9,553 Hits in 4.5 sec

Unsupervised Learning of Compositional Energy Concepts [article]

Yilun Du, Shuang Li, Yash Sharma, Joshua B. Tenenbaum, Igor Mordatch
2021 arXiv   pre-print
So far, unsupervised discovery of concepts has focused on either modeling the global scene-level or the local object-level factors of variation, but not both.  ...  Finally, discovered visual concepts in COMET generalize well, enabling us to compose concepts between separate modalities of images as well as with other concepts discovered by a separate instance of COMET  ...  A.1 Unsupervised Learning of Compositional Energy Concepts Appendix In this supplement, we provide additional empirical visualizations of our approach in Section A.1.1.  ... 
arXiv:2111.03042v1 fatcat:s2qxv34xvfadlebkrhsgk7n53y

Unsupervised Structure Learning: Hierarchical Recursive Composition, Suspicious Coincidence and Competitive Exclusion [chapter]

Long Zhu, Chenxi Lin, Haoda Huang, Yuanhao Chen, Alan Yuille
2008 Lecture Notes in Computer Science  
We describe a new method for unsupervised structure learning of a hierarchical compositional model (HCM) for deformable objects.  ...  The learning is unsupervised in the sense that we are given a training dataset of images containing the object in cluttered backgrounds but we do not know the position or boundary of the object.  ...  This gives an effective way for computing the full energy, by combining the energy of the subgraphs, and is exploited during inference and learning.  ... 
doi:10.1007/978-3-540-88688-4_56 fatcat:2wzvj5hocjgojedfmpqmeapdii

Damage assessment of smart composite structures via machine learning: a review

Asif Khan, Nayeon Kim, Jae Kyong Shin, Heung Soo Kim, Byeng Dong Youn
2019 JMST Advances  
This article focuses on a review of discriminative features and the corresponding machine learning algorithms (both supervised and unsupervised), for various types of damage in smart composite structures  ...  This paper reports on the use of machine learning techniques for the damage assessment (i.e., detection, quantification, and localization) of smart composite structures.  ...  In this review article, supervised and unsupervised machine learning techniques have been reviewed for the damage assessment of smart composite structures, with special emphasis on the types of damage  ... 
doi:10.1007/s42791-019-0012-2 fatcat:xp423kbqhba57pxmyxavuekm34

Learning Paired-associate Images with An Unsupervised Deep Learning Architecture [article]

Ti Wang, Daniel L. Silver
2014 arXiv   pre-print
Experiments show that the multi-modal learning system generates models that are as accurate as back-propagation networks but with the advantage of a bi-directional network and unsupervised learning from  ...  This paper presents an unsupervised multi-modal learning system that learns associative representation from two input modalities, or channels, such that input on one channel will correctly generate the  ...  Abstract concepts are learned and recalled through the composition of simpler concepts [1] .  ... 
arXiv:1312.6171v2 fatcat:3loqpnxftrcsvmsvr4qesawioe

Unsupervised learning models of invariant features in images: Recent developments in multistage architecture approach for object detection

Sonia Mittal
2016 Zenodo  
high level task of object recognition  ...  Here we present a survey of one particular approach that has proved very promising for invariant feature recognition and which is a key initial stage of multi-stage network architecture methods for the  ...  In this type of networks depth of architecture indicates the number of levels of composition of non-linear operations in the function learned.  ... 
doi:10.5281/zenodo.3661859 fatcat:q6naogm5nzd7vf3fautqtiwbpa

Jointly Learning Sentence Embeddings and Syntax with Unsupervised Tree-LSTMs [article]

Jean Maillard, Stephen Clark, Dani Yogatama
2017 arXiv   pre-print
It can therefore be seen as a tree-based RNN that is unsupervised with respect to the parse trees.  ...  We use Tree-LSTM as our composition function, applied along a tree structure found by a fully differentiable natural language chart parser.  ...  The unsupervised Tree-LSTM uses an analogous chart to guide the order of composition.  ... 
arXiv:1705.09189v1 fatcat:p6st2uaddvg2dimsp2u3hk52vm

Element selection for functional materials discovery by integrated machine learning of atomic contributions to properties [article]

Andrij Vasylenko, Dmytro Antypov, Vladimir Gusev, Michael W. Gaultois, Matthew S. Dyer, Matthew J. Rosseinsky
2022 arXiv   pre-print
This end-to-end machine learning approach (PhaseSelect) first derives the atomic characteristics from the compositional environments in all computationally and experimentally explored materials and then  ...  , high-temperature magnetic and targetted energy band gap materials.  ...  We demonstrate that unsupervised learning of chemical elements combined with the attention technique for learning elemental contributions can be used for the accurate classification of the materials' functional  ... 
arXiv:2202.01051v1 fatcat:4wx7lptb2bhghcke4o7sy3tify

Robust Real-Time Music Transcription with a Compositional Hierarchical Model

Matevž Pesek, Aleš Leonardis, Matija Marolt, Constantine Dovrolis
2017 PLoS ONE  
Its main features are unsupervised learning of a hierarchical representation of input data, transparency, which enables insights into the learned representation, as well as robustness and speed which make  ...  The layers are learned in an unsupervised manner from music signals.  ...  The model is constructed by unsupervised learning on a set of audio recordings and contains compositions of parts reflecting the statistical regularities in the learning set, encoding simple concepts on  ... 
doi:10.1371/journal.pone.0169411 pmid:28046074 pmcid:PMC5207709 fatcat:tan77334aze7ndr5jfglaf4l6e

Page 533 of Psychological Abstracts Vol. 9, Issue 10 [page]

1935 Psychological Abstracts  
other aspects (if any) of stu- dents’ writing ability or progress are related to this individuality, (3) to learn how individual characteris- tics express themselves in the writing of compositions, and  ...  Her emotional and motor energies were not satisfied by her musical activities.  ... 

Learning Deep Architectures for AI

Y. Bengio
2009 Foundations and Trends® in Machine Learning  
This monograph discusses the motivations and principles regarding learning algorithms for deep architectures, in particular those exploiting as building blocks unsupervised learning of single-layer models  ...  The focus of deep architecture learning is to automatically discover such abstractions, from the lowest level features to the highest level concepts.  ...  Depth of architecture refers to the number of levels of composition of non-linear operations in the function learned.  ... 
doi:10.1561/2200000006 fatcat:pqujlozkonasra65suwxpvvuou

Knowledge Guided Learning: Towards Open Domain Egocentric Action Recognition with Zero Supervision [article]

Sathyanarayanan N. Aakur, Sanjoy Kundu, Nikhil Gunti
2022 arXiv   pre-print
Advances in deep learning have enabled the development of models that have exhibited a remarkable tendency to recognize and even localize actions in videos.  ...  Building upon the compositional representation offered by Grenander's Pattern Theory formalism, we show that attention and commonsense knowledge can be used to enable the self-supervised discovery of novel  ...  of deep neural networks to overcome the dependence on annotated training data, (iii) we show that using compositional concept representations with Pattern Theory can learn semantic correspondences across  ... 
arXiv:2009.07470v2 fatcat:flrj3hhlune55atxc4h2isv66u

A three-step unsupervised neural model for visualizing high complex dimensional spectroscopic data sets

Emilio Corchado, Juan C. Perez
2010 Pattern Analysis and Applications  
learning (CMLHL), which is characterized by its capability to preserve a degree of global ordering in the data.  ...  such as maximum likelihood Hebbian learning (MLHL) and the application of the SOM without a pre-processing step.  ...  A wide variety of unsupervised learning architectures have been proposed to date.  ... 
doi:10.1007/s10044-010-0187-5 fatcat:hffmlytac5aexfql7scrfflrfy

Jointly learning sentence embeddings and syntax with unsupervised Tree-LSTMs

Jean Maillard, Stephen Clark, Dani Yogatama
2019 Natural Language Engineering  
They can therefore be seen as tree-based recurrent neural networks that are unsupervised with respect to the parse trees.  ...  AbstractWe present two studies on neural network architectures that learn to represent sentences by composing their words according to automatically induced binary trees, without ever being shown a correct  ...  The unsupervised Tree-LSTM uses an analogous chart to guide the order of composition.  ... 
doi:10.1017/s1351324919000184 fatcat:7scvo7uz2rchfpij3kvy7w24ru

National projects and government programmes: functional algorithm for evaluating and modelling using the Data Science methodology

Olim Astanakulov, Academy of Public Administration under President of Republic of Uzbekistan (APA)
2020 Economic Annals-ХХI  
, as well as achievement of goals, development of a system of performance indicators, and so on.  ...  Programme and target planning procedures in Russia have a lot of shortcomings, related to the selection of priority goals, establishment of criteria for evaluating the effectiveness of target programmes  ...  They learn simpler concepts first, and build on them to learn more abstract concepts. This strategy, studied in detail here, has not yet been much exploited in machine learning.  ... 
doi:10.21003/ea.v183-05 fatcat:zgxuu3bwtjfxlkt7l7adky2opq

Recent approaches to non-intrusive load monitoring techniques in residential settings

Yung Fei Wong, Y. Ahmet Sekercioglu, Tom Drummond, Voon Siong Wong
2013 2013 IEEE Computational Intelligence Applications in Smart Grid (CIASG)  
The concept of Smart Grids is closely related to energy conservation and load shedding concepts.  ...  Fluctuations in the aggregate power consumption signals are used to mathematically estimate the composition of operation of appliances.  ...  Unsupervised Learning In contrast, unsupervised learning as its name implies does not require any sort of training procedure before the system goes online.  ... 
doi:10.1109/ciasg.2013.6611501 dblp:conf/ciasg/WongSDW13 fatcat:cn7vlb4ktvbgvpr75xesgoh75m
« Previous Showing results 1 — 15 out of 9,553 results