Online MAP Inference and Learning for Nonsymmetric Determinantal Point\n Processes

2021/11/29 by A. Rama Mohan Reddy, Reddy, Aravind, Ryan A. Rossi +15
Computer Science · Mathematics · #Artificial Intelligence (cs.AI) #Bayesian Methods and Mixture Models #Data Structures and Algorithms (cs.DS) #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Markov Chains and Monte Carlo Methods #Point processes and geometric inequalities

paper · pdf · doi:10.48550/arxiv.2111.14674

openalex publication_date 2021/11/29 · openalex created_date 2022/11/07 · openalex updated_date 2026/07/28

Abstract

In this paper, we introduce the online and streaming MAP inference and\nlearning problems for Non-symmetric Determinantal Point Processes (NDPPs) where\ndata points arrive in an arbitrary order and the algorithms are constrained to\nuse a single-pass over the data as well as sub-linear memory. The online\nsetting has an additional requirement of maintaining a valid solution at any\npoint in time. For solving these new problems, we propose algorithms with\ntheoretical guarantees, evaluate them on several real-world datasets, and show\nthat they give comparable performance to state-of-the-art offline algorithms\nthat store the entire data in memory and take multiple passes over it.\n

Related

• Scalable Learning and MAP Inference for Nonsymmetric Determinantal Point\n Processes
• Faster Greedy MAP Inference for Determinantal Point Processes
• Lazy and Fast Greedy MAP Inference for Determinantal Point Process
• Learning Nonsymmetric Determinantal Point Processes
• Fixed-point algorithms for learning determinantal point processes