On June 11th, the 2025 "ERNIE China Tour - Nanjing Station", guided by the National Engineering Research Center for Deep Learning Technology and Applications and hosted by Baidu Online Network Technology (Beijing) Co., Ltd., was successfully held. Industry elites, experts, scholars, and AI developers gathered to delve into topics such as algorithm application innovation in the era of large models, data value mining, and interdisciplinary talent cultivation. They explored how artificial intelligence technology can drive the in - depth integration of the real economy and the digital economy through the government - industry - university - research collaborative innovation mechanism.

Luculent was invited to participate in the conference and became one of the "First Batch of Baidu AI Technology Ecosystem Partner Enterprises in Nanjing". It also delivered a keynote speech and showcased a series of AI product achievements.

Li Fengcheng, an expert from the company, delivered a keynote report titled "Application and Practice of Luculent SuShine Yaoguang Large Model in the Energy Field" at the main forum of the conference. Focusing on the core pain points of industrial enterprises in crossing the gap from "digitization" to "intelligent digitization", through the collaborative application of the "combination punch" of the large - model's general knowledge and the small - model's specialization in specific scenarios, it drives the "intelligent" transformation and upgrading of energy enterprises, making the value of AI tangible and measurable.

"Like an industrial AI expert who understands both theory and practice, the Yaoguang Large Model, relying on the training of massive multi - modal data, has decision - making capabilities such as industrial knowledge reasoning, production risk identification, and process simulation optimization. By intelligently disassembling tasks and accurately scheduling corresponding tools and small models, it can quickly respond to personalized needs such as unit safety status monitoring and intelligent equipment fault early - warning, truly achieving 'big empowerment from a small breakthrough' and bridging the 'last mile' of enterprise intelligent digitization implementation," Li Fengcheng introduced.

The large model not only constructs an intelligent operation and maintenance closed - loop system from perception, analysis to decision - making with three intelligent agents of "intelligent inquiry - intelligent diagnosis - intelligent strategy" as the core. It also break - throughly extends the value of AI to more high - value business areas: at the production operation end, it drives scenarios such as intelligent monitoring, fan energy - saving, and cold - end optimization for "energy efficiency improvement"; at the safety management and control end, it builds a "zero - hidden - danger" defense line for intelligent inspection and work permit review; at the business management end, it provides "intelligent digital assistants" services such as contract intelligent review and decision - making assistance, injecting intelligent momentum into enterprise management.

At the new - type power sub - forum of the conference, Li Hui, another expert from the company, delivered a report titled "PaddlePaddle Empowers the Upgrade of the Power Generation Intelligent Security System", sharing the "New Intelligent Security System" jointly created by Luculent's AI vision algorithm matrix and Baidu PaddlePaddle platform.

In response to the current safety management dilemma in factories, this system deeply integrates Luculent's independently developed key technologies such as the image comparison algorithm and monocular 3D recognition algorithm with Baidu PaddlePaddle's core algorithm capabilities in image classification, object detection, semantic segmentation, etc.

Like a vigilant "safety guard", it "captures" various potential abnormal risks such as coal smoldering, leakage, and hot - work operations 24/7, and quickly responds through the LiEMS control platform, achieving accurate information push, rapid problem handling, and comprehensive post - event evaluation, realizing the closed - loop implementation from "algorithm early - warning" to "business linkage". It makes traditional safety management more than just "monitoring" but also "early - warning", and makes machine vision more than just "seeing" but also "understanding", truly becoming the "guardian" of factory safety and the "booster" of production efficiency.

"In the future, we will further promote the integration of AI vision and large - model capabilities and continuously inject the industrial knowledge system to create an 'industrial digital butler' with more 'foresight thinking' and 'decision - making ability', accurately analyzing where, when, and why abnormalities and risk events occur in the factory, making safety emergency response more scientific and precise, and preventing factory safety incidents before they happen!"

Li Hui said in an interview, "Luculent will carry out all - round and in - depth cooperation with Baidu around scenarios, technologies, and applications, and continue to make efforts and innovate in areas such as intelligent material annotation and secondary verification of visual reasoning results. AI is Luculent's unwavering long - term strategic track. The company will strive to build a leading AI system in the energy industry through the new cooperation paradigm of 'technology platform + scenario application + industrial ecosystem' to help enterprises' intelligent transformation reach a new level!"