Singapore Management University

LMFLab for Multimodality
and Forecasting

We connect multimodal understanding and modeling with forecasting at macro and micro scales, grounded in real-world problems.

Led by Yunshan Ma · School of Computing and Information Systems

Meet the faculty, postdoc, PhD students, visiting PhD students, and alumni of LMF.

What we study

01

Multimodality

Understanding and fusing text, images, audio, graphs, and symbolic data; modeling how their elements compose and combine.

Explore this direction →
02

Forecasting

Anticipating macro-level events and micro-level behavior, from financial markets and geopolitics to users and cyber attacks.

Explore this direction →

Where we apply it

Fashion

Fashion understanding, personalized styling, outfit and wardrobe generation, and trend forecasting.

Finance

Financial forecasting, explainable stock prediction, time-series reasoning, and financial analysis.

Cybersecurity

Cyber threat intelligence, attack graph construction, and cyber attack sequence prediction.

Social Events

Understanding and forecasting social and geopolitical events through graph modeling, open-ended hypothesis generation, and multi-expert reasoning.

Music

Connecting symbolic music and audio with language models for music understanding and personalization.

Connect with LMF

For research inquiries, contact Yunshan Ma.

ysma@smu.edu.sg