• Corpus ID: 43357859

Learning Semantic Maps with Topological Spatial Relations Using Graph-Structured Sum-Product Networks

@article{Zheng2017LearningSM,
  title={Learning Semantic Maps with Topological Spatial Relations Using Graph-Structured Sum-Product Networks},
  author={Kaiyu Zheng and Andrzej Pronobis and Rajesh P. N. Rao},
  journal={ArXiv},
  year={2017},
  volume={abs/1709.08274}
}
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… 

Figures and Tables from this paper

Sum-Product Networks: A Survey

TLDR
A survey of SPNs, including their definition, the main algorithms for inference and learning from data, several applications, a brief review of software libraries, and a comparison with related models are offered.

Semantic Mapping Based on Spatial Concepts for Grounding Words Related to Places in Daily Environments

TLDR
A novel statistical semantic mapping method called SpCoMapping is proposed, which integrates probabilistic spatial concept acquisition based on multimodal sensor information and a Markov random field applied for learning the arbitrary shape of a place on a map.

SPNC: An Open-Source MLIR-Based Compiler for Fast Sum-Product Network Inference on CPUs and GPUs

TLDR
SPNC is presented, the first tool flow for generating fast native code for SPN inference on both CPUs and GPUs, including the use of vectorized/SIMD execution.

References

SHOWING 1-10 OF 27 REFERENCES

Learning Relational Sum-Product Networks

TLDR
This paper introduces Relational Sum-Product Networks (RSPNs), a new tractable first-order probabilistic architecture that generalizes SPNs by modeling a set of instances jointly, allowing them to influence each other's probability distributions, as well as modeling probabilities of relations between objects.

Large-scale semantic mapping and reasoning with heterogeneous modalities

TLDR
A probabilistic framework combining heterogeneous, uncertain, information such as object observations, shape, size, appearance of rooms and human input for semantic mapping that relies on the concept of spatial properties which make the semantic map more descriptive, and the system more scalable and better adapted for human interaction.

Voronoi Random Fields: Extracting Topological Structure of Indoor Environments via Place Labeling

TLDR
This paper introduces Voronoi random fields (VRFs), a novel technique for mapping the topological structure of indoor environments and shows that the technique is able to label unknown environments based on parameters learned from other environments.

Deep Spatial Affordance Hierarchy : Spatial Knowledge Representation for Planning in Large-scale Environments

TLDR
DASH is designed to represent space from the perspective of a mobile robot executing complex behaviors in the environment, and directly encodes gaps in knowledge and spatial affordances, and leverage the deep model of generic spatial concepts to infer latent and missing information at all abstraction levels.

On the Latent Variable Interpretation in Sum-Product Networks

TLDR
It is shown that the Viterbi-style algorithm for MPE proposed in literature was never proven to be correct, and is found to be a sound derivation of the EM algorithm for SPNs when applied to augmented SPNs.

Online Structure Learning for Sum-Product Networks with Gaussian Leaves

TLDR
This paper describes the first online structure learning technique for continuous SPNs with Gaussian leaves and introduces an accompanying new parameter learning technique.

Supervised semantic labeling of places using information extracted from sensor data

Learning the Structure of Sum-Product Networks

TLDR
This work proposes the first algorithm for learning the structure of SPNs that takes full advantage of their expressiveness, and shows that the learned SPNs are typically comparable to graphical models in likelihood but superior in inference speed and accuracy.

Structured Prediction Energy Networks

TLDR
This work introduces structured prediction energy networks (SPENs), a flexible framework for structured prediction that is able to apply to multi-label problems with substantially larger label sets than previous applications of structured prediction, while modeling high-order interactions using minimal structural assumptions.

Learning deep generative spatial models for mobile robots

TLDR
Experiments on laser-range data from a mobile robot show that the proposed universal model obtains performance superior to state-of-the-art models fine-tuned to one specific task, such as Generative Adversarial Networks (GANs) or SVMs.