Accepted Papers
Factor Investing with a Deep Multi-Factor Model.pdfFactor Investing with a Deep Multi-Factor Model
Zikai Wei (The Chinese University of Hong Kong)*; Bo Dai (Shanghai AI Lab); Dahua Lin (The Chinese University of Hong Kong)
Understanding stock market instability via graph auto-encoders.pdfUnderstanding stock market instability via graph auto-encoders
Dragos Gorduza (University of Oxford)*; Xiaowen Dong (Oxford); Stefan Zohren (Oxford)
Improving Protein Subcellular Localization Prediction with Structural Prediction & Graph Neural Networks.pdfImproving Protein Subcellular Localization Prediction with Structural Prediction & Graph Neural Networks
Geoffroy Dubourg-Felonneau (Shiru, INC)*; Arash Abbasi (Shiru, INC); Eyal Akiva (Shiru, INC); Lawrence Lee (Shiru, INC)
Homological Neural Networks.pdfHomological Neural Networks
Tomaso Aste (University College London)*; Yuanrong Wang (UCL)
Learning on Graphs for Mineral Asset Valuation Under Supply and Demand Uncertainty .pdfLearning on Graphs for Mineral Asset Valuation Under Supply and Demand Uncertainty
Yassine Yaakoubi (McGill University)*; Hager Radi (Mila, Quebec AI Institute); Roussos Dimitrakopoulos (McGill University)
Graph Q-Learning for Combinatorial Optimization.pdfGraph Q-Learning for Combinatorial Optimization
Victoria M Dax (Stanford University)*; Jiachen Li (Stanford University); Kevin Leahy (MIT Lincoln Laboratory); Mykel J Kochenderfer (Stanford University)
Dual GNNs: Learning Graph Neural Networks with Limited Supervision.pdfDual GNNs: Learning Graph Neural Networks with Limited Supervision
Abdullah Alchihabi (Carleton University )*; Yuhong Guo (Carleton University)
Battery GraphNets : Relational Learning for Lithium-ion Batteries(LiBs) Life Estimation.pdfBattery GraphNets: Relational Learning for Lithium-ion Batteries(LiBs) Life Estimation
Sagar Srinivas Sakhinana (Tata Research Development and Design Center)*; Rajat Kumar Sarkar (Tata Research Development and Design Centre); Venkataramana Runkana (Tata Consultancy Services Limited)
Dissecting In-the-Wild Stress from Multimodal Sensor Data.pdfDissecting In-the-Wild Stress from Multimodal Sensor Data
Sujay Nagaraj (University of Toronto)*; Thomas Hartvigsen (MIT); Adrian Boch (Evidation); Luca Foschini (Evidation); Marzyeh Ghassemi (University of Toronto, Vector Institute); Sarah Goodday (4youandme); Stephen Friend (4youandme); Anna Goldenberg (SickKids Research Institute/University of Toronto)
Sample-Specific Contextualized Graphical Models Using Clinical and Molecular Data Reveal Transcriptional Network Heterogeneity
Caleb Ellington (Carnegie Mellon University)*; Benjamin J Lengerich (MIT); Thomas B.K. Watkins (UCL Cancer Institute); Jiekun Yang (MIT); Manolis Kellis (Massachusetts Institute of Technology); Eric P. Xing (MBZUAI, CMU, and Petuum Inc.)
Abstract: Cancers are shaped by somatic mutations, microenvironment, and patient background, each altering both gene expression and gene regulatory networks (GRNs) in complex ways, resulting in highly-variable cellular states and dynamics. Inferring GRNs from expression data can help characterize this regulation-driven heterogeneity, but network inference is intractable without many statistical samples, limiting GRNs to cluster-level analyses that ignore intra-cluster heterogeneity. We propose to move beyond cluster-based analyses by using contextualized learning, a meta-learning paradigm, to generate sample-specific network models from sample contexts. We unify three network classes (correlation, Markov, Bayesian) and estimate sample-specific GRNs for 7000 tumours across 25 tumor types, with each network contextualized by copy number and driver mutation profiles, tumor microenvironment, and patient demographics. Sample-specific networks provide a de-noised view of expression dynamics at sample-specific resolution, which reveal co-expression modules in correlation networks (CNs), cliques and independent regulatory elements in Markov Networks (MNs), and topological ordering and probability factorization in Bayesian Networks (BNs). Sample-specific networks enable GRN-based precision oncology, including brain tumor subtyping that improves survival prognosis.