Alibaba DAMO Academy (DAMO), a global research initiative by Alibaba Group, has released its forecast of the leading trends it believes will shape the tech industry in 2022.
The DAMO list of 10 technology trends for the next two to five years is the result of analysing millions of public papers and patent filings over the past three years and conducting interviews with nearly 100 scientists.
"Over the past century, the evolution of digital technologies has accelerated technological progress and industrial development," says DAMO head, Jeff Zhang.
"The boundary of technologies is extended from the physical world to mixed reality, while more and more cutting-edge technologies find their way to industrial applications. Digital technology plays an important role in powering a green and sustainable future, whether in green data centers and energy-efficient manufacturing or day-to-day activities like paperless offices. With technology, we will create a better future."
He says in the next two years they expect to see a surge of applications running on top of the new computing system, including:
DAMO expects the rapid development of new network technologies will fuel the evolution of cloud computing towards a new computing system, cloud-network-device convergence. With the new system, clouds, networks, and devices have a more clearly defined division of labour.
"Cloud-network-device convergence is the catalyst that will drive the emergence of new applications to fulfil more demanding tasks, such as high-precision industrial simulation, real-time industrial quality inspection, and mixed reality," says Zhang. "In the next two years, we expect to see a surge of applications running on top of the new computing system.
"In the next three years, we expect to see AI broadly applied in the research process of applied science, the widespread use of silicon photonic chips in large-scale data centers, AI paving the way for the integration of renewable energy sources into the power grid, people-centric precision medicine becoming a major trend, groundbreaking improvements in the performance and interpretability of privacy-preserving computation, as well as a new generation of XR glasses."
AI for science
For hundreds of years, the scientific community has had two basic paradigms, experimental science and theoretical science. DAMO says the advancement of AI is making new scientific paradigms possible. Machine learning can process massive amounts of multidimensional and multimodal data and solve complex scientific problems, allowing scientific exploration to flourish in areas previously thought impossible.
DAMO believes AI will accelerate the speed of scientific research and help discover new scientific laws. Over the next three years, it expects AI to be broadly applied in applied science research and be used as a production tool in some basic sciences.
Silicon photonic chips
As the size of transistors approaches physical limits, the speed of electronic chip development can no longer meet the increasing data throughput demand brought by the rise of high-performance computing.
Unlike electronic chips, silicon photonic chips use photons instead of electrons to transmit data. Photons do not directly interact and can travel longer distances, so silicon photonic chips can provide higher computing density and energy efficiency.
"The rise of cloud computing and AI drives the rapid development of silicon photonics technology," says Zhang.
"In the next three years, we can expect to see the widespread use of silicon photonic chips in high-speed data transmission in large-scale data centers."
AI for renewable energy
The recent and rapid development of technology in renewable energy such as wind and solar power has made renewables a tempting energy source to add to the power grid. However, DAMO says issues such as difficulty in grid integration, low energy utilisation rates, and storage of excess energy are significant roadblocks along the way.
It says due to the unpredictable nature of renewable energy power generation, integrating renewable energy sources into the power grid presents challenges that affect the safety and reliability of the grid. The application of AI in the industry is pivotal in improving the efficiency and automation of electric power systems, maximising resource usage and stability.
DAMO expects the convergence of AI and precision medicine to boost expertise and new auxiliary diagnostic technologies and serve as a high-precision compass for clinical medicine. Using this compass, it says doctors will diagnose diseases and make medical decisions as quickly and accurately as possible.
"These advances will allow us to quantify, compute, predict, and prevent severe diseases," says Zhang. "In the next three years, we expect to see people-centric precision medicine become a major trend that will span multiple fields of healthcare, including disease prevention, diagnosis, and treatment."
For a long time, the application of privacy-preserving computation has been limited to a narrow scope of small-scale computation. This is due to performance bottlenecks, lack of confidence in the technology, and standardisation issues.
As more and more integrated technologies, such as dedicated chips, cryptographic algorithms, whitebox implementation, and data trusts emerge, privacy-preserving computation will be adopted in scenarios such as processing massive amounts of data and integrating data from all domains.
DAMO says the adoption will boost new productivity that is powered by data from all domains. In the next three years, it says we will witness groundbreaking improvements in the performance and interpretability of privacy-preserving computation and the emergence of data trust entities that provide data sharing services based on the technology.
Extended reality (XR)
The development of cloud-edge computing, network communications, and digital twins brings XR into full bloom. XR glasses promise to make immersive mixed reality Internet possible. DAMO says the technology plants the seed for a new industrial ecosystem encompassing electronic components, devices, operating systems, and applications. It says XR will reshape digital applications and revolutionise the way people interact with technology in scenarios such as entertainment, social networking, office, shopping, education, and healthcare.
"In the next three years, we expect to see a new generation of XR glasses that have an indistinguishable look and feel from ordinary glasses entering the market and serving as a key entry point to the next generation of Internet," says Zhang.
"In the next five years, we expect to see perceptive soft robotics replacing conventional robots in the manufacturing industry, and satellites and terrestrial systems working as computing nodes providing ubiquitous connectivity."
Perceptive soft robotics
Unlike conventional robots, perceptive soft robots have physically flexible bodies and enhanced perceptibility towards pressure, vision, and sound. These robots take advantage of state-of-the-art technologies such as flexible electronics, pressure adaptive materials, and AI, which allow them to perform highly specialised and complex tasks and adapt to different physical environments.
DAMO expects the emergence of perceptive soft robotics to change the course of the manufacturing industry, from the mass-production of standardised products towards specialised, small-batch products.
Satellite-terrestrial integrated computing
Terrestrial networks and computing systems provide digital services for densely populated areas, while no service is available in sparsely inhabited areas such as deserts, seas, and space. STC connects high-Earth orbit (HEO) and low-Earth orbit (LEO) satellites and terrestrial mobile communications networks, achieving seamless and multidimensional coverage.
STC also creates a computing system that integrates satellites, satellite networks, terrestrial communications systems, and cloud computing technologies. This makes digital services more accessible and inclusive across the globe.
DAMO says in the next five years, satellites and terrestrial systems will work as computing nodes to constitute an integrated network system providing ubiquitous connectivity. It expects to see the future of AI shifting to the co-evolution of large and small scale models via clouds, edges, and devices.
Co-evolution of large and small scale AI models
Large-scale pre-training models, also known as the foundation models, are the grounding breakthrough technique from weak AI to general AI, which relatively boosts the performance of various applications using conventional deep learning.
However, says DAMO, the merit in the higher performance and the drawback in the power consumption are not well-balanced, limiting the exploration of large-scale models. The future of AI is shifting from the race on the scalability of foundation models to the co-evolution of large and small scale models via clouds, edges, and devices, which is more useful in practice.