Sharing Perspectives on the Future of AI

Shenzhen Loop Area Institute  

Center for Language, Intelligence and Machines (LIMA)

Highlights from the LIMA Open Day

Language, Intelligence and Machines

On December 19, 2025, the Center for Language, Intelligence and Machines (LIMA) at the Shenzhen Loop Area Institute successfully held its 2025 Open Day.

（Group Photo of the Event）

Through a variety of keynote speeches and on-site demonstrations of research achievements and practical application cases, the Open Day comprehensively showcased the LIMA's accomplishments over the past half-year in scientific research, talent development, academic exchange, and student cultivation. The event established a communication and collaboration platform for academia, industry, and students to engage in cross-boundary exchange and co-creation. It supported the frontier exploration and development of research directions such as multimodal large dialogue models, brain-inspired auditory intelligence, and speech brain-computer interfaces. Furthermore, it contributed to advancing high-quality industry-academia-research collaboration and the cultivation of innovative talent in the field of artificial intelligence within the Greater Bay Area.

The event brought together over 300 attendees, including Professor Xuejin Li, Associate Vice President and Dean of Research of the CUHK(SZ), and Bichen Jing, Deputy Director of the China Merchants Lion Rock Artificial Intelligence Laboratory. Participants came from more than 10 government and public institutions, over 30 universities and research institutes, as well as more than 40 enterprises and partner organizations.

At the beginning of the event, Professor Li Haizhou, Director of the LIMA, Fellow of the Singapore Academy of Engineering, and Dean of the School of Artificial Intelligence at the CUHK-Shenzhen, extended a warm welcome to the attendees. He provided an overview of the Shenzhen Loop Area Institute and the LIMA.

Professor Li pointed out that AI aims to endow machines with human-like thinking and action capabilities to accomplish tasks autonomously. He described language models as the cognitive core of AI, while human-computer interaction focuses on optimizing the experience between humans and machines. The LIMA has organized 3 core research teams focusing on large language models, brain-like computing, and brain-computer interfaces. It is committed to advancing the cutting-edge integration of language models and human-computer interaction. He systematically elaborated on the research value and development trends in these related fields.

He stated that this event aimed to facilitate academic exchange among faculty and students, establish a cross-disciplinary communication platform connecting academia, industry, and students, and promote the two-way flow of knowledge and resources. Furthermore, it sought to stimulate greater enthusiasm for exploration in the field of AI across various sectors and create favorable opportunities to attract more outstanding talent to this frontier domain.

Following the opening, Professor Min Zhang, a joint appointment professor at the LIMA, specially appointed Presidential Assistant of the Harbin Institute of Technology (Shenzhen), and Director of the Institute of Computing and Intelligence, delivered a keynote titled Next-Generation Artificial Intelligence with Natural Language. He provided an in-depth analysis of the developmental pathway for next-generation AI centered on natural language within the wave of AI advancement. His presentation systematically outlined the historical context, methodological frameworks, and paradigm shifts in the evolution of artificial intelligence. A key focus was placed on discussing novel architectures for native multimodal large models that use language as their kernel.

Professor Huajin Tang, also a joint appointment professor at the LIMA, a Qiushi Distinguished Professor at Zhejiang University, and a national high-level talent, shared insights on the topic Neural Mechanisms of Brain-Inspired Computing and AI Applications. Professor Tang pointed out that the neural information processing and computational mechanisms of the brain are a core issue of common interest to both neuroscience and computer science. Brain-inspired computing uses neurons and synapses as fundamental computational units. By drawing inspiration from their information processing mechanisms, as well as the neural circuits underlying learning, memory, and higher cognition, it builds sparse event-based representations and learning methods using spikes. This provides a crucial pathway toward highly energy-efficient computing. Simultaneously, by simulating complex neural circuit mechanisms such as perception-learning-decision, it holds key significance for theoretical innovation in next-generation AI and offers important modeling methods for deeply analyzing the emotional mechanisms of the brain.

After the demonstration highlights session, invited guest Professor Liping Wang, Founding Director of the Brain Cognition and Brain Disease Institute at the Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, and Founding Dean of the Shenzhen-Hong Kong Institute of Brain Science Innovation, gave a presentation titled Instinct and Emergence: How to Digitally Present the Functional Duet of the Brain?

Professor Wang noted that survival and emergence are two core aspects for understanding life. The homeostasis maintenance of living organisms encompasses multi-layered manifestations, from instinctive responses to complex social behaviors. His team employs technologies like optogenetics and circuit tracing to analyze the underlying neural circuit mechanisms. Combined with cross-scale detection methods, they aim to construct a foundational structure that reveals the intrinsic architecture of life. When facing internal and external environmental pressures, organisms follow the basic pattern of brain perception - whole-body response, and adaptively exhibit emergent new behavioral traits.

（Professor Haizhou Li, Director of the LIMA, presented a commemorative trophy to Professor Wang Liping.）

To systematically study this process, Professor Liping Wang's team established a computer vision-based digital behavioral analysis system. This system enables the precise characterization of behavioral dynamics and emergence patterns. By further integrating free-moving behavioral studies with multimodal physiological recordings, the research achieved a fully digital analysis of the mechanisms behind "behavior-homeostasis interaction and complex behavioral emergence." This interdisciplinary framework, which merges neuroscience, computational science, artificial intelligence, and engineering, provides a novel pathway for deepening the understanding of brain function and behavioral mechanisms.

Professor Zhiguo Zhang delivered a report titled Brain-Computer Interface: The Path to Human-Machine Shared Intelligence. Professor Zhang holds a joint appointment at the LIMA, serves as Associate Dean of the School of Intelligent Science and Engineering at the Harbin Institute of Technology (Shenzhen), and is a national-level young talent. He pointed out that the relationship between artificial intelligence and human intelligence is evolving from emulation to symbiosis.

Brain-computer interfaces are a key representative of this progression. They utilize advanced machine learning techniques to achieve efficient decoding of neural signals. A direct communication bridge is established between the brain and external devices. This enables the accurate recognition of intent, emotion, and cognitive states. The report systematically outlined the co-evolutionary path of human and machine intelligence. It provided an in-depth analysis of the current state and core challenges of brain-computer interface technology. Finally, it offered a perspective on the future development towards human-machine shared intelligence.

A dedicated exhibition area was set up, showcasing 16 demonstrations of cutting-edge achievements. Key projects included emotionally and cognitively enabled bionic robots, an Omni-Modal Intelligent Interaction Experience Platform, brain-controlled embodied robots, the Speech BCI project, and a smarter local AI multi-model fusion intelligent system. These offered participants hands-on experience and opportunities for in-depth exchange.

（Demo）

The event featured a vote to select the Best Demo Awards. The awards were presented to the Speech BCI project team led by Professor Siqi Cai, the brain-controlled embodied robot team led by Professor Zhiguo Zhang, and the Omni-Modal Intelligent Interaction Experience Platform team led by Professor Baotian Hu. On behalf of the center, Director Professor Haizhou Li presented awards to the three winning teams and encouraged more faculty and student teams to continue co-creating and exploring the limitless possibilities of AI+.

（Award Presented to Professor Siqi Cai's Team）

（Award Presented to Professor Zhiguo Zhang's Team）

（Award Presented Professor Baotian Hu's Team）

The 2025 Open Day of the LIMA at Shenzhen Loop Area Institute concluded amid a technology-rich demo showcase and a relaxed, engaging atmosphere of exchange.

（Group photo of the LIMA）

Looking ahead, the center looks forward to more faculty and student teams collaborating in fields such as multimodal large dialogue models, brain-inspired auditory intelligence, and speech brain-computer interfaces. Through the mutual advancement of academia and industry, it will continue to foster frontier exploration and the translation of research outcomes, jointly creating a new AI+ future.

The Center for Language Intelligence and Machines (LIMA) is one of the five core research centers at the Shenzhen Loop Area Institute. It is dedicated to advancing the frontier integration of language models and human-computer interaction. The center is led by Professor Li Haizhou, Director of LIMA, who also serves as Dean of the School of Artificial Intelligence, Presidential Chair Professor in Chi Kin, and Fellow of the Singapore Academy of Engineering at The Chinese University of Hong Kong, Shenzhen (CUHK-Shenzhen). The center brings together over twenty professors from around the globe to jointly promote cutting-edge exploration and development in research areas including multimodal large dialogue models, brain-inspired auditory intelligence, and speech brain-computer interfaces.

LIMA is entrusted with the mission of the Shenzhen Loop Area Institute as a pilot zone for cultivating top innovative talent in artificial intelligence. It aims to nurture leading innovators in the AI field by integrating world-class scientific research resources, Shenzhen-Hong Kong cross-border innovation elements, and advanced industrial ecosystems.