[ad_1]
When Hongzhi Gao was young, he lived with his family in Gansu, a province located in central northern China, next to the Tengger Desert. When he remembers his childhood, he remembers that there was a constant, steady earth wind outside their house and lasted no more than a minute after stepping outside most months of the year, and sand filling up any empty space and seeping into it. pockets, boots and mouth. The monotony of the desert stuck with him for years, and at university he turned that memory into the idea of building a machine that could bring plant life into the desert landscape.
Efforts to stop desertification, the process by which fertile land turns into desert, have focused primarily on expensive manual solutions. Hongzhi designed a robot with deep learning technology to automate the tree planting process: from identifying optimal spots to planting tree seedlings and watering. Despite having no experience with artificial intelligence, as an undergraduate, Hongzhi used Baidu’s deep learning platform PaddlePaddle to combine different modules to create a robot with better object detection capabilities than similar machines already available on the market. It took less than a year for Hongzhi et al to spin the final product and put it to work.
Hongzhi’s desert robot serves as a striking example of the increasing accessibility of artificial intelligence.
Today, more than four million developers use Baidu’s open-source AI technology to create solutions that can improve the lives of people in their communities, and many have little or no technical expertise in the field. “In the next ten years, AI will be the source of changes occurring in every fabric of our society, changing the way industries and businesses operate. “The technology will expand the human experience by taking us deeper into the digital world,” Baidu CEO Robin Li said at Baidu Create 2021, an AI developer conference.
As we enter a new chapter in the evolution of AI, Baidu CTO Haifeng Wang identified two key trends that support the progress of the industry: AI will continue to mature and increase its technical complexity. At the same time, the cost of deployment and the barrier to entry will be reduced – benefiting both businesses developing AI-powered solutions and software developers exploring the world of AI.
Combining knowledge and data with deep learning
The integration of knowledge and data with deep learning has significantly increased the efficiency and accuracy of AI models. Baidu’s AI infrastructure has been taking new information and integrating it into a large-scale knowledge graph since 2011. Currently, this infographic has more than 550 billion facts covering all aspects of daily life as well as industry-specific topics including manufacturing, pharmaceuticals, legal, financial services, technology and media and entertainment.
Together, this infographic and massive data points form the building blocks of Baidu’s newly released pre-trained language model PCL-BAIDU Wenxin (ERINIE 3.0 Titan version). The model outperforms other language models without knowledge graphics on 60 natural language processing (NLP) tasks, including reading comprehension, text classification, and semantic similarity.
Cross-modal learning
Cross-modal learning is a new field of AI research that aims to improve the cognitive understanding of machines and better mimic the adaptive behavior of humans. Examples of research efforts in this area include automated text-to-image synthesis, where a model is trained to generate images from text-only descriptions, and algorithms built to understand and verbalize visual content. The challenge with these tasks is that machines make semantic connections between different sets of data (eg, images, text) and understand the interdependence between them.
The next step for AI is to combine AI technologies such as computer vision, speech recognition and natural language processing to create a multimodal system.
On this front, Baidu has unveiled a variant of NLP models that connect language and visual semantic understanding. Examples of real-world applications for such a model include digital avatars that can perceive their environment as humans and handle customer support for businesses, and algorithms that can “draw” artworks and write poems based on their understanding of the artwork created. .
There are even more creative, impactful potential implications for this technology. The PaddlePaddle platform can make semantic connections between vision and language, which prompted a group of graduate students in China to create a dictionary to protect endangered languages in regions like Yunnan and Guangxi by more easily translating them into simplified Chinese.
AI integration across software and hardware and industry-specific use cases
As AI systems are applied to solve increasingly complex and industry-specific problems, more emphasis is placed on optimizing the software (deep learning framework) and hardware (AI chip) as a whole, rather than optimizing each individually. as computing power, power consumption and latency.
Also, tremendous innovation is taking place in the platform layer of Baidu’s AI infrastructure, where third-party developers use deep learning capabilities to create new applications tailored to specific use cases. The PaddlePaddle platform has a set of APIs to support AI applications in newer technologies such as quantum computing, life sciences, computational fluid mechanics and molecular dynamics.
AI also has practical uses. For example, in Shouguang, a small city in Shandong Province, artificial intelligence is being used to streamline the fruit and vegetable industry. It only takes two people and an app to manage dozens of vegetable barns.
“Despite the increasing complexity of AI technology, the open-source deep learning platform combines the processor and applications such as an operating system, lowering entry barriers for companies and individuals who want to incorporate AI into their business,” Wang says.
The barrier to entry for developers and end users has been reduced
On the technology front, pre-training of large models like PCL-BAIDU Wenxin (version ERNIE 3.0 Titan) has solved many common bottlenecks faced by traditional models. For example, these general-purpose models helped lay the foundation for different types of NLP tasks, such as text classification and question answering, to run in one consolidated place, whereas in the past all types of tasks should have been solved. with a separate model
PaddlePaddle also has a number of developer-friendly tools, such as model compression technologies, to tune general-purpose models to suit more specific use cases. With over 400 models ranging from large to small, the platform provides an officially supported industrial-grade model library that preserves only a fraction of the size of general-purpose models, but reduces model development and deployment costs by providing comparable performance.
Today, Baidu’s open-source deep learning technology supports a community of more than four million AI developers who have collectively created 476,000 models, contributing to the AI-driven transformation of 157,000 businesses and institutions. The examples listed above are the result of innovations in all layers of the Baidu AI infrastructure, integrating technologies such as voice recognition, computer vision, AR/VR, knowledge graphics, and pre-training large models that are one step closer to perception. world as human.
In its current state, AI has reached a level of maturity that allows it to do great things. For example, the recent launch of Metaverse XiRang wouldn’t have been possible without PaddlePaddle’s platform for creating digital avatars for attendees around the world to connect from their devices. Also, future breakthroughs in areas such as quantum computing could significantly improve the performance of metadatabases. This shows how Baidu’s different offerings are intertwined and interdependent.
In a few years, AI will be close to the core of our human experience. It will be to our society what steam power, electricity and the internet were to previous generations. As AI gets more complex, developers like Hongzhi will work more in the capacity of artists and designers, given the creative freedom to explore use cases previously thought only theoretically possible. The sky is the limit.
This content was produced by Baidu. It was not written by the editorial staff of MIT Technology Review.
[ad_2]
Source link