Thank you very much for attending NEC Innovation Day. This year marks our third year. This year, not just NEC's CTO, but our CDO, Yoshizaki, will join us to explain our state-of-the-art technology, our activities for our new businesses, and R&D activities to drive advanced technologies and creation of new businesses. Now, the flow for today's presentation. First, I will explain NEC's innovation creation scheme, then strengths of NEC's R&D, the contribution to current businesses through global number one technologies. For NEC, this means our IT service business and social infrastructure business. In the area of IT service business, we made a major investment back in July, along with Yoshizaki, in regards to generative AI, and I would like to make our next new announcement. Finally, creation of new growth businesses. This is also something new, and currently, the commercialization of IT licensing business is doing quite well. I will report to you the progress. Then I will give you updates since last year regarding expanding business innovation, including healthcare and life science, along with corporate values. This is a structure to create innovation, NEC's organizational structure. Along with business units covering each business area, we have common units underpinning this. One of them is the Global Innovation Business Unit. As a strong characteristic of NEC, we have corporate R&D, new business development, and IP strategy unit working as one. IP that is created from these three units, seeds for new businesses, as well as the so-called essence of various businesses, is offered to existing businesses to strengthen competitiveness. We also offer outcomes to the next growth business as well. In terms of our larger global structure, we have seven R&D sites in Japan and around the world. One point that stands out is that NEC's overseas R&D sites are extremely large. This has been a tradition since 1990, and 40% of our researchers are outside of Japan. We also have another four sites focusing on new business development. Globally, we are exploring or cultivating the possibilities of new businesses. Units that supports these endeavors, our expert team, research centers, business acceleration, IP, all need expertise. This means that it is critical to grow or secure such talents. We have 2,000 specialized members on board, and they are not just doing their respective jobs. They are also involved in company-wide R&D, business acceleration, offer training about IP, create impact, and offer professional services in expert domains. This is the strength of NEC's R&D. For high technological competitiveness on a global scale, we must have a platform technology, and this will be reflected on academic records. NEC has been concentrating on this recently. Let me give an example. The number of papers accepted at leading academic conferences for machine learning for 20 years in total, NEC has been in the top 10. Other ranking companies are GAFA and many Chinese companies. NEC is the only constant Japanese company ranked. The leading academic conferences are listed here, which I believe are self-explanatory, and we also have a similar track record for video image processing, security, and communication. This is a platform of our strength. What about patents? A strong technology is reflected in a strong patent. In terms of numbers for biometric authentication, video recognition, and analysis, prescriptive AI, we are global number one in the number of accumulated patents filed. We multiply the number of U.S. patents with right holder scores to benchmark how we can impact our business. For these three areas, we are aiming to attain global number one in 2025. When I present this outside of Japan, I feel the strong impact of this message. Especially with biometrics, face, iris, and fingerprint recognition, there is a specific institution that conducts benchmarks, and NIST in North America has consecutively acknowledged that we are world number one. Moving into the future, in terms of researchers and the quality of the research itself, it will be crucial to have a computing facility to realize this. We boast the largest AI supercomputer in Japan. We have 1,000 GPU units. Our investments began back in 2021, and parallel operations accompany many challenges and demands know-how. March 2022 and then on to 2023, it took two years to conduct parallel operations of 1,000 GPU units. This will allow the acceleration of AI R&D. For instance, generative AI was a technology breakthrough that happened a year ago. NEC caught up and launched in a matter of a few months, and we now can make more new announcements. This is the underlying platform. For such advanced technologies, we at NEC allow portable usage in various businesses, and this is possible because it is implemented on our NDP, NEC Digital Platform. Now, to contribute to current businesses leveraging our global number one technologies, I would like to explain this with our IT service business. Down below, you see our various businesses, although all are impacted, but I will highlight IT services…. This is our technology vision introduced last year. Please assume that all of NEC solutions are mapped. We observe data from the real world, utilize digital twins, and conduct modeling in the virtual world. If we can further do simulations, this will allow the implementation of optimal algorithms. We can abstract on what the optimal control should be and return that to the real world. Creating this continuous cycle is NEC's approach, and the fact that NEC is underpinned with computer, communication, AI, and software technologies remains unchanged. I believe that all of you are aware that the world of AI is entering the era of foundation models. Machine learning itself carries extreme value, but the new generation of AI, the foundation model, emerged exactly one year ago. To understand the impact of the foundation model, it is best to make comparisons with the Internet. Consider the days before the Internet and after. Many services emerged at extremely low costs, and you saw the GAFAs making many service offerings. Many new applications were launched on the Internet. Without the Internet, there was no such concept. Now, with the Foundation Model, it's the same story. AI models used to cater to the extremely specialized. Now anyone could create them. What will the impact be? I have served as an engineer and researcher for many years, and when I consider epoch-making moments in life, I believe I have witnessed two: the emergence of the Internet and the Foundation Model. How will then the Foundation Model impact NEC's technology vision, inclusive of Digital Twin technology? We believe it will contribute to further advancements, automation, and scaling of digital systems. However, the overall structure remains the same. It contributes to a major update or advancement. Next, allow me to explain in more detail how it will impact NEC's digital systems. These five pillars are extremely important. First, the proprietary Foundation Model. Foundation models are not limited to just GPTs like ChatGPT. It is about providing reason and intellect and logic that humans think about in the form of software. Now, this is cloud-enabled, but in the future, it will be implemented everywhere, including devices. There is huge potential. It has the potential for various applications. That is all the more why that we believe that if you do not have the competency to use it or unable to hack it yourself, we cannot provide the next generation product. With the announced Foundation Model, the LLM size can scale flexibly from small to large. We are not limited to just the various specialized LLMs. We can combine our multimodal face recognition engine and image recognition to create a new architecture and craft a purpose-specific AI model, so the flexibility of the model is quite high. Second is Multimodal AI. Audio, image, and face recognition, sensing technologies are all important for NEC. This can be combined with LLM. This allows the understanding of the real world at the same level as humans or perhaps understand better than humans and process events. At the exhibition, many cases will be introduced. Next, for safety and security in the LLM era, it is true that advanced cybersecurity can be considered, but for LLM itself, for instance, hallucinations—are there any lies? Is the response in accordance with ethics? Was the data used valid? Was any personal information leaked? We must address security. Needless to say, NEC will ensure safety and security. With automation of system construction and operations, we also need to make sure that software development efficiency is realized, along with reducing power consumption. Now, lastly, for orchestration function, we started this at the beginning of the year. To actually use such AI technologies, each process needs to be broken down into tasks automatically and in accordance, linked to AI models. This will lend attention to networks and security. LLMs will be on devices and in the cloud, and these will be connected. This is where orchestration is needed, and this will be the prerequisite for various real-world processes to become automated. So the overall structure, our technology vision, remains the same, but with the foundation model and AI orchestration, five areas of key technologies will be used here. And today, we plan to show you seven actual demonstrations afterwards. But let me highlight that a few of them now. The first is what we announced in July, NEC's large language model, 13B. In comparison to our peers, 1/13 in size is quite small. That is why it can be used in on-premise servers, and it's so small that it can probably fit into next-generation device processors as well. These are areas that we have enhanced. Please refer to the left. The data volume that we introduced back in July, we now have double that data, high-quality data, and Japanese data. Now, as a result of that, note in the middle, Japanese language processing has improved, and when compared to other LLMs, you can note that it outperforms. Please note that scoring is the dark and light blue bars combined. The next technology breakthrough is next, processing longer text. In other words, longer prompts... we can handle up to 300,000-character prompts. One challenge of LLMs are, although they are good at handling what is in the system when it was created, there is no update data, or the LLM is created with general data available on the internet and does not reflect individual company data that the customer may have, or any personal data. To train further will take time and money, but using these long prompts will overcome this challenge. So in enabling on-demand with 300,000 characters, has enormous potential. So putting it simply, it is equivalent to 15 paperbacks or quite many new hardcover books. So something similar in volume, unlike customer manual documents or information, updated information can be utilized, and an on-demand specialized LLM can be realized. The core technology to support this, and we are still at our R&D stage, is the flexible combination of the large language model and accommodating AI. The scale-up technology noted on the left is leveraging multiple 13Bs. We are in the midst of developing a control function for large-scale model 100B. In the center, you see NEC AI and third-party AI, and we have the function to combine these. As needed, we can also allocate third-party LLM or industry-specific LLM and combine. This will foster an open innovation type connection as well. The third is the control function. Small scale LLMs can be used at the edge or in devices and put on the cloud, and this function collaborates all of this. As a system vendor, this implies that we are currently challenging ourselves with an extremely flexible architecture. Here are some examples. The first is a combination of video recognition and LLM. NEC's strength is video recognition AI, which takes real-world video, generates text, which describes it, and then converts it into meaningful information into text by LLM. For example, we can analyze image data from a huge number of cameras, identify all possible causes of, for example, a traffic accident, and generate a written report. This is real-world awareness, not just image recognition. In addition, a textual query of all relevant images is possible as well. This is truly an example of what is possible with the fusion between image processing and LLM processing. On the right is a disaster management solution. Four years ago, we introduced a data-driven DX solution called dotData Insight. We have also linked this with LLM. In the past, only experts in data analysis could understand things like feature amount and the parameters, but by using LLM, business insights can be obtained from data analysis without being an expert in data analysis. Business hypotheses can be refined in an interactive manner. Business line leaders and managers can understand the outcome without having a scientific background. For example, for cybersecurity, we have developed a security-specific LLM. Traditionally, cybersecurity has been an area for specialized experts. However, we have developed a new security-specific LLM. We have a concept called NEC Zero User, which means that we will build a product first for the internal use. According to the report, vulnerability assessment requires a high level of expertise, but we can save up to 80% man-hour. Security tasks, such as attack routine, attack route analysis, and countermeasure planning, can be done with half the resources. The figure on the right-hand side show HPC technology that accelerates data analysis. This is a technology originally developed by NEC. For example, table processing and deep learning processing is 4-16 times faster. This technology is also used in AI drug discovery. We are involved in the development of a vaccine for betacoronavirus, together with CEPI. With CEPI, we are doing this. We can simulate a vaccine in one week instead of three months. A beta version of this technology is already available. Now, Mr. Yoshizaki, our CDO, who has been working on this area for many years, will talk about the future direction of this AI. Mr. Yoshizaki, please. Yes, thank you very much. I'm CDO Yoshizaki, and I also lead NEC's hardware, software, network, product, and IT service units concurrently. Very pleased to talk to you. Let me first explain about this slide. As mentioned earlier, NEC announced the Japanese language LLM for the first time on July the 6th. At the same time, we explained about the service launch here in Japan. The photo on the right-hand side is the photos from the press conference. I'm sure that some of you who are here today were also attendant at that meeting. Thank you very much. ... Well, going back two years ago, we developed the LLM using our supercomputer at the research center. On July the sixth, we also announced the services of the data center at Inzai Data Center, which I will explain later. Personally, I speak with global leaders at the Davos meeting, which I attend every January. The last time, last year, I went there with President Morita regarding the foundation model. Mr. Nishihara mentioned several times before, we realized that, we will need a general purpose model and also specialized model for each country, each region, and each industry. So we at NEC, we decided to focus on the specialized model, and we made an announcement. That was the background. That was how it started. Let me now talk about how NEC plans to promote generative AI and how we did it in the past. This is what I wrote in handwriting, and this slide presents the services of the data center at Inzai. As I said earlier, we will focus on the Japanese language model of NEC's LLM and the industry-specific model at Inzai. On the right-hand side of this diagram, you can see hyperscalers. Those are global companies. People can use those global competitors' LLM, or they can use our LLM. We anticipate that our customers will choose one or the other according to their needs. That is our intent as well. For example, the defense industry or government agencies who handle sensitive data can use our LLM and Inzai services that do not involve the Internet. On the other hand, we hope that the general customers will also use best LLM depending on their needs. NEC has decided to give the AI its own name, cotomi. Our researchers and developers have worked very hard to develop this, and so we decided to give it a name, because the development is now completed. Actually, in naming it cotomi, more than 50 candidates were proposed by nearly 100 researchers, developers, marketers, salespeople, and so forth. And we also asked NEC's generative AI as well. And cotomi was the name we have chosen eventually, and this means things will become to be coming to fruition. We think it's a very good name. The name cotomi is used consistently going forward, not only for LLM, but also for multimodal, different offerings, including videos and images. cotomi can transform our business in any industry. In software development and the product development, there are some commonalities regardless of industry. Each industry sector has its own priorities, and we would like to create a model according to them, one at a time. Also, we would like to include cotomi in all NEC products going forward. Now, let me talk about the scenario of how we will expand its penetration. Global players tend to focus on foundation model, chips, and hardware. What is more important to me is the level above that. Of course, the demand for chips and so forth is important, but how to use them is more important. That's why we are considering three phases, and we are collaborating with 12 private companies and three universities for phase II and III. We are working very hard on the framework and the model. So once the model is complete, we will address each of the industries, and we would like to spread the model to as many companies as possible in each industry. For phase III, we are thinking of developing business using cotomi, in cooperation with various companies, partners, including partners and software companies. We would like to go to phase II and phase III in next year onward. Now, I have been talking about models, but I would like to talk about what model is. Let me set the tone here. A model is a method or a mechanism that uses input data to derive a result. A model consists of three elements in our mind. There are the foundation model, the industry-specific model, and the business-specific model. The bottom layer is the foundation model, which is about the choice of LLM. There is, of course, cotomi, or, there are other options like OpenAI and Google's AI. The middle layer is the industry model, and so there are various priorities in each industry. For example, priorities for manufacturing and finance may be different. The top layer is a business-specific model. Each model has different functions. For example, search function, inquiry function, and use of different networks, and so forth. The functionality determines the application and the functions to be used.... First, as I can see on the left, there are a lot of packages that we already have, manufacturing packages and medical packages that NEC has been traditionally using. We want to build cotomi into those or newer versions of those packages. We can call this, we call this a cotomi insight. Let me explain three specific examples. The first one is healthcare. This is a project that is being underway at the Tohoku University Hospital. In this example on the left, we are trying to reduce the doctor's workload by using voice recognition for the medical records. Because it's voice recognition, it's possible to create a medical record by converting it into text and then reading it into the system. We were able to reduce the actual working hours for 10 doctors by 116 hours. This is the actual reduction because we measured it together with the customers. On the right-hand side, we have an example that is typical of generative AI. Doctors spend a lot of time outside of consultation time to create various medical documentation and the paperwork. In the future, the legal system will become more stringent on doctors and other medical professionals working overtime, so reducing their workload is critical. There is a social demand to do this, so that the doctors can spend less time on creating documents and medical documents and reports, including insurance claims. We have actually done it, and the results show that the working time was cut by half. This type of challenge is very suitable for generative AI. At the same time, we have already announced this, Hashimoto Municipal Hospital is using our package called MegaOak, and they installed the AI into their existing package and is using it at various locations of the hospital. The second example is non-life insurance, and the third example is a local government. As described, in a recently published article, Mitsui Sumitomo Insurance has various agents and experts answering various questions and inquiries day by day, and their responses to those inquiries are now learned by generative AI, so that the AI can answer questions on behalf of the humans. Similarly, Sagamihara City also use the inquiry response function. The target audience is citizens in this case, but the AI learns the knowledge of veterans and experienced people and answers various inquiries from citizens on their behalf. This saves time and effort. It also buys time for training new employees. These recently announced examples are the most common use cases we expect to see as we enter into the diffusion phase going forward. So those were the examples that just announced from NEC. So now, how can we cover a large number of customers? We talked earlier about the Inzai Data Center services, and Mr. Nishihara also talked about NDP or NEC Digital Platform. We will be incorporating various functions into NDP, and the NDP will consist of five layers, from the application layer to the network layer, and we sell all the hardware and considerable network knowledge together. We are going to enhance cotomi's various plug-ins and APIs as managed APIs, and this will allow, for example, IaaS and PaaS, global cloud companies, to utilize our cotomi plug-ins and APIs. When we are promoting phase II and phase III, this is what we want to do in the partnership. Another key to expanding the market in the future is safety and security or risk management. There is a huge difference between general use of AI and use by industrial sectors and the personnel. In this sense, we are considering whether we can provide services by monitoring our clients' models or data and assessing risk together with Robust Intelligence, which is currently attracting attention globally for its global risk assessment capability. We are planning to release these as models which are already checked for risks, for financial, municipal, and medical applications, for example. We believe this is very important. We are also evaluating our own LLMs. We will certainly provide that as our insights from the NDP. As announced on July 6th, a team called NEC Generative AI Hub was set up under myself, as CDO. The team consists of 110 members, including data scientists, data engineers, and prompt engineers, all of whom have important expertise in the company's business of generative AI. In addition, under myself as CDO, there are 40 global external relations, legal, and governance professionals who specialize in digital trust. This is an important team as well. The team of the promotion engine and its governance engine must come hand in hand to launch an industry-specific offering. We need both. On top of that, we have an enterprise organization that is responsible for a large number of customers. We have created a team of 80 ambassadors, including engineers and salespeople. Since this is a customer-facing team, they will receive various skills to be transferred from the Generative AI Hub team, and those are the new, the established teams that we have recently established. We launched NEC Generative AI Hub concerning NDP, that I mentioned before, and another highlight is Generative AI Center, led by Mr. Nishihara team. Mr. Nishihara, please. We set up a new generative AI center this time, which includes a team of approximately 100 members of Japanese and international researchers. This and NEC Generative AI Hub of Mr. Yoshizaki is linked to cover all the way from new technical research, prototype, alpha version product, beta version product, to delivery, to managed services for customers seamlessly. It will start to operate from the second half of this fiscal year. Thank you. I talked about the hardware chip before, and going forward, upper layer algorithm and the software will be transformed rapidly. We announced a new controller this time. How we can reflect this speed into commercialization is what matters. I think that the only player which runs this cycle quickly will survive. So to that, I am serving as a speaker, but, I'd like to promote DX through collaboration between R&D and the business, not only my division, but various divisions in NEC from perspectives of product, services, and business. This concludes my presentation. Mr. Yoshizaki, thank you. Next, I'd like to present other technologies of growth business. Another AI technology, which is different from generative AI, is biometrics, which is our point of priority. Biometrics, including face recognition, showed remarkable progress in business application. We had number one market share in diverse domestic market, as shown here, and as shown in the right bottom blue part. In 80 global airports, NEC technology is adopted in immigration and boarding procedures, and NEC technology is used for a national ID, 1.6 billion population in India, Vietnam, and other countries, as well as the use for public safety. According to the survey by Frost & Sullivan, many consulting companies rated us as global number one in this field, as well as our technology business. On NID, this time, we present the new technology, gateless access control system using biometric recognition. Last year, we presented the gateless biometrics, which enabled biometric recognition without gate, and the system presented this time is equipped with more sophisticated AI edges. Wherever you install the camera and edge, any place can be the point of gateless access control with biometric recognition. These biometrics can be deployed in much more extensive areas. I think you might have experienced it at the demonstration when you entered that venue, and I hope you to experience again when you leave here. As shown on the right, in addition to the face recognition, various vital informations, including pulse rate, SpO2, respiratory rate, is available immediately from facial video. I hope you to experience this also. This technology won an innovation award in CES. Next, I'll explain the contribution to the social infrastructure business on key points, though there is no exhibition regarding this. Last year, we presented 28 gigahertz technology of distributed MIMO in the 6G beyond 5G part. It was used for base station, and this year, we had a successful experiment with 40 gigahertz. This will enable for further high speed, high capacity communication, working with NTT and NTT DOCOMO. Beamforming technology at base station enables to maintain the track, moving, terminal devices through beams to secure communications, which is comparable to the invisible wired optical communications, and it is in high capacity. High capacity is maintained while moving and in blocked environment or in reflective environment by switching antennas at base station to maintain tracking. In 6G beyond 5G, another highlight is Space, Air, Ground Integrated Network. It requires diverse communication technologies of optical, wireless, and satellite. We established a co-creation institute with Tohoku University, with which we conduct various joint researches on multiple layers on daily basis. Researchers of NEC and Tohoku University work together to create new technology in space, air, ground integrated network. Another highlight is multi-core optical fiber. We have been working with Google over 10 years on optical fiber cable. We succeeded in doubling the number of cores in optical fibers, reducing cost per bit by half or more, enabling higher capacity transmission at a lower cost per bit. This is used in Taiwan, Philippines, and U.S. submarine cable system. Another one is a quantum cryptography, shown on the right, which is regarded as the ultimate security system. The devices will be delivered in this and next fiscal year. We have completed commercialization, as shown in the exhibition. It won the Design Concept Division award. We are now working to adapt quantum cryptography technology in general optical fiber communication and downsize the module. If they are realized, quantum cryptography can be used in desktop-to-desktop optical fiber communication. Finally, let me present the creation of new growth businesses. First, IP licensing business. IP is essential for a technology company, but previously, as shown below, we were striving to strengthen IPs through IPCs in R&D, business development, and business activities, and they led to business defense, co-creation based on the NEC's technology and the developed runs. They are very important for technology, but they did not generate cash by themselves. However, it is very clear that our ICT technology can be adapted in diverse industries. For example, they are adapted in streaming, automotives, devices, et cetera, and they steadily began to earn revenue through licensing fees. We provide technology to society and earn IP revenue. First, we'll continue to strengthen patent, enhancing NEC's core technologies through patent towards 2025. As shown here, compared to 2018, the percentage of patents in those areas will double. As shown on the right bottom, through patent pools, we will provide licensing to other industries. Through those organizations, our technologies will be deployed in various market, including AI, biometric authentication, and autonomous vehicle markets. We will promote them in the framework of commercialization. To be more specific, when you compare the IP revenue in 2021-2025, with that in the previous five years of 2016-2020, it is almost expected to double. IP revenue tends to be one-off, but we'll stabilize the revenue recognition in accounting. Then, we can ensure the steady growth, which leads to stable revenue. As for the size of licensing revenue, it will be about tens of billions of JPY per year. As it will be directly reflected in profit, it gives material impact. Regarding the progress in other growth businesses, NEC has been struggling to create new businesses and established new business innovation process. Business ideas come from NEC group, partner companies, Silicon Valley entrepreneurs, and open contest as well. Then, through the scope of generate, ideate, develop, and launch, we steadily develop business while properly controlling risks. They may lead to establishment of new business within NEC, startups through carving and spinout and JV. This structure, which was established in five years, is described in the book, shown on the right top, entitled as The Innovation by Large Enterprise, written by Mr. Kitase, Corporate SVP, who leads this project. This is available in the venue, so please take a look. Let me share with you the specific results. Last year, we announced a new business value target of JPY 300 billion for 2025. Definition of business value is a value of the business when sold to the third party. It is comparable to the company's offer price in M&A. It was JPY 78 billion last year, and it is estimated to be JPY 140 billion this year. This is not based on our internal assessment, but on the assessment by the external consulting firm, which examined the audit of our new businesses. Within the mix, healthcare and life science accounts for a big part, with about JPY 200 billion in 2025 as a target. It includes AI drug discovery and OncoImmunity, which is shown in the demonstration. Besides, DX, agritech, and quantum computer businesses are making progress in commercialization, and they began to contribute to sales and profit, as well as quantum cryptography. For healthcare and life science business, let me explain the highlights only. There are three sub-segments, and one is medical care. This is a large business of NEC, and we offer electronic medical record to hospitals nationwide. Through MegaOak, we promote hospital DX. Mr. Yoshizaki presented the technology to improve hospital management through LLM before, and it will be a part of MegaOak next year. Lifestyle support segment of health promotion and testing services is shown in the middle. They exhibited in the full visuals. AI drug discovery is shown on the right. We personalize and universalize vaccines with AI. We personalize cancer vaccine and betacoronavirus vaccine, among others, with about 10 vaccines in the pipeline. Data-driven DX business and quantum computing technology are shown in the exhibit, so please visit them. Finally, this is our external open innovation activity. NEC X is a company that NEC established four years ago in Silicon Valley in the U.S. This is a very unique business based on NEC's IP, with external entrepreneurs for commercialization and with external VCs for investment, and this business shows prominent growth. In addition to ignite type support to make a startup from scratch, we began to approach to growing startups by offering NEC's technology, and they received many inquiries. In the last 2 years, we had 10 successful startup launches. We'd like to further promote this initiative. The second one is Bird Initiative, a domestic open innovation initiative. NEC X is a wholly owned subsidiary of NEC, but this business includes fund of other companies as well. How we can combine technology, assets, and HR of large corporations with external companies is one of the major challenges in Japan, and one of the objectives of this initiative is to take a lead in addressing the challenge. We succeeded in carve-out of two companies, automatic negotiation, AI with drone, and AI simulator, both composed of NEC technology and AIST technology. Our hypothesis was proven to be right, because we received many inquiries on the idea to create a new business by combining the technology of large corporations with startups, and we have 63 cases of new business consulting. This is another pillar of Bird Initiative. We continue to create new businesses on our own, as well as brushing up the consulting business. My presentation included R&D, new businesses, and announcement of generative AI. Thank you very much for your attention!