BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:AI-Powered Graph Representation Learning for Robust and Efficient 
 Urban and Social Science - Qianru Zhang
DTSTART:20250728T160000Z
DTEND:20250728T170000Z
UID:TALK234436@talks.cam.ac.uk
CONTACT:Pietro Lio
DESCRIPTION:The increasing availability of human trajectory and social dat
 a\, fueled by GPS and social networks\, presents a unique opportunity for 
 scientific discovery. However\, existing data analysis methods struggle to
  provide robust\, efficient\, and generalizable graph representations\, hi
 ndering their applicability in urban and social sciences. This research ad
 dresses this challenge by developing novel machine learning algorithms spe
 cifically tailored for graph-structured data in these domains. This resear
 ch tackles three key challenges: (1) Sparse Data and Data Distribution Het
 erogeneity:  Current methods often struggle with sparse data and varying d
 ata distributions\, limiting their ability to capture diverse patterns and
  hindering scalability. This research proposes novel approaches for flexib
 le\, adaptive\, and generalizable representations in urban planning and so
 cial sciences. (2) Non-General Representation and Difficulty Adapting to N
 ew Data:  Existing methods often lack the ability to generalize across dif
 ferent datasets and struggle to adapt to new data\, hindering their effect
 iveness in real-world applications. This research aims to develop methods 
 that can learn robust and efficient representations that generalize across
  different datasets and adapt to new data. (3) Trade-off Between Efficienc
 y and Effectiveness: Balancing processing speed\, accuracy\, and reliabili
 ty is crucial in urban and social science data analysis. This research add
 resses this challenge by developing innovative algorithms that optimize fo
 r both efficiency and effectiveness. This research leverages contrastive l
 earning and information bottleneck techniques to develop robust and effici
 ent graph representation learning methods for spatial-temporal data and re
 commender systems. The developed methods have demonstrated significant imp
 rovements in downstream tasks such as traffic prediction\, crime predictio
 n\, and anomaly detection. This research lays a strong foundation for futu
 re work in graph-structured data analysis across various domains\, includi
 ng urban science\, social science\, and scientific discovery. Future resea
 rch will focus on extending these methods to multi-modal datasets\, enabli
 ng zero-shot learning\, and developing novel approaches for understanding 
 complex biological systems. The Zoom link is shown as follows:https://hku.
 zoom.us/j/93862113743?pwd=5pt5SNvWUtmS7xnjdrzevajlGnwjIw.1
LOCATION:Computer Laboratory\, William Gates Building\, Lecture Theatre 1
END:VEVENT
END:VCALENDAR