scholarly journals Efficient Inference for Untied MLNs

2017 ◽  
Author(s):  
Somdeb Sarkhel ◽  
Deepak Venugopal ◽  
Nicholas Ruozzi ◽  
Vibhav Gogate
Keyword(s):  
Graphical Models ◽  
Scaling Up ◽  
Partition Functions ◽  
Markov Logic ◽  
Practical Applications ◽  
Lifted Inference ◽  
Hard Time ◽  
Inference Algorithms ◽  
Order Of Magnitude ◽  
Symmetric Approximation

We address the problem of scaling up local-search or sampling-based inference in Markov logic networks (MLNs) that have large shared sub-structures but no (or few) tied weights. Such untied MLNs are ubiquitous in practical applications. However, they have very few symmetries, and as a result lifted inference algorithms--the dominant approach for scaling up inference--perform poorly on them. The key idea in our approach is to reduce the hard, time-consuming sub-task in sampling algorithms, computing the sum of weights of features that satisfy a full assignment, to the problem of computing a set of partition functions of graphical models, each defined over the logical variables in a first-order formula. The importance of this reduction is that when the treewidth of all the graphical models is small, it yields an order of magnitude speedup. When the treewidth is large, we propose an over-symmetric approximation and experimentally demonstrate that it is both fast and accurate.

scholarly journals Lifted Variable Elimination: Decoupling the Operators from the Constraint Language

2013 ◽  
Vol 47 ◽  
pp. 393-439 ◽  
Author(s):  
N. Taghipour ◽  
D. Fierens ◽  
J. Davis ◽  
H. Blockeel
Keyword(s):  
Graphical Models ◽  
Probabilistic Inference ◽  
Relational Algebra ◽  
Exact Inference ◽  
Semantic Level ◽  
Lifted Inference ◽  
Variable Elimination ◽  
Constraint Language ◽  
Inference Algorithms ◽  
Powerful Constraint

Lifted probabilistic inference algorithms exploit regularities in the structure of graphical models to perform inference more efficiently. More specifically, they identify groups of interchangeable variables and perform inference once per group, as opposed to once per variable. The groups are defined by means of constraints, so the flexibility of the grouping is determined by the expressivity of the constraint language. Existing approaches for exact lifted inference use specific languages for (in)equality constraints, which often have limited expressivity. In this article, we decouple lifted inference from the constraint language. We define operators for lifted inference in terms of relational algebra operators, so that they operate on the semantic level (the constraints' extension) rather than on the syntactic level, making them language-independent. As a result, lifted inference can be performed using more powerful constraint languages, which provide more opportunities for lifting. We empirically demonstrate that this can improve inference efficiency by orders of magnitude, allowing exact inference where until now only approximate inference was feasible.

scholarly journals Lifted Message Passing for Hybrid Probabilistic Inference

2019 ◽  
Author(s):  
Yuqiao Chen ◽  
Nicholas Ruozzi ◽  
Sriraam Natarajan
Keyword(s):  
Message Passing ◽  
Computational Cost ◽  
Markov Logic Networks ◽  
Markov Logic ◽  
Lifted Inference ◽  
Particle Sampling ◽  
Variable Elimination ◽  
Inference Algorithms ◽  
Arbitrary Potential ◽  
Inference Methods

Lifted inference algorithms for first-order logic models, e.g., Markov logic networks (MLNs), have been of significant interest in recent years.  Lifted inference methods exploit model symmetries in order to reduce the size of the model and, consequently, the computational cost of inference.  In this work, we consider the problem of lifted inference in MLNs with continuous or both discrete and continuous groundings. Existing work on lifting with continuous groundings has mostly been limited to special classes of models, e.g., Gaussian models, for which variable elimination or message-passing updates can be computed exactly.  Here, we develop approximate lifted inference schemes based on particle sampling.  We demonstrate empirically that our approximate lifting schemes perform comparably to existing state-of-the-art for models for Gaussian MLNs, while having the flexibility to be applied to models with arbitrary potential functions.

scholarly journals On Lifted Inference Using Neural Embeddings

2019 ◽  
Vol 33 ◽  
pp. 7916-7923
Author(s):  
Mohammad Maminur Islam ◽  
Somdeb Sarkhel ◽  
Deepak Venugopal
Keyword(s):  
Neural Networks ◽  
State Of The Art ◽  
Markov Logic Networks ◽  
Markov Logic ◽  
Lifted Inference ◽  
Inference Algorithms ◽  
Art Methods ◽  
Dense Representation

We present a dense representation for Markov Logic Networks (MLNs) called Obj2Vec that encodes symmetries in the MLN structure. Identifying symmetries is a key challenge for lifted inference algorithms and we leverage advances in neural networks to learn symmetries which are hard to specify using hand-crafted features. Specifically, we learn an embedding for MLN objects that predicts the context of an object, i.e., objects that appear along with it in formulas of the MLN, since common contexts indicate symmetry in the distribution. Importantly, our formulation leverages well-known skip-gram models that allow us to learn the embedding efficiently. Finally, to reduce the size of the ground MLN, we sample objects based on their learned embeddings. We integrate Obj2Vec with several inference algorithms, and show the scalability and accuracy of our approach compared to other state-of-the-art methods.

scholarly journals Model-based machine learning

2013 ◽  
Vol 371 (1984) ◽  
pp. 20120222 ◽  
Author(s):  
Christopher M. Bishop
Keyword(s):  
Machine Learning ◽  
Graphical Models ◽  
Large Scale ◽  
Commercial Application ◽  
Software Environment ◽  
Probabilistic Programming ◽  
Practical Applications ◽  
Model Based ◽  
Inference Algorithms ◽  
Modelling Environment

Several decades of research in the field of machine learning have resulted in a multitude of different algorithms for solving a broad range of problems. To tackle a new application, a researcher typically tries to map their problem onto one of these existing methods, often influenced by their familiarity with specific algorithms and by the availability of corresponding software implementations. In this study, we describe an alternative methodology for applying machine learning, in which a bespoke solution is formulated for each new application. The solution is expressed through a compact modelling language, and the corresponding custom machine learning code is then generated automatically. This model-based approach offers several major advantages, including the opportunity to create highly tailored models for specific scenarios, as well as rapid prototyping and comparison of a range of alternative models. Furthermore, newcomers to the field of machine learning do not have to learn about the huge range of traditional methods, but instead can focus their attention on understanding a single modelling environment. In this study, we show how probabilistic graphical models, coupled with efficient inference algorithms, provide a very flexible foundation for model-based machine learning, and we outline a large-scale commercial application of this framework involving tens of millions of users. We also describe the concept of probabilistic programming as a powerful software environment for model-based machine learning, and we discuss a specific probabilistic programming language called Infer.NET , which has been widely used in practical applications.

Diagnostics of metal samples using the results of experimental and theoretical study of positively charged microparticles formed upon sample destruction

2018 ◽  
Vol 84 (10) ◽  
pp. 23-28
Author(s):  
D. A. Golentsov ◽  
A. G. Gulin ◽  
Vladimir A. Likhter ◽  
K. E. Ulybyshev
Keyword(s):  
Experimental Data ◽  
Early Stage ◽  
Experimental Studies ◽  
Material Model ◽  
Total Charge ◽  
Cross Sectional ◽  
Practical Applications ◽  
Mechanical And Electrical Properties ◽  
Order Of Magnitude ◽  
Using Data

Destruction of bodies is accompanied by formation of both large and microscopic fragments. Numerous experiments on the rupture of different samples show that those fragments carry a positive electric charge. his phenomenon is of interest from the viewpoint of its potential application to contactless diagnostics of the early stage of destruction of the elements in various technical devices. However, the lack of understanding the nature of this phenomenon restricts the possibility of its practical applications. Experimental studies were carried out using an apparatus that allowed direct measurements of the total charge of the microparticles formed upon sample rupture and determination of their size and quantity. The results of rupture tests of duralumin and electrical steel showed that the size of microparticles is several tens of microns, the particle charge per particle is on the order of 10–14 C, and their amount can be estimated as the ratio of the cross-sectional area of the sample at the point of discontinuity to the square of the microparticle size. A model of charge formation on the microparticles is developed proceeding from the experimental data and current concept of the electron gas in metals. The model makes it possible to determine the charge of the microparticle using data on the particle size and mechanical and electrical properties of the material. Model estimates of the total charge of particles show order-of-magnitude agreement with the experimental data.

Scaling Up Markov Logic Probabilistic Inference for Social Graphs

2017 ◽  
Vol 29 (2) ◽  
pp. 433-445 ◽  
Author(s):  
Haiquan Chen ◽  
Wei-Shinn Ku ◽  
Haixun Wang ◽  
Liang Tang ◽  
Min-Te Sun
Keyword(s):  
Probabilistic Inference ◽  
Scaling Up ◽  
Markov Logic ◽  
Social Graphs

scholarly journals Engineering of self-rectifying filamentary resistive switching in LiNbO3 single crystalline thin film via strain doping

2019 ◽  
Vol 9 (1) ◽  
Author(s):  
Tiangui You ◽  
Kai Huang ◽  
Xiaomeng Zhao ◽  
Ailun Yi ◽  
Chen Chen ◽  
...  
Keyword(s):  
Thin Films ◽  
Resistive Switching ◽  
Schottky Contact ◽  
Homogeneous Distribution ◽  
Single Crystalline ◽  
Wafer Scale ◽  
Practical Applications ◽  
Crystalline Oxide ◽  
Order Of Magnitude ◽  
Switching Characteristics

AbstractThe abilities to fabricate wafer scale single crystalline oxide thin films on metallic substrates and to locally engineer their resistive switching characteristics not only contribute to the fundamental investigations of the resistive switching mechanism but also promote the practical applications of resistive switching devices. Here, wafer scale LiNbO3 (LNO) single crystalline thin films are fabricated on Pt/SiO2/LNO substrates by ion slicing with wafer bonding. The lattice strain of the LNO single crystalline thin films can be tuned by He implantation as indicated by XRD measurements. After He implantation, the LNO single crystalline thin films show self-rectifying filamentary resistive switching behaviors, which is interpreted by a model that the local conductive filaments only connect/disconnect with the bottom interface while the top interface maintains the Schottky contact. Thanks to the homogeneous distribution of defects in single crystalline thin films, highly reproducible and uniform self-rectifying resistive switching with large on/off ratio over four order of magnitude was achieved. Multilevel resistive switching can be obtained by varying the compliance current or by using different magnitude of writing voltage.

scholarly journals THE CAPPELLI–ITZYKSON–ZUBER A-D-E CLASSIFICATION

2000 ◽  
Vol 12 (05) ◽  
pp. 739-748 ◽  
Author(s):  
TERRY GANNON
Keyword(s):  
Mathematical Structure ◽  
The Other ◽  
Field Theories ◽  
Partition Functions ◽  
Conformal Field Theories ◽  
Modular Invariant ◽  
Conformal Field ◽  
Affine Algebras ◽  
Order Of Magnitude ◽  
The Rich

In 1986 Cappelli, Itzykson and Zuber classified all modular invariant partition functions for the conformal field theories associated to the affine A1 algebra; they found they fall into an A-D-E pattern. Their proof was difficult and attempts to generalise it to the other affine algebras failed — in hindsight the reason is that their argument ignored most of the rich mathematical structure present. We give here the "modern" proof of their result; it is an order of magnitude simpler and shorter, and much of it has already been extended to all other affine algebras. We conclude with some remarks on the A-D-E pattern appearing in this and other RCFT classifications.

scholarly journals Quantifying the multi-scale performance of network inference algorithms

2014 ◽  
Vol 13 (5) ◽  
Author(s):  
Chris J. Oates ◽  
Richard Amos ◽  
Simon E.F. Spencer
Keyword(s):  
Graphical Models ◽  
Network Inference ◽  
Multiple Scales ◽  
Biological Systems ◽  
Higher Order ◽  
Wisdom Of Crowds ◽  
Multi Scale ◽  
Reverse Engineer ◽  
Multiple Length Scales ◽  
Inference Algorithms

AbstractGraphical models are widely used to study complex multivariate biological systems. Network inference algorithms aim to reverse-engineer such models from noisy experimental data. It is common to assess such algorithms using techniques from classifier analysis. These metrics, based on ability to correctly infer individual edges, possess a number of appealing features including invariance to rank-preserving transformation. However, regulation in biological systems occurs on multiple scales and existing metrics do not take into account the correctness of higher-order network structure. In this paper novel performance scores are presented that share the appealing properties of existing scores, whilst capturing ability to uncover regulation on multiple scales. Theoretical results confirm that performance of a network inference algorithm depends crucially on the scale at which inferences are to be made; in particular strong local performance does not guarantee accurate reconstruction of higher-order topology. Applying these scores to a large corpus of data from the DREAM5 challenge, we undertake a data-driven assessment of estimator performance. We find that the “wisdom of crowds” network, that demonstrated superior local performance in the DREAM5 challenge, is also among the best performing methodologies for inference of regulation on multiple length scales.

Wafer-scale heterostructured piezoelectric bio-organic thin films

Science ◽  
2021 ◽  
Vol 373 (6552) ◽  
pp. 337-342
Author(s):  
Fan Yang ◽  
Jun Li ◽  
Yin Long ◽  
Ziyi Zhang ◽  
Linfeng Wang ◽  
...  
Keyword(s):  
Thin Films ◽  
Electricity Generation ◽  
Large Scale ◽  
Organic Thin Films ◽  
Electrical Energy ◽  
Wafer Scale ◽  
Practical Applications ◽  
Piezoelectric Phase ◽  
Order Of Magnitude

Piezoelectric biomaterials are intrinsically suitable for coupling mechanical and electrical energy in biological systems to achieve in vivo real-time sensing, actuation, and electricity generation. However, the inability to synthesize and align the piezoelectric phase at a large scale remains a roadblock toward practical applications. We present a wafer-scale approach to creating piezoelectric biomaterial thin films based on γ-glycine crystals. The thin film has a sandwich structure, where a crystalline glycine layer self-assembles and automatically aligns between two polyvinyl alcohol (PVA) thin films. The heterostructured glycine-PVA films exhibit piezoelectric coefficients of 5.3 picocoulombs per newton or 157.5 × 10−3 volt meters per newton and nearly an order of magnitude enhancement of the mechanical flexibility compared with pure glycine crystals. With its natural compatibility and degradability in physiological environments, glycine-PVA films may enable the development of transient implantable electromechanical devices.

Sign in / Sign up

Export Citation Format

Share Document