish Network they have this um um the CU the D or the DU or the CU coming from uh a commercial offthe shelf servers like Dell and then they install a software of mineer or on
• 46:30 - 47:00 this uh dell Hardwares and then on the radio side they can uh use mavenir radio and Fujitsu radio it's their choice no they are not uh tightly coupled or um locked in into a single vendor so you can mix and match different vendors okay so what is open run no so open run is uh or an alliance is the one um that is um the key players no in
• 47:00 - 47:30 producing the guidelines and specifications for open run uh system so Oran Alliance is uh a group of telecom companies that aim to improve the performance and openness of their Network so it's a group of telecom companies so they call for more intelligence in the network through information collection of these virtualized Network elements and artificial intelligence and machine learning can be applied to this collected uh information now let's uh
• 47:30 - 48:00 see Oran Alliance in deployment or producing these kind of guidelines and uh where they um use artificial intelligence uh in terms of open run transforming radio access networks to be open intelligent virtualized and fully interoperable is not a far off vision is not a far vision o ran Alliance has made this Vision real since 2018 the adoption
• 48:00 - 48:30 of the O ran Alliance by the industry has been incredible it has grown to a worldwide community of operators vendors and research and academic institutions jointly developing a sustainable oan ecosystem through expanding traditional mobile networks the oan alliance has created its oan architecture to become a foundation for truly open and intelligent radio access networks the oan architecture improves the ran in
• 48:30 - 49:00 several areas first the oan architecture supports disaggregation of software from the hardware and network virtualization underlying open reference design hardware and the O Cloud provide a high performing and scalable computation layer second the oan architecture is based on well-defined open interfaces to enable a multivendor interoperable supply chain ecosystem third the oran architecture
• 49:00 - 49:30 empowers mobile networks through artificial intelligence and machine learning oan networks can self optimize to ensure the best possible service for the end customers and automated operations and maintenance enable more efficient Network rollout and operations defining the complex oan architecture is a step-by-step process resulting in hundreds of technical documents to ensure compatibility and to avoid duplication of work the oan specification effort Builds on common standards and cares about alignment with
• 49:30 - 50:00 other industry bodies oan specifications serve as a basis for open software reference design the oan software Community host developers brought in by many Oran players cooperating on tens of software projects oan specifications in open software are enabling a broad industry ecosystem companies of different sizes and competencies can use them to create their innovative and competitive open Solutions in addition to specifications
• 50:00 - 50:30 oan Alliance supports the ecosystem in testing and integration oan Global plug fests allow companies to demonstrate component functionality as well as multivendor interoperability moreover the open testing and integration centers around the world will offer a collaborative and impartial testing environment all Rand supports implementations using open source modules proprietary modules or any combination operators and suppliers can create different paths and approaches to
• 50:30 - 51:00 implementing or optimize for their specific Network and customer needs the rapidly emerging or ecosystem has enabled service providers on multiple continents to start field trials and Commercial deployments aiming to build significant parts of their networks based on or technology open intelligent virtualized and fully interoperable oan powered mobile networks will soon be serving billions of people around the world the oan
• 51:00 - 51:30 alliance Alliance okay so that is Oran Alliance so as you can see you Oran or open run um uh system is uh uh started to deploy in different parts of the world and some of the vendors or operators are now exploring to deploy open run okay now if you if you want to learn more about open uh Oran Alliance so link so I will also
• 51:30 - 52:00 share it with you ww. or.org and then you can also see the different places or different uh countries uh started implementing open run hopefully in the Philippines we could see some applications uh here uh as well in the country okay okay so we'll uh need to uh Fast Track discussion so this is logical architecture of Oran so where in this
• 52:00 - 52:30 architecture you can find the artificial intelligence machine learning you can find it here in the non-real time Rick and the near realtime Rick okay so these are the Run intelligent controller [Music] soal and then if you look at the actual open run setup so this a typical example from mavenir so you have the CU and the DU you have the radio and then you have the antenna so that this type of antenna is like an omnidirectional antennas but there are different types of antennas as
• 52:30 - 53:00 well okay so what are the different applications of uh uh of artificial intelligence or U use cases uh in the Telecommunications uh using open run so we have the traffic steering so this an evolution of Mobile load balancing that we have right now so traffic steering is that um you steer traffic now from one cell coverage to another cell coverage it depends on the operator strategy on how to handle traffic uh uh steering
• 53:00 - 53:30 okay okay so this is where the Run intelligent controller will come in so the Run intelligent controller will Define um the timing and how fast it could uh handle traffic steering procedure system okay and then the other one is the quality of service use case so uh the network will prioritize in real time uh the quality of service or
• 53:30 - 54:00 or the high priority users so VIP users or Emergency Response Team so they have a higher priority system so for example the Philippines uh radio resources for um firemen or respondents should be
• 54:00 - 54:30 prioritized bloggers creators so that's qos no so what's the future of AI telecommunications yeah so AI will uh will have a continuing role AI will drive the evolution of Telecom with smarter uh more autonomous networks that adapt to Growing demands and then it
• 54:30 - 55:00 will also have a hyper personalized services that will deliver tailored bandwidth and service um to users based on individual needs um enhanced network security okay uh intelligent networks for future so it will uh power networks that support smart cities autonomous vehicles and Next Generation industrial iot applications and of course there's a lot of endless possibilities probably the possibilities right now we cannot see it
• 55:00 - 55:30 in our lifetime no I see in in the Q&A or chat box that there's a lot of futuristic possibilities that they're thinking actually I cannot uh tell if that will happen in our lifetime uh but yeah probably no uh because the possibilities are endless so the integration of AI in a current network uh for example in open R probably in 6G in 2030 will lead to networks that are more Dynamic reliable and flexible than ever
• 55:30 - 56:00 before okay so yeah I'll end this uh session with the summary okay so we have discussed about the artificial intelligence and how they enable machines to mimic human intelligence okay we have check also applications of AIML across Industries uh application of AI and ml in telecommunications and in radio access and network particularly in open run and possibilities of uh AI in Telecom uh
• 56:00 - 56:30 industry in the future