Online Structure Learning for Sum-Product Networks with Gaussian Leaves

Hsu, Wilson; Kalra, Agastya; Poupart, Pascal

doi:10.48550/arxiv.1701.05265

Cited by 5 publications

(4 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While it is possible to specify SPNs by hand, weight learning is additionally required to obtain a probability distribution, but also the specification of SPNs has to obey conditions of completeness and decomposability, all of which makes structure learning an obvious choice. Since SPNs were introduced, a number of structure learning frameworks have been developed for those and related data structures, e.g., [20,24,33].…”

Section: Related Workmentioning

confidence: 99%

Fairness in machine learning with tractable models

Varley

Belle

2021

Knowledge-Based Systems

View full text Add to dashboard Cite

Section: Related Workmentioning

confidence: 99%

Fairness in machine learning with tractable models

Varley

Belle

2021

Knowledge-Based Systems

View full text Add to dashboard Cite

“…SPNs model joint or conditional distributions and can be learned generatively [15] or discriminatively [16] using Expectation Maximization (EM) or gradient descent (GD). Additionally, several algorithms were proposed for simultaneous learning of network parameters and structure [21][22] [23]. In this work, we use a simple structure learning technique [12] which begins by initializing the SPN with a random dense structure that is later pruned.…”

Section: A Sum-product Networkmentioning

confidence: 99%

From Pixels to Buildings: End-to-end Probabilistic Deep Networks for Large-scale Semantic Mapping

Zheng

Pronobis

2019

2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

View full text Add to dashboard Cite

We introduce TopoNets, end-to-end probabilistic deep networks for modeling semantic maps with structure reflecting the topology of large-scale environments. TopoNets build a unified deep network spanning multiple levels of abstraction and spatial scales, from pixels representing geometry of local places to high-level descriptions of semantics of buildings. To this end, TopoNets leverage complex spatial relations expressed in terms of arbitrary, dynamic graphs. We demonstrate how TopoNets can be used to perform end-toend semantic mapping from partial sensory observations and noisy topological relations discovered by a robot exploring large-scale office spaces. Thanks to their probabilistic nature and generative properties, TopoNets extend the problem of semantic mapping beyond classification. We show that TopoNets successfully perform uncertain reasoning about yet unexplored space and detect novel and incongruent environment configurations unknown to the robot. Our implementation of TopoNets achieves real-time, tractable and exact inference, which makes these new deep models a promising, practical solution to mobile robot spatial understanding at scale.

show abstract

“…Parameters of an SPN can be learned generatively (Poon and Domingos 2011) or discriminatively (Gens and Domingos 2012) using Expectation Maximization (EM) or gradient descent. Additionally, several algorithms were proposed for simultaneous learning of network parameters and structure (Hsu, Kalra, and Poupart 2017;Gens and Domingos 2013). In this work, we use a simple structure learning technique to build template SPNs.…”

Section: Sum-product Networkmentioning

confidence: 99%

Learning Graph-Structured Sum-Product Networks for Probabilistic Semantic Maps

Zheng

Pronobis

Rao

2018

AAAI

View full text Add to dashboard Cite

We introduce Graph-Structured Sum-Product Networks (GraphSPNs), a probabilistic approach to structured prediction for problems where dependencies between latent variables are expressed in terms of arbitrary, dynamic graphs. While many approaches to structured prediction place strict constraints on the interactions between inferred variables, many real-world problems can be only characterized using complex graph structures of varying size, often contaminated with noise when obtained from real data. Here, we focus on one such problem in the domain of robotics. We demonstrate how GraphSPNs can be used to bolster inference about semantic, conceptual place descriptions using noisy topological relations discovered by a robot exploring large-scale office spaces. Through experiments, we show that GraphSPNs consistently outperform the traditional approach based on undirected graphical models, successfully disambiguating information in global semantic maps built from uncertain, noisy local evidence. We further exploit the probabilistic nature of the model to infer marginal distributions over semantic descriptions of as yet unexplored places and detect spatial environment configurations that are novel and incongruent with the known evidence.

show abstract

Online Structure Learning for Sum-Product Networks with Gaussian Leaves

Cited by 5 publications

References 7 publications

Fairness in machine learning with tractable models

Fairness in machine learning with tractable models

From Pixels to Buildings: End-to-end Probabilistic Deep Networks for Large-scale Semantic Mapping

Learning Graph-Structured Sum-Product Networks for Probabilistic Semantic Maps

Contact Info

Product

Resources

About