Diverting the river into the sea, TCL Electronics embraces big changes in the field of AI

@新tropy original

Edit丨伊page

The Yellow River originates from the Qinghai-Tibet Plateau at an altitude of 4,800 meters, flows through nine provinces and regions, changes its course six times, and finally converges at the Bohai Sea.

Change goes far and all things are prosperous. This river is so similar to TCL, which has more than 40 years of technological manufacturing development history.

The water gathered from the source of the plateau winds through the valley in the middle reaches, gathers at the dragon gate in the lower reaches, and pours down rapidly, running thousands of miles along the way. Relying on the self-reliance of reform and innovation and the courage to move forward, TCL started from the initial telephone business and gradually moved to a broader range of smart terminals, semiconductor displays, new energy photovoltaics and other fields, constantly challenging in various scenarios. Its own boundary leads the river to the sea meanderingly and firmly.

In this era where everything can be "AI", TCL Electronics (01070.HK), which focuses on smart terminals, is further exploring new areas of artificial intelligence and making waves in the ocean of AI variables.

Fit the scene and find the boundaries of large model capabilities

AI large-scale model is a super large-scale artificial intelligence model built based on large amounts of data using deep learning technology. These models can handle massive amounts of data and have very strong intelligence and learning capabilities. ChatGPT's strong "out of the circle" at the end of 2022 has caused people to focus on large AI models and brought new opportunities and challenges to the development of artificial intelligence.

Domestic Alibaba Damo Academy, Baidu, Byte, Huawei, Pengcheng Laboratory and other companies are all developing large models on their own, and these things all developed within a few months after the rise of ChatGPT. The real difficulty of AI large models is how enterprises choose the most suitable large model based on their actual business volume and application scenarios. This undoubtedly requires enterprises to create the boundaries of large model capabilities and the fit points of enterprise products and services.

The layout is orderly and the moves are sound. TCL Electronics has found the direction to determine what the large AI model needs to do, that is, by exploring the boundaries of the large model's capabilities, and laying out the two major tributary scene directions of outward and inward：

Externally, facing consumers, we use the capabilities of large models to upgrade the interactive experience of smart terminals, and use large models to enrich the content generation of smart terminals to further enhance the user's product experience.

In terms of intelligence, various products such as TVs, tablets, mobile phones, and smart glasses form a rich and diverse combination of application scenarios, giving users a better experience in different scenarios and creating products that can understand and It can identify complex and fuzzy instructions issued by users and proactively understand user needs; at the same time, the rich content formats meet consumers' needs for quick access to new information and are also an important means of increasing user stickiness, which can feed back TCL Electronics' intelligent business.

Inward, facing enterprise operations, strengthen the penetration of human-machine collaborative Co-working business scenarios in intelligent manufacturing, supply chain, marketing, finance, human resources, etc. AI large models can effectively improve operational management efficiency and customer satisfaction degree and reduce operating costs. According to Xinhua News Agency, research results from Stanford University and MIT show that with the optimization of large model parameters and the gradual improvement of computing power utilization, the use of mature large models will increase the labor productivity of technical support employees by an average of 14%, processing complex Scenario business capabilities will also be greatly enhanced.

Bengyong: Multi-modal AIGC application, rich large model ecology

"Data-Model-Application" is the key to the success of large model companies, which solves the problem of what AI large models need to do. With the gradual formation of AI large model business models, the AIGC era is just around the corner.

Let’s take the iconic ChatGPT as an example. ChatGPT is just an interface for To C. The large model behind it is GPT, which is Generative Pre-trained Transformer. It is an extremely large parameter that uses large-scale corpus for unsupervised pre-training. Deep learning models at scale. In the field of artificial intelligence, the GPT series is one of the best-performing large models currently and one of the representative products of AIGC (Artificial Intelligence Generated Content).

In the past, human-computer interaction mainly used text for basic communication, using simple language and only using a single modal model. With the advancement of technology and increasing computing power, the AIGC model is gradually improving its ability to understand images, enabling it to understand and process images. Benefiting from the continuously evolving text, picture, and video content generation capabilities of AI, AIGC will be the first to be implemented in education, games, finance, law, media, consumer electronics and other fields, promoting cost reduction and efficiency improvement in various industries, enhancing the richness of content, and creating Brand new content form. Multimodal interaction has become an important direction for future AIGC applications.

So, how to apply?

The answer is simple and straightforward - wherever there is a screen, there must be content. From personal computers in the PC era, to smartphones in the mobile era, and even future AR/VR devices, the core carrier is the screen. Therefore, for a company like TCL Electronics that has screens and interactive display terminals, the intelligent generation of content is a huge bonus. AIGC supports the generation of cross-modal content such as text, images, videos, etc., allowing dominant companies in the display field like TCL Electronics to quickly provide consumers with lower content production thresholds and higher production efficiency. Bringing a new content ecosystem and a more natural and personalized interactive experience.

It is understood that TCL Electronics continues to invest in research in the field of AI in order to explore AIGC models that significantly improve consumer perception and experience, enhance the understanding of human-computer interaction intentions, better match and connect richer content, and The results are displayed in front of consumers.

A few days ago, TCL Electronics announced that its domestic Internet business entity Thunderbird Technology has launched the first children's exclusive column based on a large AI model - "Super Smart Painting" AI story collection. This is the first time in the Internet TV industry that a customized content column generated by artificial intelligence large model technology has appeared. On the one hand,

can quickly generate a large amount of diversified creative content, improve interactive experience and user stickiness, thereby promoting transaction conversion

; On the one hand, as the industry's first column based on a large AI model, the launch of "Super Smart Painting" marks TCL's success in leading a new model of Internet TV ecology.

In addition to TVs, AI large models can also enable powerful product matrices such as TCL mobile phones, tablets, commercial display conference screens, AR/XR smart glasses, etc. to achieve greater applications by comprehensively utilizing multi-modal modes such as voice, vision, and video. . When the above product matrix occupies the modal input of the user's visual and auditory senses, AR glasses and microdisplays can comprehensively utilize multi-modal modes such as voice, vision, and video to provide users with a closed loop of expected results and truly release the large model ecosystem. .

In fact, TCL Electronics has already started ecological experiments. For example, the RayNeo X2 augmented reality (AR) smart glasses released by Thunderbird Innovation incubated by TCL Electronics at CES 2023 this year use a revolutionary binocular full-color Micro-LED optical waveguide display and a series of new interactive functions to create Unparalleled AR experience. The glasses are the world's first consumer-grade AR glasses that use binocular full-color Micro-LED optical waveguide displays, which can provide high contrast and brightness for indoor and outdoor use, attracting more user groups. At the same time, the glasses also provide a The series integrates assistant functions, including intelligent navigation, automatic translation, photography and music playback, and provides a hands-free integrated camera, allowing users to create still images, videos and time-lapse photography from a unique first-person perspective, enriching the user experience. As proof of its commitment to AR innovation, RayNeo X2 is expected to launch a developer project in the first half of 2023, calling on innovative developers to join in and build a creative AR glasses ecosystem.

AIoT2.0 era, TCL Electronics’ “ambition”

The river that continues to surge will surely produce countless waves, just like the opportunities that companies face when facing the storm. If the large model is compared to a "brain", it is not enough to just have a "brain" to sit and talk. Knowledge and action must be integrated, and there are "hands and feet" to functionally realize the intention.

As artificial intelligence enters the era of large models, it will not only become a more intelligent conversation tool, but will also be implemented in many fields such as AIoT and enterprise management. In terms of current technological development, any digital content needs Through a product into the hands of users, that is to say, the large model must be combined with AIoT products. This is a logical necessity.

AIoT stands for AI IoT, which refers to the integration of artificial intelligence technology and the Internet of Things in practical applications. The field of integration of AI and IoT has been booming in recent years. In the wave of AIoT, TCL Electronics is undoubtedly the first to ride the wave. As early as 2019,

TCL Electronics has been developing the "AI x IoT" strategy. Different from AI IoT's simple empowerment of individual devices, "AI x IoT" is a new concept proposed by TCL Electronics, which refers to user needs, IoT device capabilities and AI capabilities interact and are deeply integrated,

and are considered the "AIoT 2.0" era. This is also the reason why TCL Electronics can gain momentum quickly in the era of large models. To simply understand, the AI ​​x IoT strategy is the brain of the AI ​​large model plus the "hands and feet" connected through IoT, integrating home services, software, Hardware, etc., can perfectly connect with the user's intention, give full play to the ability of Internet perception, form a closed loop, and unleash the greater value of integration.

The key to the connection is the Honghu Laboratory of TCL Electronics’ parent company, TCL Industrial.

Honghu Lab focuses on technology research and development and product innovation applications such as AI, IoT and cloud services, providing global users with "AI x IoT" full-scenario intelligent solutions. It is the "data service" and "smart scene" of TCL smart terminals core provider. Since its establishment, TCL Honghu Laboratory has been committed to solving three major industry problems: smartphones and smart TVs have seized the entrance, resulting in high "learning costs" and hindering the popularization of AIoT; poor interoperability, low intelligence, and low intelligence of equipment The number of scenarios is small; the interconnection is difficult and there are few scenarios with poor compatibility.

Based on this, TCL built the Honghu Cloud platform for global business, developed its own Wi-Fi module, realized experience optimization such as distributed distribution network, and more than 200 intelligent functions, and successfully implemented 7 types of intelligent units. Taste. At the same time, TCL Electronics has built multiple smart scenarios for full-category interconnection such as cross-screen circulation, and finally realized the full implementation of AIoT capabilities such as interconnection and voice interaction at the product level. In addition, TCL Electronics has also realized the unification of the account system and user service system. Through the TCL Electronics official APP, users can experience the control and scene realization of IoT anytime and anywhere, allowing interconnection and interoperability to go hand in hand, truly achieving "AI x IoT" .

"Ask the channel how clear it is, and let the source of living water flow." TCL Electronics' parent company TCL Industries has 25 R&D centers and 6 ecological laboratories around the world, focusing on Mini LED display technology, artificial intelligence, IoT , AR/VR, cloud computing and other technologies have built strong ecological and technological competitiveness, and are like a golden key, opening the door for TCL Electronics to enter the era of truly active and intelligent AIoT 2.0, realizing the "ambition of the swan".

Convergence: Responsible for users, interconnection under the framework of Matter protocol

Even if the wave of large models drives TCL Electronics to have a broader space for imagination, and to explore and deepen the road to create a quality life for consumers by continuing to deepen the "AI x IoT" strategy, TCL Electronics still has one last thing to solve. Question - user choice.

Consumers have the right to choose, but because there is no extensive interconnection between products of different companies, it is actually difficult for consumers to achieve true interconnection and interoperability between cross-brand devices. TCL Electronics adopts a responsible "convergence" strategy to achieve cross-brand collaboration and seek solutions in the AIoT field.

In emerging markets, no company can survive alone, especially when new opportunities are just emerging, competition is bound to be fierce. However, as hundreds of rivers merge into the sea, versatility is TCL Electronics’ commitment to consumers. Therefore, TCL Electronics’ products are competing to obtain the world’s first batch of Matter certifications.

Matter's Internet, as an important cross-brand basic agreement, can effectively break the "isolated island" situation. It will serve as an ecological alliance, like the sea, bringing together the tributaries of many products, allowing consumers to sail the same boat. , you can enjoy the full scene space.

In this technical framework, there is a standard for cross-brand device connection. Consumers can control all devices whether through TCL Home APP, Amazon Alexa, Google assistant, or Apple Home, Siri, etc. Brands support such products to truly achieve a cross-brand interconnected user experience.

At present, TCL brand air conditioners and TVs have obtained the first batch of product certifications. After certification, users can easily interconnect with these devices whether they are using Apple, Google mobile phones, or Amazon Alexa devices.

As an "ocean" that accepts all products and equipment, TCL Electronics not only has breadth but also depth in interconnection.

TCL Electronics and its parent company TCL Industries are not only one of the first companies to promote and participate in the development of the Matter protocol, but are also important members of the 3GPP and AVS standards. They are currently promoting wide area network and local area network communication technologies, as well as audio and video editing. The work of international standardization of decoding technology revolves around all types of interconnected devices, overlaying all IoT scenarios, and user-centered audio and video streaming.

Conclusion: Lead the river into the sea and stay on the tide forever

When the sea is flowing, a hero shows his true colors. Faced with the variables in the AI ​​field, when each company builds its own technological moat, TCL Electronics is thinking about how to guide the river into the sea and sail out to sea. This responsibility is worthy of appreciation.

In the face of the AI ​​era that is surging with hope and power, from efficiency leadership to ecological leadership, from "bravely standing at the forefront" to "eternally standing at the forefront", TCL Electronics has great ambitions under the still water. , also waiting for time’s answer in the story.

The above is the detailed content of Diverting the river into the sea, TCL Electronics embraces big changes in the field of AI. For more information, please follow other related articles on the PHP Chinese website!