BEGIN:VCALENDAR
VERSION:2.0
PRODID:icalendar-ruby
CALSCALE:GREGORIAN
X-WR-CALNAME:CIS Seminar Series  2026 -Mengjia Xu\, Ph.D.-Ying Wu College o
 f Computing\, NJIT
X-WR-TIMEZONE:Eastern Time (US & Canada)
BEGIN:VEVENT
DTSTAMP:20260909T160226Z
UID:tag:localist.com\,2008:EventInstance_52055978270656
DTSTART:20260220T163000Z
DTEND:20260220T173000Z
DESCRIPTION:Hyperbolic Neural Architectures for Hierarchical Language Model
 ing and Brain Network Representation\n\n \n\nABSTRACT\n\nMost deep learnin
 g models are typically built on Euclidean representations\, which are ofte
 n poorly suited for capturing the hierarchical structure and relational or
 ganization that inherent in real-world data such as language\, electronic 
 health record\, and brain connectivity. This talk presents a unified line 
 of work introducing hyperbolic geometry as an inductive bias for structure
 d representation learning in sequence modeling and neuroimaging data analy
 tics.\n\nI will first present an overview of hyperbolic large language mod
 els\, where linguistic hierarchies and compositional dependencies are embe
 dded in negatively curved space to improve representational efficiency and
  long-context reasoning. I then describe hierarchy-aware sequence modeling
  in hyperbolic geometry\, integrating state-space architectures with geome
 tric structure to produce scalable language embeddings. Finally\, I discus
 s fully hyperbolic neural networks for brain network modeling\, where func
 tional brain connectivity is embedded in hyperbolic space to better captur
 e hierarchical organization and inter-regional relationships in neuroimagi
 ng data for aging trajectory detection and subjective cognitive decline de
 tection. \n\nTogether\, these works highlight hyperbolic neural architectu
 res as a geometric foundation for hierarchical reasoning in LLMs and relat
 ional data modeling. I conclude by outlining key challenges and emerging o
 pportunities toward geometry-aware\, interpretable\, and long-context inte
 lligent systems.\n\n \n\nBIOGRAPHY\n\nDr. Mengjia Xu is an Assistant Profe
 ssor in the Department of Data Science at the Ying Wu College of Computing
 \, New Jersey Institute of Technology (NJIT). Her research focuses on mach
 ine learning\, geometric deep learning\, and multimodal large language mod
 els\, with emphasis on graph representation learning and hierarchical reas
 oning for graph-structured\, temporal\, and language data. She develops pr
 incipled learning frameworks for applications\, including neurodegenerativ
 e disease and brain aging\, electronic health records and genomics for can
 cer detection\, and scientific discovery in solar physics. Prior to joinin
 g NJIT\, she held a joint postdoctoral position at the McGovern Institute 
 for Brain Research at MIT and the Division of Applied Mathematics at Brown
  University\, mentored by Prof. Tomaso Poggio and Prof. George Em Karniada
 kis. She received her Ph.D. in Computer Science from Northeastern Universi
 ty (China)\, including  a two-year joint PhD training at Brown University.
GEO:39.680588;-75.754318
LOCATION:Smith Hall\, 120
SUMMARY:CIS Seminar Series  2026 -Mengjia Xu\, Ph.D.-Ying Wu College of Com
 puting\, NJIT
URL;VALUE=URI:https://events.udel.edu/event/cis-seminar-series-2026-mengjia
 -xu-phd-ying-wu-college-of-computing-njit
CATEGORIES:Academics
CATEGORIES:College of Engineering
END:VEVENT
END:VCALENDAR