Top 10 tech developments of 2023 according to Alibaba DAMO Academy

DAMO Academy, Alibaba Group’s internal research institute, presents its forecast for the coming year for the fifth year in a row. According to him, ” From generative artificial intelligence to cloud-native security, 2023 will be the year that advances technology, reshapes industries and fuels innovation.” In his detailed report, he reveals the top 10 technology trends for 2023.

The Alibaba DAMO Academy compiled this list of top technology trends by analyzing research papers and interviewing nearly 100 subject matter experts. Scholars have considered many factors, including the impact of theories, technological feasibility, degree of industrialization, and social values.

Alibaba DAMO Academy Head Jeff Zhang says:

“In 2023, the development of various technologies will facilitate the deployment of artificial intelligence and other digital technologies in vertical markets and strengthen the cooperation of public and private sectors and individuals in security technologies and security management. Innovation driven by advanced technologies and their specific application for industry has become an irreversible trend..

Top 10 Upcoming Tech Trends

• Generative AI It’s been in the news a lot over the past year, allowing anyone to produce images, text, software code, drawings and illustrations, video and audio. It has already experienced significant growth and should continue to grow. With advances in technology and future cost reductions, generative AI will become an inclusive technology and can dramatically increase the diversity, creativity and efficiency of content creation. According to the DAMO Academy, ” “Over the next three years, generative artificial intelligence will develop human-like content creation capabilities to facilitate digital content creation.”.

• Dual Engine Business Intelligence DAMO is an important emerging technology for the Academy. Powered by both operations optimization and machine learning, the dual-engine decision intelligence system enables dynamic, comprehensive, and real-time allocation of resources such as real-time power dispatch, port throughput optimization, airport stand allocation, and production process improvement. According to the institute, it will be applied in more scenarios and predicts that ” Business intelligence powered by machine learning will play a larger role in carbon reduction programs.”.

• Cloud-native security will be more versatile and more easily adaptable to multi-cloud architectures and more conducive to building dynamic, precise and enforceable security systems in hybrid environments. According to Gartner’s 2021 report, more than 95% of new digital workloads in 2021 will be deployed on on-premise cloud platforms by 2025, compared to 30% in . For DAMO Academy “In the future, companies will adopt a cloud-native security approach that integrates security into all phases of the software development cycle. Instead of having a security system as an add-on, security will be built into the product design from day one.”.

• Pre-designed multimodal base models According to Huang Fei, head of the Language Technology Laboratory of DAMO Academy, it will serve as the basic infrastructure for artificial intelligence systems for tasks related to image, text and audio, thinking about them, answering questions , will provide cognitive intelligence opportunities to generalize and create. – modal models for understanding, searching, creating and answering questions. For DAMO, “By giving enterprises access to advanced models and data analytics, multi-module pre-training models will be key to improving business productivity in the future.”

• CIPU (Integrated Hardware-Software Architecture for Cloud Computing) will become the norm cloud computing creation and will provide new development opportunities for core software R&D and custom chip design. According to DAMO, “Exponential growth in data processing needs has made centralized processing units (CPUs) inadequate.”He predicts: ” Over the next three years, cloud computing will adopt a new CIPU-centric infrastructure system. This will create growth opportunities for core software research and development and the specialized chip industry.”.

• predictable fabric It is expected to disrupt the TCP-based network architecture and become part of the core network of next-generation data centers. According to DAMO, cloud providers do not offer guaranteed network resources to customers for this: ” As the computing power grid becomes a key part of the nation’s infrastructure, predictable demand will undoubtedly be the number one priority.”Submits a report that“In the next two to three years, a predictable network powered by edge network integration will become the norm.”

• digital imaging will continue to revolutionize traditional imaging technologies and lead to innovative applications such as lensless imaging and non-line-of-sight (NLOS) imaging, the report states: ” The emergence of computational imagery is expected to change the way humans and machines perceive the world.”

• Chips It could bring a new wave of change to the IC R&D process and reshape the landscape of the chip industry. To make computer chips faster, designers must pack more transistors onto a single piece of silicon. Instead of producing larger chips, they now favor an alternative approach of breaking the chips into smaller pieces, chiplets, and then assembling them. These offer various advantages: they are less prone to manufacturing defects and they are cheaper. Custom chip manufacturers can also add or remove chiplets to processors to tailor the chips to their needs. A new standard for the design of chiplet-based processors was also introduced in 2022. For the DAMO Academy, “These small but powerful components will continue to transform chip research and development in the coming year.”

• Processing-in-memory (PIM) chips should be used in more powerful applications such as cloud-based output. This will have a positive impact on sectors such as cloud computing, AI and the Internet of Things (IoT). In-memory processing facilitates the transfer of data: it allows calculations and processing to be carried out directly where the data is stored (computer, server). ) as a result, energy consumption decreases and system performance increases. The report states that ” in-memory processing will become the dominant computing architecture in the era of artificial intelligence. Developers have designed in-memory computing chips to power a wide range of artificial intelligence applications, from virtual and augmented reality to computing astronomical data..

• Large scale Urban Digital Twins It will be more autonomous and multidimensional in a new approach to city management: traffic management, natural disaster prevention and management, carbon peaking and neutrality. According to the DAMO Academy, ” large-scale urban digital twins have made significant advances in traffic management, natural disaster prevention, and carbon emissions management. In the future, they can help urban planners develop and test autonomous vehicles, drones and robots to build smart cities.”

See the full report