On Nov. 8, 2023, Professor Qiaoyu Tan gave a speech on machine learning as part of the Research and Life Lessons Series. The event was an insightful exploration into the evolving landscape of machine learning, particularly in the context of a connected world. It commenced with a comprehensive introduction to cutting-edge deep neural networks that leverage interconnected data points or relationships among entities, departing from the traditional paradigm of independent data modeling.

The presentation highlighted the transformative potential of this approach in various domains such as social media, transportation systems, and e-commerce platforms. By harnessing interconnectedness, these innovative models demonstrated their capacity to learn from neighboring data points, akin to learning from friends in a network.

Moreover, the discussion delved into the integration of textual information from vast sources available on the internet, emphasizing the role of large language models like ChatGPT in augmenting the capabilities of deep neural networks. The utilization of textual data was showcased as a significant enhancement, particularly in the era of expansive language models, opening new possibilities for enhancing machine learning algorithms.

Concluding the event, the speaker shared valuable insights and experiences regarding the preparation for graduate studies. 

Overall, the event was an enlightening journey through the evolving methodologies and applications in machine learning, offering attendees a deeper understanding of leveraging connectedness and textual information to advance the capabilities of modern models.

Author: Junhao Zhu

Photo credited to Junhao Zhu