Unnamed: 0.1
int64
0
41k
Unnamed: 0
int64
0
41k
author
stringlengths
9
1.39k
id
stringlengths
11
18
summary
stringlengths
25
3.66k
title
stringlengths
4
258
year
int64
1.99k
2.02k
arxiv_url
stringlengths
32
39
info
stringlengths
523
3.18k
embeddings
stringlengths
16.9k
17.1k
700
700
['A. J. Tallón-Ballesteros', 'P. A. Gutiérrez-Peña', 'C. Hervás-Martínez']
1205.3336v1
This paper deals with the distributed processing in the search for an optimum classification model using evolutionary product unit neural networks. For this distributed search we used a cluster of computers. Our objective is to obtain a more efficient design than those net architectures which do not use a distributed p...
Distribution of the search of evolutionary product unit neural networks for classification
2,012
http://arxiv.org/pdf/1205.3336v1
Title Distribution search evolutionary product unit neural network classification Summary paper deal distributed processing search optimum classification model using evolutionary product unit neural network distributed search used cluster computer objective obtain efficient design net architecture use distributed proce...
[-0.017015954479575157, 0.04218994826078415, -0.0480910949409008, 0.035894256085157394, -0.011669846251606941, -0.019090872257947922, 0.07465261965990067, 0.0159330852329731, -0.03907737880945206, -0.002233824459835887, 0.02942030131816864, -0.011982315219938755, 0.005564785096794367, 0.051176197826862335, 0.0086611984...
701
701
['Baochen Sun', 'Jiashi Feng', 'Kate Saenko']
1612.01939v1
In this chapter, we present CORrelation ALignment (CORAL), a simple yet effective method for unsupervised domain adaptation. CORAL minimizes domain shift by aligning the second-order statistics of source and target distributions, without requiring any target labels. In contrast to subspace manifold methods, it aligns t...
Correlation Alignment for Unsupervised Domain Adaptation
2,016
http://arxiv.org/pdf/1612.01939v1
Title Correlation Alignment Unsupervised Domain Adaptation Summary chapter present CORrelation ALignment CORAL simple yet effective method unsupervised domain adaptation CORAL minimizes domain shift aligning secondorder statistic source target distribution without requiring target label contrast subspace manifold metho...
[-0.003157225204631686, 0.055424828082323074, -0.01645541936159134, 0.009064463898539543, -0.008572671562433243, 0.014969167299568653, 0.03380617871880531, -0.011032774113118649, -0.016263090074062347, 0.012759760022163391, -0.0734294205904007, 0.04532918706536293, 0.027217239141464233, 0.01444198563694954, -0.04497967...
702
702
['Gundram Leifert', 'Tobias Strauß', 'Tobias Grüning', 'Roger Labahn']
1605.08412v1
We describe CITlab's recognition system for the HTRtS competition attached to the 13. International Conference on Document Analysis and Recognition, ICDAR 2015. The task comprises the recognition of historical handwritten documents. The core algorithms of our system are based on multi-dimensional recurrent neural netwo...
CITlab ARGUS for historical handwritten documents
2,016
http://arxiv.org/pdf/1605.08412v1
Title CITlab ARGUS historical handwritten document Summary describe CITlabs recognition system HTRtS competition attached 13 International Conference Document Analysis Recognition ICDAR 2015 task comprises recognition historical handwritten document core algorithm system based multidimensional recurrent neural network ...
[0.006485320162028074, 0.011991418898105621, 0.03605543076992035, 0.10382351279258728, -0.054830402135849, 0.0312416423112154, 0.047403573989868164, 0.05970511958003044, 0.031199734658002853, -0.02386263944208622, 0.013471421785652637, -0.02181967720389366, 0.04529687762260437, -0.0016658701933920383, -0.03147829324007...
703
703
['Keyu Lu', 'Jian Li', 'Xiangjing An', 'Hangen He']
1610.09609v1
Vision-based object detection is one of the fundamental functions in numerous traffic scene applications such as self-driving vehicle systems and advance driver assistance systems (ADAS). However, it is also a challenging task due to the diversity of traffic scene and the storage, power and computing source limitations...
Generalized Haar Filter based Deep Networks for Real-Time Object Detection in Traffic Scene
2,016
http://arxiv.org/pdf/1610.09609v1
Title Generalized Haar Filter based Deep Networks RealTime Object Detection Traffic Scene Summary Visionbased object detection one fundamental function numerous traffic scene application selfdriving vehicle system advance driver assistance system ADAS However also challenging task due diversity traffic scene storage po...
[-0.026212334632873535, 0.0003751225012820214, -0.013209350407123566, 0.05791609361767769, -0.042799949645996094, 0.0005910324980504811, 0.03361477330327034, -0.002968650311231613, -0.0026382768992334604, -0.003895937232300639, 0.03201752156019211, -0.004890859127044678, -0.023898780345916748, 0.08317878097295761, 0.01...
704
704
['Weishan Dong', 'Ting Yuan', 'Kai Yang', 'Changsheng Li', 'Shilei Zhang']
1701.01272v1
In this paper, we study learning generalized driving style representations from automobile GPS trip data. We propose a novel Autoencoder Regularized deep neural Network (ARNet) and a trip encoding framework trip2vec to learn drivers' driving styles directly from GPS records, by combining supervised and unsupervised fea...
Autoencoder Regularized Network For Driving Style Representation Learning
2,017
http://arxiv.org/pdf/1701.01272v1
Title Autoencoder Regularized Network Driving Style Representation Learning Summary paper study learning generalized driving style representation automobile GPS trip data propose novel Autoencoder Regularized deep neural Network ARNet trip encoding framework trip2vec learn driver driving style directly GPS record combi...
[-0.03630707785487175, 0.038288988173007965, -0.03017522022128105, 0.04267675429582596, 0.026386911049485207, 0.012322071008384228, 0.02979334257543087, -0.004941744729876518, -0.04290768876671791, -0.02354791760444641, 0.02713371440768242, 0.01855294406414032, 0.0042855404317379, 0.04404690861701965, 0.058824878185987...
705
705
['Prutha Date', 'Ashwinkumar Ganesan', 'Tim Oates']
1707.09899v1
Convolutional Neural Networks have been highly successful in performing a host of computer vision tasks such as object recognition, object detection, image segmentation and texture synthesis. In 2015, Gatys et. al [7] show how the style of a painter can be extracted from an image of the painting and applied to another ...
Fashioning with Networks: Neural Style Transfer to Design Clothes
2,017
http://arxiv.org/pdf/1707.09899v1
Title Fashioning Networks Neural Style Transfer Design Clothes Summary Convolutional Neural Networks highly successful performing host computer vision task object recognition object detection image segmentation texture synthesis 2015 Gatys et al 7 show style painter extracted image painting applied another normal photo...
[-0.00823851116001606, 0.062297869473695755, 0.004127266816794872, 0.0646948516368866, 0.0053047193214297295, -0.0337434858083725, 0.020034287124872208, 0.00061299919616431, -0.0553683266043663, 0.0040498278103768826, -0.006196638569235802, 0.02744789980351925, 0.00698753260076046, 0.06729361414909363, 0.04759786650538...
706
706
['Seungkyun Hong', 'Seongchan Kim', 'Minsu Joh', 'Sa-kwang Song']
1708.03417v1
Advances in remote sensing technologies have made it possible to use high-resolution visual data for weather observation and forecasting tasks. We propose the use of multi-layer neural networks for understanding complex atmospheric dynamics based on multichannel satellite images. The capability of our model was evaluat...
GlobeNet: Convolutional Neural Networks for Typhoon Eye Tracking from Remote Sensing Imagery
2,017
http://arxiv.org/pdf/1708.03417v1
Title GlobeNet Convolutional Neural Networks Typhoon Eye Tracking Remote Sensing Imagery Summary Advances remote sensing technology made possible use highresolution visual data weather observation forecasting task propose use multilayer neural network understanding complex atmospheric dynamic based multichannel satelli...
[0.011978700757026672, 0.03270982578396797, 0.0042501394636929035, 0.0399932935833931, 0.03708725795149803, 0.030548449605703354, 0.015155835077166557, -0.09495334327220917, -0.023419389501214027, 0.05626843869686127, -0.002228537807241082, -0.05608472228050232, 0.02461966685950756, 0.04351809248328209, 0.0705731287598...
707
707
['Dat Thanh Tran', 'Alexandros Iosifidis', 'Moncef Gabbouj']
1709.09902v3
The excellent performance of deep neural networks has enabled us to solve several automatization problems, opening an era of autonomous devices. However, current deep net architectures are heavy with millions of parameters and require billions of floating point operations. Several works have been developed to compress ...
Improving Efficiency in Convolutional Neural Network with Multilinear Filters
2,017
http://arxiv.org/pdf/1709.09902v3
Title Improving Efficiency Convolutional Neural Network Multilinear Filters Summary excellent performance deep neural network enabled u solve several automatization problem opening era autonomous device However current deep net architecture heavy million parameter require billion floating point operation Several work d...
[-0.0230800062417984, 0.06493087112903595, -0.007753004319965839, 0.052645500749349594, 0.00351553107611835, -0.03316081687808037, 0.06656789034605026, -0.014636198990046978, -0.04129568859934807, -0.00214259116910398, -0.017294492572546005, 0.03075144626200199, 4.964257823303342e-05, 0.03679436445236206, 0.03370646387...
708
708
['Devinder Kumar', 'Graham W. Taylor', 'Alexander Wong']
1710.10675v1
Objective: Radiomics-driven Computer Aided Diagnosis (CAD) has shown considerable promise in recent years as a potential tool for improving clinical decision support in medical oncology, particularly those based around the concept of Discovery Radiomics, where radiomic sequencers are discovered through the analysis of ...
Discovery Radiomics with CLEAR-DR: Interpretable Computer Aided Diagnosis of Diabetic Retinopathy
2,017
http://arxiv.org/pdf/1710.10675v1
Title Discovery Radiomics CLEARDR Interpretable Computer Aided Diagnosis Diabetic Retinopathy Summary Objective Radiomicsdriven Computer Aided Diagnosis CAD shown considerable promise recent year potential tool improving clinical decision support medical oncology particularly based around concept Discovery Radiomics ra...
[0.004513891413807869, -0.01456287782639265, -0.0058048442006111145, -0.010569184087216854, 0.026145469397306442, 0.04538270831108093, 0.014838271774351597, 0.07058724015951157, -0.007606544066220522, 0.05095557123422623, 0.0695001408457756, 0.0015230110147967935, 0.026726379990577698, 0.07343531399965286, -0.019839175...
709
709
['Emad Barsoum', 'John Kender', 'Zicheng Liu']
1711.09561v1
Predicting and understanding human motion dynamics has many applications, such as motion synthesis, augmented reality, security, and autonomous vehicles. Due to the recent success of generative adversarial networks (GAN), there has been much interest in probabilistic estimation and synthetic data generation using deep ...
HP-GAN: Probabilistic 3D human motion prediction via GAN
2,017
http://arxiv.org/pdf/1711.09561v1
Title HPGAN Probabilistic 3D human motion prediction via GAN Summary Predicting understanding human motion dynamic many application motion synthesis augmented reality security autonomous vehicle Due recent success generative adversarial network GAN much interest probabilistic estimation synthetic data generation using ...
[-0.0058366525918245316, 0.06477829813957214, -0.04030698537826538, 0.006125688087195158, 0.0358552560210228, 0.0020677705761045218, -0.002153107663616538, 0.018649939447641373, -0.009829223155975342, 0.004496832378208637, -0.009524628520011902, -0.0560121014714241, 0.005378892179578543, 0.010715723037719727, 0.0897707...
710
710
['Qiangeng Xu', 'John Kender']
1712.09709v2
In the research of the impact of gestures using by a lecturer, one challenging task is to infer the attention of a group of audiences. Two important measurements that can help infer the level of attention are eye movement data and Electroencephalography (EEG) data. Under the fundamental assumption that a group of peopl...
Report: Dynamic Eye Movement Matching and Visualization Tool in Neuro Gesture
2,017
http://arxiv.org/pdf/1712.09709v2
Title Report Dynamic Eye Movement Matching Visualization Tool Neuro Gesture Summary research impact gesture using lecturer one challenging task infer attention group audience Two important measurement help infer level attention eye movement data Electroencephalography EEG data fundamental assumption group people would ...
[0.02106880210340023, 0.0004780478775501251, -0.03516867756843567, 0.017498597502708435, 0.004785249475389719, 0.04110534489154816, 0.011310875415802002, 0.06284011155366898, -0.003938641864806414, -0.024829421192407608, -0.026670614257454872, 0.013832620345056057, 0.061912354081869125, 0.044799890369176865, 0.06801147...
711
711
['Audrey G. Chung', 'Paul Fieguth', 'Alexander Wong']
1802.03318v1
Evolutionary deep intelligence synthesizes highly efficient deep neural networks architectures over successive generations. Inspired by the nature versus nurture debate, we propose a study to examine the role of external factors on the network synthesis process by varying the availability of simulated environmental res...
Nature vs. Nurture: The Role of Environmental Resources in Evolutionary Deep Intelligence
2,018
http://arxiv.org/pdf/1802.03318v1
Title Nature v Nurture Role Environmental Resources Evolutionary Deep Intelligence Summary Evolutionary deep intelligence synthesizes highly efficient deep neural network architecture successive generation Inspired nature versus nurture debate propose study examine role external factor network synthesis process varying...
[-0.024060428142547607, 0.05777814984321594, -0.057195376604795456, 0.0167063120752573, -0.006461993791162968, -0.015112362802028656, 0.032274022698402405, -0.02153223194181919, -0.040376462042331696, 0.020567379891872406, -0.03807233273983002, 0.006133691407740116, -0.021327001973986626, 0.050032876431941986, 0.036894...
712
712
['Akisato kimura', 'Derek Pang', 'Tatsuto Takeuchi', 'Kouji Miyazato', 'Junji Yamato', 'Kunio Kashino']
1004.0085v1
Recent studies in the field of human vision science suggest that the human responses to the stimuli on a visual display are non-deterministic. People may attend to different locations on the same visual input at the same time. Based on this knowledge, we propose a new stochastic model of visual attention by introducing...
A stochastic model of human visual attention with a dynamic Bayesian network
2,010
http://arxiv.org/pdf/1004.0085v1
Title stochastic model human visual attention dynamic Bayesian network Summary Recent study field human vision science suggest human response stimulus visual display nondeterministic People may attend different location visual input time Based knowledge propose new stochastic model visual attention introducing dynamic ...
[-0.010042090900242329, -0.02305353246629238, -0.019794777035713196, -0.0253529641777277, 0.010284334421157837, -0.005860269069671631, 0.004688510671257973, 0.0102235684171319, -0.007809461560100317, -0.035747282207012177, 0.025163181126117706, -0.04932325333356857, 0.047812964767217636, 0.06632932275533676, 0.05536041...
713
713
['Tee Connie', 'Mundher Al-Shabi', 'Michael Goh']
1612.09506v2
With rapid development of the Internet, web contents become huge. Most of the websites are publicly available, and anyone can access the contents from anywhere such as workplace, home and even schools. Nevertheless, not all the web contents are appropriate for all users, especially children. An example of these content...
Smart Content Recognition from Images Using a Mixture of Convolutional Neural Networks
2,016
http://arxiv.org/pdf/1612.09506v2
Title Smart Content Recognition Images Using Mixture Convolutional Neural Networks Summary rapid development Internet web content become huge website publicly available anyone access content anywhere workplace home even school Nevertheless web content appropriate user especially child example content pornography image ...
[0.053268421441316605, 0.04292959347367287, -0.025858022272586823, 0.04832489788532257, -0.04007880762219429, 0.05105172470211983, 0.07456593215465546, 0.04035274311900139, 0.017997123301029205, -0.048988133668899536, -0.028000997379422188, -0.020716283470392227, -0.01718912459909916, 0.04956079646945, -0.0219766721129...
714
714
['Giacomo Cocci', 'Davide Barbieri', 'Giovanna Citti', 'Alessandro Sarti']
1407.0733v2
The visual systems of many mammals, including humans, is able to integrate the geometric information of visual stimuli and to perform cognitive tasks already at the first stages of the cortical processing. This is thought to be the result of a combination of mechanisms, which include feature extraction at single cell l...
Cortical spatio-temporal dimensionality reduction for visual grouping
2,014
http://arxiv.org/pdf/1407.0733v2
Title Cortical spatiotemporal dimensionality reduction visual grouping Summary visual system many mammal including human able integrate geometric information visual stimulus perform cognitive task already first stage cortical processing thought result combination mechanism include feature extraction single cell level g...
[-0.029008204117417336, -0.04715174436569214, -0.0460982471704483, 0.024282919242978096, -0.018221618607640266, 0.03414541482925415, 0.05285128951072693, -0.008611221797764301, 0.0069997296668589115, 0.018568407744169235, -0.02713211253285408, -0.014310279861092567, 0.057770781219005585, 0.057154033333063126, 0.0658645...
715
715
['Can Xu', 'Suleyman Cetintas', 'Kuang-Chih Lee', 'Li-Jia Li']
1411.5731v1
Images have become one of the most popular types of media through which users convey their emotions within online social networks. Although vast amount of research is devoted to sentiment analysis of textual data, there has been very limited work that focuses on analyzing sentiment of image data. In this work, we propo...
Visual Sentiment Prediction with Deep Convolutional Neural Networks
2,014
http://arxiv.org/pdf/1411.5731v1
Title Visual Sentiment Prediction Deep Convolutional Neural Networks Summary Images become one popular type medium user convey emotion within online social network Although vast amount research devoted sentiment analysis textual data limited work focus analyzing sentiment image data work propose novel visual sentiment ...
[0.026351680979132652, 0.06894616782665253, -0.00021954286785330623, 0.04812313988804817, -0.02541949413716793, 0.028199469670653343, -0.009609222412109375, 0.01943696103990078, 0.014297880232334137, -0.02676563896238804, -0.011985491029918194, 0.015088360756635666, -0.03671792522072792, 0.05788110941648483, 0.01268449...
716
716
['Fei Zhu', 'Abderrahim Halimi', 'Paul Honeine', 'Badong Chen', 'Nanning Zheng']
1602.01729v1
In hyperspectral images, some spectral bands suffer from low signal-to-noise ratio due to noisy acquisition and atmospheric effects, thus requiring robust techniques for the unmixing problem. This paper presents a robust supervised spectral unmixing approach for hyperspectral images. The robustness is achieved by writi...
Correntropy Maximization via ADMM - Application to Robust Hyperspectral Unmixing
2,016
http://arxiv.org/pdf/1602.01729v1
Title Correntropy Maximization via ADMM Application Robust Hyperspectral Unmixing Summary hyperspectral image spectral band suffer low signaltonoise ratio due noisy acquisition atmospheric effect thus requiring robust technique unmixing problem paper present robust supervised spectral unmixing approach hyperspectral im...
[-0.01128374319523573, 0.008524861186742783, 0.006219339091330767, 0.035685356706380844, 0.011910916306078434, 0.012478083372116089, -0.0077212220057845116, -0.007883419282734394, -0.004385383799672127, 0.031793300062417984, -0.019009126350283623, 0.05858924612402916, 0.016866693273186684, 0.026558484882116318, -0.0159...
717
717
['Farhad Arbabzadah', 'Grégoire Montavon', 'Klaus-Robert Müller', 'Wojciech Samek']
1606.07285v2
This paper focuses on the problem of explaining predictions of psychological attributes such as attractiveness, happiness, confidence and intelligence from face photographs using deep neural networks. Since psychological attribute datasets typically suffer from small sample sizes, we apply transfer learning with two ba...
Identifying individual facial expressions by deconstructing a neural network
2,016
http://arxiv.org/pdf/1606.07285v2
Title Identifying individual facial expression deconstructing neural network Summary paper focus problem explaining prediction psychological attribute attractiveness happiness confidence intelligence face photograph using deep neural network Since psychological attribute datasets typically suffer small sample size appl...
[0.01171519048511982, 0.10384302586317062, -0.03140038624405861, 0.006184595637023449, -0.0005059328977949917, 0.08105877041816711, 0.04882337898015976, -0.01604016311466694, 0.012622016482055187, 0.036673445254564285, -0.04585159569978714, -0.0045763589441776276, 0.02939782291650772, 0.027108021080493927, 0.0313397124...
718
718
['Jing Yu Koh', 'Wojciech Samek', 'Klaus-Robert Müller', 'Alexander Binder']
1606.09187v3
Semantic boundary and edge detection aims at simultaneously detecting object edge pixels in images and assigning class labels to them. Systematic training of predictors for this task requires the labeling of edges in images which is a particularly tedious task. We propose a novel strategy for solving this task, when pi...
Object Boundary Detection and Classification with Image-level Labels
2,016
http://arxiv.org/pdf/1606.09187v3
Title Object Boundary Detection Classification Imagelevel Labels Summary Semantic boundary edge detection aim simultaneously detecting object edge pixel image assigning class label Systematic training predictor task requires labeling edge image particularly tedious task propose novel strategy solving task pixellevel an...
[0.014435085467994213, -0.009498577564954758, 0.016130737960338593, 0.0587863065302372, -0.027480414137244225, -0.04759363457560539, 0.057943981140851974, -0.020749565213918686, 0.0026499656960368156, -0.03737781569361687, -0.010966208763420582, 0.05082504823803902, -0.005800250452011824, 0.05133979022502899, -0.026573...
719
719
['Sam Kriegman', 'Marcin Szubert', 'Josh C. Bongard', 'Christian Skalka']
1706.07888v2
Satellite imagery and remote sensing provide explanatory variables at relatively high resolutions for modeling geospatial phenomena, yet regional summaries are often desirable for analysis and actionable insight. In this paper, we propose a novel method of inducing spatial aggregations as a component of the machine lea...
Evolving Spatially Aggregated Features from Satellite Imagery for Regional Modeling
2,017
http://arxiv.org/pdf/1706.07888v2
Title Evolving Spatially Aggregated Features Satellite Imagery Regional Modeling Summary Satellite imagery remote sensing provide explanatory variable relatively high resolution modeling geospatial phenomenon yet regional summary often desirable analysis actionable insight paper propose novel method inducing spatial ag...
[-0.015096725896000862, 0.0686153993010521, -0.03705476596951485, -0.02326796017587185, -0.019166309386491776, 0.039359673857688904, 0.018243933096528053, -0.07028821110725403, 0.026447806507349014, 0.025008250027894974, 0.07410223037004471, 0.004483410157263279, 0.04304669424891472, 0.07307964563369751, 0.047447010874...
720
720
['Biswa Sengupta', 'Yu Qian']
1708.06250v1
In recent work, it was shown that combining multi-kernel based support vector machines (SVMs) can lead to near state-of-the-art performance on an action recognition dataset (HMDB-51 dataset). This was 0.4\% lower than frameworks that used hand-crafted features in addition to the deep convolutional feature extractors. I...
Pillar Networks++: Distributed non-parametric deep and wide networks
2,017
http://arxiv.org/pdf/1708.06250v1
Title Pillar Networks Distributed nonparametric deep wide network Summary recent work shown combining multikernel based support vector machine SVMs lead near stateoftheart performance action recognition dataset HMDB51 dataset 04 lower framework used handcrafted feature addition deep convolutional feature extractor pres...
[-0.02186121791601181, 0.019998127594590187, -0.011713789775967598, 0.04303362965583801, 0.006795182358473539, 0.023230986669659615, 0.08459728956222534, -0.02022259682416916, -0.06779662519693375, 0.017613446339964867, -0.04614480212330818, -0.015191325917840004, 0.03995179757475853, 0.07688998430967331, 0.01177845895...
721
721
['Ivo Kwee', 'Marcus Hutter', 'Juergen Schmidhuber']
cs/0105025v1
Unlike traditional reinforcement learning (RL), market-based RL is in principle applicable to worlds described by partially observable Markov Decision Processes (POMDPs), where an agent needs to learn short-term memories of relevant previous events in order to execute optimal actions. Most previous work, however, has f...
Market-Based Reinforcement Learning in Partially Observable Worlds
2,001
http://arxiv.org/pdf/cs/0105025v1
Title MarketBased Reinforcement Learning Partially Observable Worlds Summary Unlike traditional reinforcement learning RL marketbased RL principle applicable world described partially observable Markov Decision Processes POMDPs agent need learn shortterm memory relevant previous event order execute optimal action previ...
[0.014959203079342842, 0.03265559300780296, 0.0059007154777646065, -0.07738395780324936, -0.013619828969240189, -0.009802755899727345, -0.031781792640686035, -0.009771967306733131, -0.024042461067438126, -0.006411755457520485, 0.009179960936307907, 0.054150912910699844, -0.0539899580180645, 0.08083557337522507, -0.0036...
722
722
['Andras Lorincz']
cs/0308025v1
A model of sensory information processing is presented. The model assumes that learning of internal (hidden) generative models, which can predict the future and evaluate the precision of that prediction, is of central importance for information extraction. Furthermore, the model makes a bridge to goal-oriented systems ...
Controlled hierarchical filtering: Model of neocortical sensory processing
2,003
http://arxiv.org/pdf/cs/0308025v1
Title Controlled hierarchical filtering Model neocortical sensory processing Summary model sensory information processing presented model assumes learning internal hidden generative model predict future evaluate precision prediction central importance information extraction Furthermore model make bridge goaloriented sy...
[0.0038238430861383677, -0.016976026818156242, -0.030160972848534584, -0.007917379960417747, 0.0030014747753739357, 0.012925976887345314, -0.014209343120455742, 0.021481959149241447, 0.046543776988983154, -0.013778923079371452, -0.05295243486762047, 0.015117663890123367, -0.01441823597997427, 0.06966102868318558, 0.047...
723
723
['David J. Finton']
cs/0404032v1
For an intelligent agent to be truly autonomous, it must be able to adapt its representation to the requirements of its task as it interacts with the world. Most current approaches to on-line feature extraction are ad hoc; in contrast, this paper presents an algorithm that bases judgments of state compatibility and sta...
When Do Differences Matter? On-Line Feature Extraction Through Cognitive Economy
2,004
http://arxiv.org/pdf/cs/0404032v1
Title Differences Matter OnLine Feature Extraction Cognitive Economy Summary intelligent agent truly autonomous must able adapt representation requirement task interacts world current approach online feature extraction ad hoc contrast paper present algorithm base judgment state compatibility statespace abstraction prin...
[-0.0025872213300317526, 0.04461510851979256, -0.04202467203140259, -0.0055188341066241264, -0.031070062890648842, 0.015614354982972145, 0.06376229971647263, 0.0013927760301157832, 0.05910900607705116, -0.03941641002893448, 0.0278855599462986, 0.018414955586194992, 0.0176840890198946, 0.10981688648462296, 0.00545070739...
724
724
['I. Szita', 'A. Lorincz']
cs/0410004v1
Recurrent neural networks are often used for learning time-series data. Based on a few assumptions we model this learning task as a minimization problem of a nonlinear least-squares cost function. The special structure of the cost function allows us to build a connection to reinforcement learning. We exploit this conne...
Applying Policy Iteration for Training Recurrent Neural Networks
2,004
http://arxiv.org/pdf/cs/0410004v1
Title Applying Policy Iteration Training Recurrent Neural Networks Summary Recurrent neural network often used learning timeseries data Based assumption model learning task minimization problem nonlinear leastsquares cost function special structure cost function allows u build connection reinforcement learning exploit ...
[0.0023810474667698145, 0.009202794171869755, 0.004990331828594208, 0.008443190716207027, -0.00026223910390399396, -0.02158651500940323, -0.021275408565998077, -0.02535850554704666, -0.020347122102975845, 0.012646910734474659, 0.006972047965973616, 0.015723707154393196, 0.014062496833503246, 0.05185173824429512, -0.001...
725
725
['Vitaly Schetinin', 'Joachim Schult']
cs/0504069v1
A new technique is presented developed to learn multi-class concepts from clinical electroencephalograms. A desired concept is represented as a neuronal computational model consisting of the input, hidden, and output neurons. In this model the hidden neurons learn independently to classify the electroencephalogram segm...
A Neural-Network Technique to Learn Concepts from Electroencephalograms
2,005
http://arxiv.org/pdf/cs/0504069v1
Title NeuralNetwork Technique Learn Concepts Electroencephalograms Summary new technique presented developed learn multiclass concept clinical electroencephalogram desired concept represented neuronal computational model consisting input hidden output neuron model hidden neuron learn independently classify electroencep...
[-0.026561444625258446, 0.02696835808455944, -0.013183407485485077, -0.007017889525741339, -0.026515575125813484, 0.021268894895911217, 0.06517685949802399, 0.012261785566806793, -0.013266931287944317, -0.014424309134483337, 0.003456740640103817, -0.02011762000620365, 0.04236244410276413, 0.02565435692667961, 0.0446575...
726
726
['Ridwan Al Iqbal']
1005.5556v2
Standard hybrid learners that use domain knowledge require stronger knowledge that is hard and expensive to acquire. However, weaker domain knowledge can benefit from prior knowledge while being cost effective. Weak knowledge in the form of feature relative importance (FRI) is presented and explained. Feature relative ...
Empirical learning aided by weak domain knowledge in the form of feature importance
2,010
http://arxiv.org/pdf/1005.5556v2
Title Empirical learning aided weak domain knowledge form feature importance Summary Standard hybrid learner use domain knowledge require stronger knowledge hard expensive acquire However weaker domain knowledge benefit prior knowledge cost effective Weak knowledge form feature relative importance FRI presented explain...
[-0.014742983505129814, -0.013229891657829285, -0.019791970029473305, -0.030785612761974335, -0.028199950233101845, -0.02570103108882904, 0.08041764795780182, 0.024752169847488403, 0.032797474414110184, 0.01318944338709116, 0.05035974830389023, -0.00034811056684702635, 0.01610279083251953, 0.08712884038686752, 0.032612...
727
727
['J. J. Grefenstette', 'D. E. Moriarty', 'A. C. Schultz']
1106.0221v1
There are two distinct approaches to solving reinforcement learning problems, namely, searching in value function space and searching in policy space. Temporal difference methods and evolutionary algorithms are well-known examples of these approaches. Kaelbling, Littman and Moore recently provided an informative survey...
Evolutionary Algorithms for Reinforcement Learning
2,011
http://arxiv.org/pdf/1106.0221v1
Title Evolutionary Algorithms Reinforcement Learning Summary two distinct approach solving reinforcement learning problem namely searching value function space searching policy space Temporal difference method evolutionary algorithm wellknown example approach Kaelbling Littman Moore recently provided informative survey...
[0.03982983157038689, 0.02611774392426014, -0.03365381062030792, -0.0257871076464653, -0.042755842208862305, -0.027617596089839935, -0.026217561215162277, 0.020619787275791168, -0.0354771688580513, 0.01724240370094776, 0.04367851838469505, 0.032135769724845886, -0.010958706960082054, 0.03356756642460823, -0.03311513364...
728
728
['Guillaume Desjardins', 'Aaron Courville', 'Yoshua Bengio']
1203.4416v1
The deep Boltzmann machine (DBM) has been an important development in the quest for powerful "deep" probabilistic models. To date, simultaneous or joint training of all layers of the DBM has been largely unsuccessful with existing training methods. We introduce a simple regularization scheme that encourages the weight ...
On Training Deep Boltzmann Machines
2,012
http://arxiv.org/pdf/1203.4416v1
Title Training Deep Boltzmann Machines Summary deep Boltzmann machine DBM important development quest powerful deep probabilistic model date simultaneous joint training layer DBM largely unsuccessful existing training method introduce simple regularization scheme encourages weight vector associated hidden unit similar ...
[-0.013285397551953793, -0.009640747681260109, -0.02858475036919117, 0.04183497279882431, 0.017518483102321625, 0.011517945677042007, 0.07407122850418091, -0.015777211636304855, -0.015969205647706985, 0.02935321442782879, -0.010993031784892082, -0.012566630728542805, -0.0056161535903811455, 0.06052934005856514, 0.03048...
729
729
['Farnood Merrikh-Bayat', 'Saeed Bagheri Shouraki']
1109.4609v1
Fuzzy inference systems always suffer from the lack of efficient structures or platforms for their hardware implementation. In this paper, we tried to overcome this problem by proposing new method for the implementation of those fuzzy inference systems which use fuzzy rule base to make inference. To achieve this goal, ...
Memristive fuzzy edge detector
2,011
http://arxiv.org/pdf/1109.4609v1
Title Memristive fuzzy edge detector Summary Fuzzy inference system always suffer lack efficient structure platform hardware implementation paper tried overcome problem proposing new method implementation fuzzy inference system use fuzzy rule base make inference achieve goal designed multilayer neurofuzzy computing sys...
[-0.054506924003362656, -0.008144539780914783, -0.07892411202192307, 0.0748184472322464, -0.011441108770668507, -0.013800231739878654, 0.02545432560145855, -0.0005565573810599744, 0.028147999197244644, 0.008172736503183842, 0.004141718149185181, 0.01474747620522976, -0.03236301615834236, 0.07134664803743362, 0.01824261...
730
730
['Sebastián Basterrech', 'Gerardo Rubino']
1212.6276v1
In the last decade, a new computational paradigm was introduced in the field of Machine Learning, under the name of Reservoir Computing (RC). RC models are neural networks which a recurrent part (the reservoir) that does not participate in the learning process, and the rest of the system where no recurrence (no neural ...
Echo State Queueing Network: a new reservoir computing learning tool
2,012
http://arxiv.org/pdf/1212.6276v1
Title Echo State Queueing Network new reservoir computing learning tool Summary last decade new computational paradigm introduced field Machine Learning name Reservoir Computing RC RC model neural network recurrent part reservoir participate learning process rest system recurrence neural circuit occurs approach grown r...
[-0.04387815669178963, -0.0168597511947155, -0.046949636191129684, -0.007753667421638966, -0.04462617263197899, -0.01876181922852993, 0.012994497083127499, -0.04813118651509285, -0.04522783309221268, 0.0007096731569617987, 0.00988808274269104, -0.04099180921912193, 0.003933105152100325, 0.05911489576101303, 0.006672496...
731
731
['David Silver', 'Hado van Hasselt', 'Matteo Hessel', 'Tom Schaul', 'Arthur Guez', 'Tim Harley', 'Gabriel Dulac-Arnold', 'David Reichert', 'Neil Rabinowitz', 'Andre Barreto', 'Thomas Degris']
1612.08810v3
One of the key challenges of artificial intelligence is to learn models that are effective in the context of planning. In this document we introduce the predictron architecture. The predictron consists of a fully abstract model, represented by a Markov reward process, that can be rolled forward multiple "imagined" plan...
The Predictron: End-To-End Learning and Planning
2,016
http://arxiv.org/pdf/1612.08810v3
Title Predictron EndToEnd Learning Planning Summary One key challenge artificial intelligence learn model effective context planning document introduce predictron architecture predictron consists fully abstract model represented Markov reward process rolled forward multiple imagined planning step forward pas predictron...
[0.0056665921583771706, -0.006368445698171854, -0.021463168784976006, -0.013271157629787922, -0.01797153428196907, -0.03033430315554142, -0.02080725133419037, -0.024465681985020638, -0.019865550100803375, 0.025212226435542107, -0.001936144195497036, 0.07284921407699585, -0.028282826766371727, 0.07988160848617554, -0.00...
732
732
['Deepak Kumar', 'A G Ramakrishnan']
1407.6315v1
Particle swarm optimization is used in several combinatorial optimization problems. In this work, particle swarms are used to solve quadratic programming problems with quadratic constraints. The approach of particle swarms is an example for interior point methods in optimization as an iterative technique. This approach...
Quadratically constrained quadratic programming for classification using particle swarms and applications
2,014
http://arxiv.org/pdf/1407.6315v1
Title Quadratically constrained quadratic programming classification using particle swarm application Summary Particle swarm optimization used several combinatorial optimization problem work particle swarm used solve quadratic programming problem quadratic constraint approach particle swarm example interior point metho...
[-0.03507625684142113, 0.029305221512913704, -0.02904825657606125, -0.016544800251722336, 0.057384174317121506, 0.010084465146064758, 0.005685581360012293, 0.02721221186220646, 0.016257494688034058, -0.00974691566079855, -0.018319960683584213, 0.04222463443875313, 0.012938722968101501, 0.008114258758723736, -0.02714114...
733
733
['Wojciech Zaremba', 'Ilya Sutskever']
1410.4615v3
Recurrent Neural Networks (RNNs) with Long Short-Term Memory units (LSTM) are widely used because they are expressive and are easy to train. Our interest lies in empirically evaluating the expressiveness and the learnability of LSTMs in the sequence-to-sequence regime by training them to evaluate short computer program...
Learning to Execute
2,014
http://arxiv.org/pdf/1410.4615v3
Title Learning Execute Summary Recurrent Neural Networks RNNs Long ShortTerm Memory unit LSTM widely used expressive easy train interest lie empirically evaluating expressiveness learnability LSTMs sequencetosequence regime training evaluate short computer program domain traditionally seen complex neural network consid...
[0.005658090114593506, 0.008094490505754948, -0.02637835592031479, 0.01725480705499649, -0.021237267181277275, -0.009866801090538502, 0.02591564506292343, -0.02976294979453087, -0.025311168283224106, -0.038631077855825424, 0.006735583301633596, -0.03363580256700516, 0.016377123072743416, 0.032064393162727356, 0.0411731...
734
734
['Minje Kim', 'Paris Smaragdis']
1601.06071v1
Based on the assumption that there exists a neural network that efficiently represents a set of Boolean functions between all binary inputs and outputs, we propose a process for developing and deploying neural networks whose weight parameters, bias terms, input, and intermediate hidden layer output signals, are all bin...
Bitwise Neural Networks
2,016
http://arxiv.org/pdf/1601.06071v1
Title Bitwise Neural Networks Summary Based assumption exists neural network efficiently represents set Boolean function binary input output propose process developing deploying neural network whose weight parameter bias term input intermediate hidden layer output signal binaryvalued require basic bit logic feedforward...
[-0.033424027264118195, 0.03408483415842056, -0.03259963542222977, -0.00326509028673172, -0.0326678566634655, -0.017755355685949326, 0.07308824360370636, -0.0013033131835982203, -0.04563900828361511, -0.01738762855529785, 0.05179448798298836, 0.014192993752658367, -0.015645496547222137, 0.09992236644029617, 0.038498960...
735
735
['Tom Zahavy', 'Nir Ben Zrihem', 'Shie Mannor']
1602.02658v4
In recent years there is a growing interest in using deep representations for reinforcement learning. In this paper, we present a methodology and tools to analyze Deep Q-networks (DQNs) in a non-blind matter. Moreover, we propose a new model, the Semi Aggregated Markov Decision Process (SAMDP), and an algorithm that le...
Graying the black box: Understanding DQNs
2,016
http://arxiv.org/pdf/1602.02658v4
Title Graying black box Understanding DQNs Summary recent year growing interest using deep representation reinforcement learning paper present methodology tool analyze Deep Qnetworks DQNs nonblind matter Moreover propose new model Semi Aggregated Markov Decision Process SAMDP algorithm learns automatically SAMDP model ...
[-0.04897117614746094, 0.006604126654565334, -0.014694291166961193, -0.0004318010760471225, 0.018252242356538773, 0.004054898861795664, 0.014945746399462223, -0.024500226601958275, -0.05785233899950981, 0.018399419263005257, 0.0006828439072705805, 0.02430298924446106, -0.012191477231681347, 0.08019841462373734, 0.00979...
736
736
['Randal S. Olson', 'Nathan Bartley', 'Ryan J. Urbanowicz', 'Jason H. Moore']
1603.06212v1
As the field of data science continues to grow, there will be an ever-increasing demand for tools that make machine learning accessible to non-experts. In this paper, we introduce the concept of tree-based pipeline optimization for automating one of the most tedious parts of machine learning---pipeline design. We imple...
Evaluation of a Tree-based Pipeline Optimization Tool for Automating Data Science
2,016
http://arxiv.org/pdf/1603.06212v1
Title Evaluation Treebased Pipeline Optimization Tool Automating Data Science Summary field data science continues grow everincreasing demand tool make machine learning accessible nonexperts paper introduce concept treebased pipeline optimization automating one tedious part machine learningpipeline design implement ope...
[-0.021703533828258514, 0.04981401935219765, -0.04987595975399017, -0.022133897989988327, -0.020270589739084244, 0.0028025531210005283, 0.02102573961019516, 0.005663945805281401, 0.00994651485234499, -0.029387181624770164, 0.0325920432806015, 0.0328822024166584, 0.014238768257200718, 0.11373019218444824, -0.01991215907...
737
737
['Quan Liu', 'Hui Jiang', 'Andrew Evdokimov', 'Zhen-Hua Ling', 'Xiaodan Zhu', 'Si Wei', 'Yu Hu']
1603.07704v2
In this paper, we propose a new deep learning approach, called neural association model (NAM), for probabilistic reasoning in artificial intelligence. We propose to use neural networks to model association between any two events in a domain. Neural networks take one event as input and compute a conditional probability ...
Probabilistic Reasoning via Deep Learning: Neural Association Models
2,016
http://arxiv.org/pdf/1603.07704v2
Title Probabilistic Reasoning via Deep Learning Neural Association Models Summary paper propose new deep learning approach called neural association model NAM probabilistic reasoning artificial intelligence propose use neural network model association two event domain Neural network take one event input compute conditi...
[0.018449990078806877, 0.017635352909564972, -0.001346656703390181, 0.05619872361421585, -0.046915724873542786, 0.01549064926803112, 0.014106840826570988, -0.01580650359392166, 0.02446252852678299, 0.002544377464801073, 0.02468935213983059, 0.0049185920506715775, 0.004400919657200575, 0.04806472733616829, -0.0129656419...
738
738
['Ishan P. Durugkar', 'Clemens Rosenbaum', 'Stefan Dernbach', 'Sridhar Mahadevan']
1606.04615v1
Deep reinforcement learning has been shown to be a powerful framework for learning policies from complex high-dimensional sensory inputs to actions in complex tasks, such as the Atari domain. In this paper, we explore output representation modeling in the form of temporal abstraction to improve convergence and reliabil...
Deep Reinforcement Learning With Macro-Actions
2,016
http://arxiv.org/pdf/1606.04615v1
Title Deep Reinforcement Learning MacroActions Summary Deep reinforcement learning shown powerful framework learning policy complex highdimensional sensory input action complex task Atari domain paper explore output representation modeling form temporal abstraction improve convergence reliability deep reinforcement lea...
[-0.02120278775691986, 0.005359957925975323, -0.047610633075237274, -0.005514223128557205, -0.008647213689982891, -0.0020072441548109055, 0.01049203984439373, -0.026103924959897995, -0.04826332628726959, 0.010998442769050598, -0.033669933676719666, 0.000630987691693008, -0.009201161563396454, 0.07648267596960068, 0.021...
739
739
['Edward Choi', 'Mohammad Taha Bahadori', 'Joshua A. Kulas', 'Andy Schuetz', 'Walter F. Stewart', 'Jimeng Sun']
1608.05745v4
Accuracy and interpretability are two dominant features of successful predictive models. Typically, a choice must be made in favor of complex black box models such as recurrent neural networks (RNN) for accuracy versus less accurate but more interpretable traditional models such as logistic regression. This tradeoff po...
RETAIN: An Interpretable Predictive Model for Healthcare using Reverse Time Attention Mechanism
2,016
http://arxiv.org/pdf/1608.05745v4
Title RETAIN Interpretable Predictive Model Healthcare using Reverse Time Attention Mechanism Summary Accuracy interpretability two dominant feature successful predictive model Typically choice must made favor complex black box model recurrent neural network RNN accuracy versus le accurate interpretable traditional mod...
[0.013573057949543, 0.05023243650794029, -0.022835073992609978, -0.03899867832660675, -0.005059984978288412, 0.01578308641910553, -0.024581685662269592, 0.020174631848931313, 0.01758205145597458, 0.020548513159155846, 0.03386103734374046, -0.06458643823862076, 0.03096551075577736, 0.07153351604938507, -0.00131754297763...
740
740
['Meng Joo Er', 'Rajasekar Venkatesan', 'Ning Wang']
1608.08898v1
In this paper a high speed neural network classifier based on extreme learning machines for multi-label classification problem is proposed and dis-cussed. Multi-label classification is a superset of traditional binary and multi-class classification problems. The proposed work extends the extreme learning machine techni...
A High Speed Multi-label Classifier based on Extreme Learning Machines
2,016
http://arxiv.org/pdf/1608.08898v1
Title High Speed Multilabel Classifier based Extreme Learning Machines Summary paper high speed neural network classifier based extreme learning machine multilabel classification problem proposed discussed Multilabel classification superset traditional binary multiclass classification problem proposed work extends extr...
[0.011641849763691425, -0.019172636792063713, -0.013563303276896477, -0.0055283233523368835, 0.01633145846426487, 0.04251297563314438, 0.01093616709113121, 0.0065765297040343285, 0.02636253461241722, -0.032061949372291565, -0.018313191831111908, 0.05077899247407913, -0.03119768761098385, 0.03670485317707062, 0.01005120...
741
741
['Meng Joo Er', 'Rajasekar Venkatesan', 'Ning Wang']
1609.00843v1
Classification involves the learning of the mapping function that associates input samples to corresponding target label. There are two major categories of classification problems: Single-label classification and Multi-label classification. Traditional binary and multi-class classifications are sub-categories of single...
An Online Universal Classifier for Binary, Multi-class and Multi-label Classification
2,016
http://arxiv.org/pdf/1609.00843v1
Title Online Universal Classifier Binary Multiclass Multilabel Classification Summary Classification involves learning mapping function associate input sample corresponding target label two major category classification problem Singlelabel classification Multilabel classification Traditional binary multiclass classific...
[0.00635999022051692, 0.026650086045265198, -0.005068789701908827, -0.023504823446273804, 0.010311434976756573, 0.033111415803432465, 0.06899004429578781, 0.015605117194354534, 0.028564417734742165, -0.06869341433048248, 0.0148243959993124, 0.009656363166868687, 0.0006846588221378624, 0.05772661790251732, 0.00129340554...
742
742
['Arif Budiman', 'Mohamad Ivan Fanany', 'Chan Basaruddin']
1610.01922v1
A machine learning method needs to adapt to over time changes in the environment. Such changes are known as concept drift. In this paper, we propose concept drift tackling method as an enhancement of Online Sequential Extreme Learning Machine (OS-ELM) and Constructive Enhancement OS-ELM (CEOS-ELM) by adding adaptive ca...
Adaptive Online Sequential ELM for Concept Drift Tackling
2,016
http://arxiv.org/pdf/1610.01922v1
Title Adaptive Online Sequential ELM Concept Drift Tackling Summary machine learning method need adapt time change environment change known concept drift paper propose concept drift tackling method enhancement Online Sequential Extreme Learning Machine OSELM Constructive Enhancement OSELM CEOSELM adding adaptive capabi...
[-0.018714556470513344, 0.01274622417986393, -0.026175880804657936, -0.039601970463991165, 0.021359985694289207, 0.005881111603230238, -0.005135837011039257, 0.016426902264356613, -0.029276644811034203, -0.01450412068516016, 0.03420611843466759, 0.05599755793809891, 0.005857973359525204, 0.08668451756238937, -0.0149632...
743
743
['Arif Budiman', 'Mohamad Ivan Fanany', 'Chan Basaruddin']
1610.02348v1
In big data era, the data continuously generated and its distribution may keep changes overtime. These challenges in online stream of data are known as concept drift. In this paper, we proposed the Adaptive Convolutional ELM method (ACNNELM) as enhancement of Convolutional Neural Network (CNN) with a hybrid Extreme Lea...
Adaptive Convolutional ELM For Concept Drift Handling in Online Stream Data
2,016
http://arxiv.org/pdf/1610.02348v1
Title Adaptive Convolutional ELM Concept Drift Handling Online Stream Data Summary big data era data continuously generated distribution may keep change overtime challenge online stream data known concept drift paper proposed Adaptive Convolutional ELM method ACNNELM enhancement Convolutional Neural Network CNN hybrid ...
[-0.023038430139422417, 0.016558583825826645, -0.04026084765791893, 0.013589044101536274, -0.015497859567403793, 0.009319085627794266, -0.005719159264117479, -0.011203566566109657, -0.046573903411626816, -0.005502530373632908, 0.06766276806592941, 0.019047144800424576, -0.02096056006848812, 0.0857652947306633, 0.008196...
744
744
['Daniel Hein', 'Alexander Hentschel', 'Thomas Runkler', 'Steffen Udluft']
1610.05984v5
Fuzzy controllers are efficient and interpretable system controllers for continuous state and action spaces. To date, such controllers have been constructed manually or trained automatically either using expert-generated problem-specific cost functions or incorporating detailed knowledge about the optimal control strat...
Particle Swarm Optimization for Generating Interpretable Fuzzy Reinforcement Learning Policies
2,016
http://arxiv.org/pdf/1610.05984v5
Title Particle Swarm Optimization Generating Interpretable Fuzzy Reinforcement Learning Policies Summary Fuzzy controller efficient interpretable system controller continuous state action space date controller constructed manually trained automatically either using expertgenerated problemspecific cost function incorpor...
[-0.023040346801280975, 0.003841106314212084, -0.029586870223283768, -0.017008721828460693, 0.061029527336359024, -0.024846505373716354, 0.006034351885318756, 0.02768389694392681, -0.03447912633419037, -0.001576033653691411, 0.015887467190623283, 0.06297402083873749, 0.001110281446017325, 0.06841270625591278, 0.0019515...
745
745
['Marc Pickett', 'Rami Al-Rfou', 'Louis Shao', 'Chris Tar']
1610.06402v1
The long-term memory of most connectionist systems lies entirely in the weights of the system. Since the number of weights is typically fixed, this bounds the total amount of knowledge that can be learned and stored. Though this is not normally a problem for a neural network designed for a specific task, such a bound i...
A Growing Long-term Episodic & Semantic Memory
2,016
http://arxiv.org/pdf/1610.06402v1
Title Growing Longterm Episodic Semantic Memory Summary longterm memory connectionist system lie entirely weight system Since number weight typically fixed bound total amount knowledge learned stored Though normally problem neural network designed specific task bound undesirable system continually learns open range dom...
[0.017698826268315315, 0.06042559817433357, -0.028343917801976204, 0.01917206309735775, 0.008936609141528606, 0.006127943750470877, 0.003440682077780366, -0.022848647087812424, -0.03195356950163841, -0.009028660133481026, 0.033631667494773865, -0.0035851760767400265, -0.034466084092855453, 0.0609319843351841, -0.008430...
746
746
['Wen-Chieh Fang', 'Yi-ting Chiang']
1611.02512v1
Humans can learn concepts or recognize items from just a handful of examples, while machines require many more samples to perform the same task. In this paper, we build a computational model to investigate the possibility of this kind of rapid learning. The proposed method aims to improve the learning task of input fro...
Cognitive Discriminative Mappings for Rapid Learning
2,016
http://arxiv.org/pdf/1611.02512v1
Title Cognitive Discriminative Mappings Rapid Learning Summary Humans learn concept recognize item handful example machine require many sample perform task paper build computational model investigate possibility kind rapid learning proposed method aim improve learning task input sensory memory leveraging information re...
[-0.029868775978684425, -0.07687529176473618, -0.04442603141069412, 0.012023000977933407, 0.0004023114452138543, 0.03359444439411163, 0.012297326698899269, 0.0009574511204846203, 0.05177688226103783, -0.030614037066698074, 0.005810556001961231, 0.010457903146743774, 0.018516918644309044, 0.001411361270584166, -0.002868...
747
747
['Hao Shen']
1611.05827v3
Training deep neural networks for solving machine learning problems is one great challenge in the field, mainly due to its associated optimisation problem being highly non-convex. Recent developments have suggested that many training algorithms do not suffer from undesired local minima under certain scenario, and conse...
Towards a Mathematical Understanding of the Difficulty in Learning with Feedforward Neural Networks
2,016
http://arxiv.org/pdf/1611.05827v3
Title Towards Mathematical Understanding Difficulty Learning Feedforward Neural Networks Summary Training deep neural network solving machine learning problem one great challenge field mainly due associated optimisation problem highly nonconvex Recent development suggested many training algorithm suffer undesired local...
[0.001356947934255004, 0.0023668515495955944, -0.0050458284094929695, 0.02698616497218609, -0.017796842381358147, 0.006771114654839039, 0.002815478714182973, 0.002154191955924034, -0.032144974917173386, 0.029983893036842346, 0.01813541166484356, -0.012416559271514416, 0.013703123666346073, 0.03948216885328293, 0.041411...
748
748
['Gonzalo Diaz', 'Achille Fokoue', 'Giacomo Nannicini', 'Horst Samulowitz']
1705.08520v1
A major challenge in designing neural network (NN) systems is to determine the best structure and parameters for the network given the data for the machine learning problem at hand. Examples of parameters are the number of layers and nodes, the learning rates, and the dropout rates. Typically, these parameters are chos...
An effective algorithm for hyperparameter optimization of neural networks
2,017
http://arxiv.org/pdf/1705.08520v1
Title effective algorithm hyperparameter optimization neural network Summary major challenge designing neural network NN system determine best structure parameter network given data machine learning problem hand Examples parameter number layer node learning rate dropout rate Typically parameter chosen based heuristic r...
[-0.0006385190645232797, -0.001494747120887041, -0.02385726384818554, -0.02010945975780487, 0.033184126019477844, -0.02971462346613407, 0.015958281233906746, -0.00976563710719347, 0.0197676420211792, 0.017336972057819366, -0.014380546286702156, 0.046589069068431854, 0.002216424560174346, 0.07431766390800476, 0.01886870...
749
749
['Decebal Constantin Mocanu', 'Elena Mocanu', 'Peter Stone', 'Phuong H. Nguyen', 'Madeleine Gibescu', 'Antonio Liotta']
1707.04780v1
Through the success of deep learning, Artificial Neural Networks (ANNs) are among the most used artificial intelligence methods nowadays. ANNs have led to major breakthroughs in various domains, such as particle physics, reinforcement learning, speech recognition, computer vision, and so on. Taking inspiration from the...
Evolutionary Training of Sparse Artificial Neural Networks: A Network Science Perspective
2,017
http://arxiv.org/pdf/1707.04780v1
Title Evolutionary Training Sparse Artificial Neural Networks Network Science Perspective Summary success deep learning Artificial Neural Networks ANNs among used artificial intelligence method nowadays ANNs led major breakthrough various domain particle physic reinforcement learning speech recognition computer vision ...
[-0.027887728065252304, 0.0174647718667984, -0.032964251935482025, 0.02248745784163475, -0.00527630839496851, -0.034026794135570526, 0.012799931690096855, 0.011461532674729824, -0.03405437991023064, 0.028860636055469513, -0.020940473303198814, 0.04646703973412514, -0.020626455545425415, 0.03567114844918251, 0.033183451...
750
750
['Ritambhara Singh', 'Jack Lanchantin', 'Arshdeep Sekhon', 'Yanjun Qi']
1708.00339v3
The past decade has seen a revolution in genomic technologies that enable a flood of genome-wide profiling of chromatin marks. Recent literature tried to understand gene regulation by predicting gene expression from large-scale chromatin measurements. Two fundamental challenges exist for such learning tasks: (1) genome...
Attend and Predict: Understanding Gene Regulation by Selective Attention on Chromatin
2,017
http://arxiv.org/pdf/1708.00339v3
Title Attend Predict Understanding Gene Regulation Selective Attention Chromatin Summary past decade seen revolution genomic technology enable flood genomewide profiling chromatin mark Recent literature tried understand gene regulation predicting gene expression largescale chromatin measurement Two fundamental challeng...
[-0.004478110931813717, -0.001050654798746109, -0.033037297427654266, -0.04848727956414223, 0.021697215735912323, 0.007799375336617231, 0.06353030353784561, 0.007420167792588472, 0.04580581933259964, 0.058938346803188324, -0.02583427168428898, 0.0028135734610259533, -0.005884683690965176, 0.051225434988737106, -0.01807...
751
751
['Eric Martin', 'Chris Cundy']
1709.04057v2
Recurrent neural networks (RNNs) are widely used to model sequential data but their non-linear dependencies between sequence elements prevent parallelizing training over sequence length. We show the training of RNNs with only linear sequential dependencies can be parallelized over the sequence length using the parallel...
Parallelizing Linear Recurrent Neural Nets Over Sequence Length
2,017
http://arxiv.org/pdf/1709.04057v2
Title Parallelizing Linear Recurrent Neural Nets Sequence Length Summary Recurrent neural network RNNs widely used model sequential data nonlinear dependency sequence element prevent parallelizing training sequence length show training RNNs linear sequential dependency parallelized sequence length using parallel scan a...
[0.02666763961315155, 0.032647982239723206, -0.01742280274629593, 0.04196585342288017, -0.020144179463386536, 0.006452351342886686, 0.03652665764093399, -0.009637504816055298, -0.022059965878725052, -0.01910150796175003, 0.024840177968144417, -0.03230572119355202, 0.033829059451818466, 0.011337578296661377, -0.00257403...
752
752
['Filipe Alves Neto Verri', 'Renato Tinós', 'Liang Zhao']
1710.09300v1
Data and knowledge representation are fundamental concepts in machine learning. The quality of the representation impacts the performance of the learning model directly. Feature learning transforms or enhances raw data to structures that are effectively exploited by those models. In recent years, several works have bee...
Feature learning in feature-sample networks using multi-objective optimization
2,017
http://arxiv.org/pdf/1710.09300v1
Title Feature learning featuresample network using multiobjective optimization Summary Data knowledge representation fundamental concept machine learning quality representation impact performance learning model directly Feature learning transforms enhances raw data structure effectively exploited model recent year seve...
[-0.0015126197831705213, 0.02810029499232769, -0.0386347770690918, -0.0009802425047382712, -0.003071610815823078, -0.02640114165842533, 0.022930234670639038, 0.006832682061940432, -0.020855717360973358, -0.02538805827498436, -9.471612429479137e-05, 0.011832165531814098, 0.0013946352992206812, 0.07586498558521271, -0.00...
753
753
['Chelsea Finn', 'Sergey Levine']
1710.11622v3
Learning to learn is a powerful paradigm for enabling models to learn from data more effectively and efficiently. A popular approach to meta-learning is to train a recurrent model to read in a training dataset as input and output the parameters of a learned model, or output predictions for new test inputs. Alternativel...
Meta-Learning and Universality: Deep Representations and Gradient Descent can Approximate any Learning Algorithm
2,017
http://arxiv.org/pdf/1710.11622v3
Title MetaLearning Universality Deep Representations Gradient Descent Approximate Learning Algorithm Summary Learning learn powerful paradigm enabling model learn data effectively efficiently popular approach metalearning train recurrent model read training dataset input output parameter learned model output prediction...
[-0.02336861379444599, 0.014397252351045609, -0.011752142570912838, 0.02348981611430645, 0.025858798995614052, 0.0008374102762900293, 0.006655561272054911, -0.01857181079685688, -0.06292036175727844, -0.00812552496790886, -0.02127445861697197, 0.0339849479496479, -0.01750265806913376, 0.030898740515112877, 0.0466424003...
754
754
['Paulo Rauber', 'Filipe Mutz', 'Juergen Schmidhuber']
1711.06006v1
Goal-conditional policies allow reinforcement learning agents to pursue specific goals during different episodes. In addition to their potential to generalize desired behavior to unseen goals, such policies may also help in defining options for arbitrary subgoals, enabling higher-level planning. While trying to achieve...
Hindsight policy gradients
2,017
http://arxiv.org/pdf/1711.06006v1
Title Hindsight policy gradient Summary Goalconditional policy allow reinforcement learning agent pursue specific goal different episode addition potential generalize desired behavior unseen goal policy may also help defining option arbitrary subgoals enabling higherlevel planning trying achieve specific goal agent may...
[0.03146608918905258, 0.06408703327178955, 0.010849918238818645, -0.061614830046892166, 0.009132280945777893, -0.0132643086835742, -0.0090391980484128, 0.02972756326198578, -0.03456856682896614, 0.016909440979361534, 0.0024515341501682997, 0.02041325531899929, -0.04846867918968201, 0.0543789304792881, 0.005326449871063...
755
755
['Mohammad Javad Shafiee', 'Francis Li', 'Brendan Chwyl', 'Alexander Wong']
1711.07459v1
While deep neural networks have been shown in recent years to outperform other machine learning methods in a wide range of applications, one of the biggest challenges with enabling deep neural networks for widespread deployment on edge devices such as mobile and other consumer devices is high computational and memory r...
SquishedNets: Squishing SqueezeNet further for edge device scenarios via deep evolutionary synthesis
2,017
http://arxiv.org/pdf/1711.07459v1
Title SquishedNets Squishing SqueezeNet edge device scenario via deep evolutionary synthesis Summary deep neural network shown recent year outperform machine learning method wide range application one biggest challenge enabling deep neural network widespread deployment edge device mobile consumer device high computatio...
[-0.02499484270811081, 0.024122700095176697, -0.04396184906363487, 0.05038096755743027, 0.010828804224729538, -0.0407930389046669, 0.08757714927196503, -0.017255116254091263, -0.043878424912691116, 0.005290680099278688, 0.004355383105576038, 0.004438853822648525, -0.018252452835440636, 0.08484749495983124, 0.0234755445...
756
756
['Pierre-Yves Oudeyer']
1712.01626v1
Autonomous lifelong development and learning is a fundamental capability of humans, differentiating them from current deep learning systems. However, other branches of artificial intelligence have designed crucial ingredients towards autonomous learning: curiosity and intrinsic motivation, social learning and natural i...
Autonomous development and learning in artificial intelligence and robotics: Scaling up deep learning to human--like learning
2,017
http://arxiv.org/pdf/1712.01626v1
Title Autonomous development learning artificial intelligence robotics Scaling deep learning humanlike learning Summary Autonomous lifelong development learning fundamental capability human differentiating current deep learning system However branch artificial intelligence designed crucial ingredient towards autonomous...
[0.04060013219714165, 0.03074970282614231, -0.05631304904818535, -0.047551605850458145, 0.00949346274137497, 0.03383752331137657, 0.05041634663939476, -0.04431520029902458, -0.01887732557952404, 0.005613050889223814, 0.006764853838831186, 0.03701455146074295, 0.012814365327358246, 0.05520724132657051, -0.00501590594649...
757
757
['Leonid Peshkin', 'Christian R. Shelton']
cs/0204043v1
Searching the space of policies directly for the optimal policy has been one popular method for solving partially observable reinforcement learning problems. Typically, with each change of the target policy, its value is estimated from the results of following that very policy. This requires a large number of interacti...
Learning from Scarce Experience
2,002
http://arxiv.org/pdf/cs/0204043v1
Title Learning Scarce Experience Summary Searching space policy directly optimal policy one popular method solving partially observable reinforcement learning problem Typically change target policy value estimated result following policy requires large number interaction environment different police considered present ...
[0.03234167769551277, 0.05274500325322151, 0.0061468263156712055, -0.060232844203710556, -0.016274088993668556, -0.03032250888645649, 0.027484552934765816, -0.0017381637590005994, -0.05202985182404518, 0.03140789270401001, -0.005477289669215679, 0.027840979397296906, -0.033138565719127655, 0.04808039218187332, -0.01369...
758
758
['Martin Pelikan', 'Kumara Sastry']
cs/0402032v1
This paper describes how fitness inheritance can be used to estimate fitness for a proportion of newly sampled candidate solutions in the Bayesian optimization algorithm (BOA). The goal of estimating fitness for some candidate solutions is to reduce the number of fitness evaluations for problems where fitness evaluatio...
Fitness inheritance in the Bayesian optimization algorithm
2,004
http://arxiv.org/pdf/cs/0402032v1
Title Fitness inheritance Bayesian optimization algorithm Summary paper describes fitness inheritance used estimate fitness proportion newly sampled candidate solution Bayesian optimization algorithm BOA goal estimating fitness candidate solution reduce number fitness evaluation problem fitness evaluation expensive Bay...
[-0.009404526092112064, 0.02924974448978901, -0.029009466990828514, -0.008554755710065365, 0.0031043144408613443, -0.04595917835831642, 0.0030008931644260883, 0.029038438573479652, -0.02064596489071846, -0.01607423461973667, 0.0007445363444276154, 0.02656448818743229, 0.011375466361641884, 0.008967547677457333, -0.0278...
759
759
['Vitaly Schetinin', 'Joachim Schult']
cs/0504070v1
In this paper we describe a new method combining the polynomial neural network and decision tree techniques in order to derive comprehensible classification rules from clinical electroencephalograms (EEGs) recorded from sleeping newborns. These EEGs are heavily corrupted by cardiac, eye movement, muscle and noise artif...
The Combined Technique for Detection of Artifacts in Clinical Electroencephalograms of Sleeping Newborns
2,005
http://arxiv.org/pdf/cs/0504070v1
Title Combined Technique Detection Artifacts Clinical Electroencephalograms Sleeping Newborns Summary paper describe new method combining polynomial neural network decision tree technique order derive comprehensible classification rule clinical electroencephalogram EEGs recorded sleeping newborn EEGs heavily corrupted ...
[-0.031679943203926086, 0.015686405822634697, -0.003519346471875906, 0.0034423188772052526, -0.005845001898705959, 0.054155103862285614, -0.01054330449551344, 0.019419606775045395, -0.05232229083776474, -0.016257239505648613, 0.0851726233959198, 0.014769934117794037, 0.05361296236515045, 0.017282292246818542, 0.0267698...
760
760
['Greg Hulley', 'Tshilidzi Marwala']
0709.3965v2
The ability of a classifier to take on new information and classes by evolving the classifier without it having to be fully retrained is known as incremental learning. Incremental learning has been successfully applied to many classification problems, where the data is changing and is not all available at once. In this...
Evolving Classifiers: Methods for Incremental Learning
2,007
http://arxiv.org/pdf/0709.3965v2
Title Evolving Classifiers Methods Incremental Learning Summary ability classifier take new information class evolving classifier without fully retrained known incremental learning Incremental learning successfully applied many classification problem data changing available paper comparison Learn one recent incremental...
[0.021815409883856773, 0.10014214366674423, 0.0016821863828226924, 0.010092966258525848, -0.029712988063693047, 0.003038417547941208, -0.020706139504909515, 0.06440241634845734, 0.017087634652853012, -0.03599880263209343, 0.04391868785023689, 0.04782071337103844, -0.016808969900012016, 0.04410529509186745, -0.020251888...
761
761
['Dasika Ratna Deepthi', 'G. R. Aditya Krishna', 'K. Eswaran']
0712.0938v1
This paper proposes an unsupervised learning technique by using Multi-layer Mirroring Neural Network and Forgy's clustering algorithm. Multi-layer Mirroring Neural Network is a neural network that can be trained with generalized data inputs (different categories of image patterns) to perform non-linear dimensionality r...
Automatic Pattern Classification by Unsupervised Learning Using Dimensionality Reduction of Data with Mirroring Neural Networks
2,007
http://arxiv.org/pdf/0712.0938v1
Title Automatic Pattern Classification Unsupervised Learning Using Dimensionality Reduction Data Mirroring Neural Networks Summary paper proposes unsupervised learning technique using Multilayer Mirroring Neural Network Forgys clustering algorithm Multilayer Mirroring Neural Network neural network trained generalized d...
[-0.02508358657360077, 0.015034792013466358, -0.05080763250589371, 0.07208844274282455, -0.007659881375730038, 0.021959535777568817, 0.08361513912677765, 0.01171223632991314, -0.0028858822770416737, 0.006447590421885252, -0.026457296684384346, 0.038594167679548264, 0.007938126102089882, 0.019436946138739586, 0.01855116...
762
762
['Alejandro Chinea Manrique De Lara', 'Juan Manuel Moreno', 'Arostegui Jordi Madrenas', 'Joan Cabestany']
0712.3654v1
In this paper we shall review the common problems associated with Piecewise Linear Separation incremental algorithms. This kind of neural models yield poor performances when dealing with some classification problems, due to the evolving schemes used to construct the resulting networks. So as to avoid this undesirable b...
Improving the Performance of PieceWise Linear Separation Incremental Algorithms for Practical Hardware Implementations
2,007
http://arxiv.org/pdf/0712.3654v1
Title Improving Performance PieceWise Linear Separation Incremental Algorithms Practical Hardware Implementations Summary paper shall review common problem associated Piecewise Linear Separation incremental algorithm kind neural model yield poor performance dealing classification problem due evolving scheme used constr...
[0.0001494176685810089, 0.07189055532217026, -0.022032974287867546, 0.0388922281563282, -0.040404386818408966, -0.03054353967308998, 0.07194789499044418, 0.032455530017614365, -0.017937183380126953, -0.04106805846095085, 0.06806177645921707, 0.013983677141368389, -0.03122500143945217, 0.07794579118490219, -0.0069475700...
763
763
['N. Suguna', 'K. Thanushkodi']
1006.4540v1
Feature selection refers to the problem of selecting relevant features which produce the most predictive outcome. In particular, feature selection task is involved in datasets containing huge number of features. Rough set theory has been one of the most successful methods used for feature selection. However, this metho...
A Novel Rough Set Reduct Algorithm for Medical Domain Based on Bee Colony Optimization
2,010
http://arxiv.org/pdf/1006.4540v1
Title Novel Rough Set Reduct Algorithm Medical Domain Based Bee Colony Optimization Summary Feature selection refers problem selecting relevant feature produce predictive outcome particular feature selection task involved datasets containing huge number feature Rough set theory one successful method used feature select...
[-0.01260025892406702, -0.04312650486826897, -0.026841871440410614, -0.061229243874549866, -0.031140973791480064, 0.008109571412205696, -0.01014422532171011, 0.06688329577445984, 0.028713781386613846, 0.0020852144807577133, 0.07218075543642044, -0.006249288097023964, 0.018859602510929108, 0.031781069934368134, -0.04068...
764
764
['Pekka Malo', 'Pyry Siitari', 'Ankur Sinha']
1012.0841v1
Most of the existing information retrieval systems are based on bag of words model and are not equipped with common world knowledge. Work has been done towards improving the efficiency of such systems by using intelligent algorithms to generate search queries, however, not much research has been done in the direction o...
Automated Query Learning with Wikipedia and Genetic Programming
2,010
http://arxiv.org/pdf/1012.0841v1
Title Automated Query Learning Wikipedia Genetic Programming Summary existing information retrieval system based bag word model equipped common world knowledge Work done towards improving efficiency system using intelligent algorithm generate search query however much research done direction incorporating humanandsocie...
[0.07866102457046509, 0.035998765379190445, -0.037776295095682144, 0.02672942355275154, -0.009266302920877934, 0.009866967797279358, -0.04097622632980347, 0.06424432247877121, -0.012354201637208462, -0.04645919427275658, 0.0025595563929528, 0.03770696371793747, -0.03312350809574127, 0.040822792798280716, 0.003726591821...
765
765
['Weishan Dong', 'Tianshi Chen', 'Peter Tino', 'Xin Yao']
1111.2221v1
Since Estimation of Distribution Algorithms (EDA) were proposed, many attempts have been made to improve EDAs' performance in the context of global optimization. So far, the studies or applications of multivariate probabilistic model based continuous EDAs are still restricted to rather low dimensional problems (smaller...
Scaling Up Estimation of Distribution Algorithms For Continuous Optimization
2,011
http://arxiv.org/pdf/1111.2221v1
Title Scaling Estimation Distribution Algorithms Continuous Optimization Summary Since Estimation Distribution Algorithms EDA proposed many attempt made improve EDAs performance context global optimization far study application multivariate probabilistic model based continuous EDAs still restricted rather low dimension...
[0.002503087744116783, 0.016965016722679138, -0.03229774162173271, -0.0029946565628051758, 0.03892260789871216, -0.020755266770720482, 0.022248180583119392, 0.0043020108714699745, -0.0246600154787302, 0.03191452473402023, -0.0027542049065232277, 0.014080332592129707, 0.015541508793830872, 0.049390487372875214, 0.003658...
766
766
['Martin Pelikan', 'Mark W. Hauschild', 'Pier Luca Lanzi']
1203.5443v2
An automated technique has recently been proposed to transfer learning in the hierarchical Bayesian optimization algorithm (hBOA) based on distance-based statistics. The technique enables practitioners to improve hBOA efficiency by collecting statistics from probabilistic models obtained in previous hBOA runs and using...
Transfer Learning, Soft Distance-Based Bias, and the Hierarchical BOA
2,012
http://arxiv.org/pdf/1203.5443v2
Title Transfer Learning Soft DistanceBased Bias Hierarchical BOA Summary automated technique recently proposed transfer learning hierarchical Bayesian optimization algorithm hBOA based distancebased statistic technique enables practitioner improve hBOA efficiency collecting statistic probabilistic model obtained previo...
[-0.014708032831549644, 0.009262935258448124, 0.012727460823953152, -0.007349533028900623, -0.016176505014300346, -0.023754090070724487, 0.04161599278450012, 0.02191346138715744, -0.0016984455287456512, -0.012812601402401924, -0.01368795521557331, 0.025898709893226624, 0.04316175729036331, 0.06633852422237396, -0.00330...
767
767
['Richard J. Preen', 'Larry Bull']
1204.4200v2
A number of representation schemes have been presented for use within Learning Classifier Systems, ranging from binary encodings to neural networks. This paper presents results from an investigation into using a discrete dynamical system representation within the XCS Learning Classifier System. In particular, asynchron...
Discrete Dynamical Genetic Programming in XCS
2,012
http://arxiv.org/pdf/1204.4200v2
Title Discrete Dynamical Genetic Programming XCS Summary number representation scheme presented use within Learning Classifier Systems ranging binary encoding neural network paper present result investigation using discrete dynamical system representation within XCS Learning Classifier System particular asynchronous ra...
[-0.034023359417915344, 0.030810607597231865, -0.06012888252735138, -0.0337800532579422, -0.040345024317502975, -0.03351465240120888, 0.0017823297530412674, 0.011603031307458878, -0.003092412371188402, 0.0024450630880892277, 0.10172656178474426, 0.020060474053025246, -0.02859484776854515, 0.09362750500440598, -0.029521...
768
768
['Richard J. Preen', 'Larry Bull']
1204.4202v1
A number of representation schemes have been presented for use within Learning Classifier Systems, ranging from binary encodings to Neural Networks, and more recently Dynamical Genetic Programming (DGP). This paper presents results from an investigation into using a fuzzy DGP representation within the XCSF Learning Cla...
Fuzzy Dynamical Genetic Programming in XCSF
2,012
http://arxiv.org/pdf/1204.4202v1
Title Fuzzy Dynamical Genetic Programming XCSF Summary number representation scheme presented use within Learning Classifier Systems ranging binary encoding Neural Networks recently Dynamical Genetic Programming DGP paper present result investigation using fuzzy DGP representation within XCSF Learning Classifier System...
[-0.03398854285478592, 0.04416368901729584, -0.06830047070980072, -0.02871222421526909, -0.008317552506923676, -0.01767830178141594, -0.018961185589432716, 0.02694653533399105, -0.019403336569666862, -0.0214847344905138, 0.07777810096740723, -0.002548235235735774, -0.011316522024571896, 0.07732318341732025, -0.02466365...
769
769
['Jonathan Roberts', 'Ke Chen']
1308.6415v2
Procedural content generation (PCG) has recently become one of the hottest topics in computational intelligence and AI game researches. Among a variety of PCG techniques, search-based approaches overwhelmingly dominate PCG development at present. While SBPCG leads to promising results and successful applications, it po...
Learning-Based Procedural Content Generation
2,013
http://arxiv.org/pdf/1308.6415v2
Title LearningBased Procedural Content Generation Summary Procedural content generation PCG recently become one hottest topic computational intelligence AI game research Among variety PCG technique searchbased approach overwhelmingly dominate PCG development present SBPCG lead promising result successful application po...
[0.006704776082187891, 0.028883595019578934, -0.03291783481836319, 0.02905902825295925, 0.002205322030931711, -0.006693611852824688, -0.01848847232758999, -0.00043214860488660634, -0.01983494684100151, 0.04324449598789215, -0.0035593716893345118, 0.016182001680135727, -0.025613220408558846, 0.05025029927492142, -0.0205...
770
770
['Wojciech Jaśkowski']
1406.1509v3
N-tuple networks have been successfully used as position evaluation functions for board games such as Othello or Connect Four. The effectiveness of such networks depends on their architecture, which is determined by the placement of constituent n-tuples, sequences of board locations, providing input to the network. The...
Systematic N-tuple Networks for Position Evaluation: Exceeding 90% in the Othello League
2,014
http://arxiv.org/pdf/1406.1509v3
Title Systematic Ntuple Networks Position Evaluation Exceeding 90 Othello League Summary Ntuple network successfully used position evaluation function board game Othello Connect Four effectiveness network depends architecture determined placement constituent ntuples sequence board location providing input network popul...
[-0.00758735416457057, -0.0052129290997982025, -0.035774797201156616, 0.021313093602657318, -0.02567540854215622, -0.06571634858846664, 0.05021866038441658, -0.01986871287226677, -0.03476303815841675, 0.0057578133419156075, -0.0013465570518746972, -0.009042971767485142, 0.01685352623462677, 0.05219406262040138, 0.06060...
771
771
['Önder Gürcan', 'Carole Bernon', 'Kemal S. Türker']
1207.3760v1
In this paper, the early design of our self-organized agent-based simulation model for exploration of synaptic connections that faithfully generates what is observed in natural situation is given. While we take inspiration from neuroscience, our intent is not to create a veridical model of processes in neurodevelopment...
Towards a Self-Organized Agent-Based Simulation Model for Exploration of Human Synaptic Connections
2,012
http://arxiv.org/pdf/1207.3760v1
Title Towards SelfOrganized AgentBased Simulation Model Exploration Human Synaptic Connections Summary paper early design selforganized agentbased simulation model exploration synaptic connection faithfully generates observed natural situation given take inspiration neuroscience intent create veridical model process ne...
[0.009691315703094006, -0.0249745212495327, -0.028372108936309814, -0.020776573568582535, -0.012043382041156292, -0.0023995819501578808, 0.03426482900977135, -0.026169540360569954, 0.06224702298641205, -0.021491480991244316, -0.011645463295280933, 0.02812385931611061, -0.00671433238312602, 0.022176727652549744, 0.05902...
772
772
['G. N. Tripathi', 'V. Rihani']
1207.4931v1
The paper presents the electronic design and motion planning of a robot based on decision making regarding its straight motion and precise turn using Artificial Neural Network (ANN). The ANN helps in learning of robot so that it performs motion autonomously. The weights calculated are implemented in microcontroller. Th...
Motion Planning Of an Autonomous Mobile Robot Using Artificial Neural Network
2,012
http://arxiv.org/pdf/1207.4931v1
Title Motion Planning Autonomous Mobile Robot Using Artificial Neural Network Summary paper present electronic design motion planning robot based decision making regarding straight motion precise turn using Artificial Neural Network ANN ANN help learning robot performs motion autonomously weight calculated implemented ...
[0.02890363149344921, -0.04013105854392052, 0.0036974656395614147, -0.025588860735297203, 0.00411352701485157, -0.03195500746369362, 0.010114934295415878, -0.032883137464523315, -0.026571722701191902, 0.003150728764012456, 0.02828584797680378, 0.08020555227994919, -0.015967482700943947, 0.026520198211073875, 0.03149738...
773
773
['Rónán Daly', 'Qiang Shen']
1401.3464v1
Bayesian networks are a useful tool in the representation of uncertain knowledge. This paper proposes a new algorithm called ACO-E, to learn the structure of a Bayesian network. It does this by conducting a search through the space of equivalence classes of Bayesian networks using Ant Colony Optimization (ACO). To this...
Learning Bayesian Network Equivalence Classes with Ant Colony Optimization
2,014
http://arxiv.org/pdf/1401.3464v1
Title Learning Bayesian Network Equivalence Classes Ant Colony Optimization Summary Bayesian network useful tool representation uncertain knowledge paper proposes new algorithm called ACOE learn structure Bayesian network conducting search space equivalence class Bayesian network using Ant Colony Optimization ACO end t...
[-0.021568967029452324, -0.0019393136026337743, -0.020777583122253418, -0.00777273578569293, -0.025497393682599068, -0.01368641946464777, -0.0033678885083645582, 0.03526681289076805, 0.02410024218261242, -0.04581594094634056, 0.0031160584185272455, 0.04981612786650658, -0.012527765706181526, 0.03706486523151398, -0.036...
774
774
['Kenton W. Murray', 'Jayant Krishnamurthy']
1612.00712v1
We present probabilistic neural programs, a framework for program induction that permits flexible specification of both a computational model and inference algorithm while simultaneously enabling the use of deep neural networks. Probabilistic neural programs combine a computation graph for specifying a neural network w...
Probabilistic Neural Programs
2,016
http://arxiv.org/pdf/1612.00712v1
Title Probabilistic Neural Programs Summary present probabilistic neural program framework program induction permit flexible specification computational model inference algorithm simultaneously enabling use deep neural network Probabilistic neural program combine computation graph specifying neural network operator wei...
[-0.02065419964492321, 0.05876593291759491, -0.023386411368846893, 0.005874189548194408, -0.036246638745069504, -0.03405193239450455, -0.015628725290298462, -0.06682796776294708, -0.027826080098748207, 0.016735929995775223, 0.06017959490418434, 0.018140411004424095, 0.03910701349377632, 0.08217211067676544, 0.014991165...
775
775
['András Lőrincz', 'Máté Csákvári', 'Áron Fóthi', 'Zoltán Ádám Milacski', 'András Sárkány', 'Zoltán Tősér']
1612.00745v1
Machine learning is making substantial progress in diverse applications. The success is mostly due to advances in deep learning. However, deep learning can make mistakes and its generalization abilities to new tasks are questionable. We ask when and how one can combine network outputs, when (i) details of the observati...
Cognitive Deep Machine Can Train Itself
2,016
http://arxiv.org/pdf/1612.00745v1
Title Cognitive Deep Machine Train Summary Machine learning making substantial progress diverse application success mostly due advance deep learning However deep learning make mistake generalization ability new task questionable ask one combine network output detail observation evaluated learned deep component ii fact ...
[-0.01585247926414013, 0.006229399237781763, -0.011902273632586002, -0.011448919773101807, 0.01015924196690321, 0.02907133474946022, 0.04583369567990303, -0.030547264963388443, -0.0053973812609910965, -0.013417177833616734, 0.016290709376335144, 0.040741998702287674, -0.01556606125086546, 0.04691661149263382, 0.0208981...
776
776
['Alexander L. Gaunt', 'Marc Brockschmidt', 'Rishabh Singh', 'Nate Kushman', 'Pushmeet Kohli', 'Jonathan Taylor', 'Daniel Tarlow']
1612.00817v1
We study machine learning formulations of inductive program synthesis; that is, given input-output examples, synthesize source code that maps inputs to corresponding outputs. Our key contribution is TerpreT, a domain-specific language for expressing program synthesis problems. A TerpreT model is composed of a specifica...
Summary - TerpreT: A Probabilistic Programming Language for Program Induction
2,016
http://arxiv.org/pdf/1612.00817v1
Title Summary TerpreT Probabilistic Programming Language Program Induction Summary study machine learning formulation inductive program synthesis given inputoutput example synthesize source code map input corresponding output key contribution TerpreT domainspecific language expressing program synthesis problem TerpreT ...
[-0.019659869372844696, -0.004071400035172701, -0.02797825261950493, 0.0072052963078022, -0.022938398644328117, -0.0064613064751029015, -0.029354672878980637, -0.013239962980151176, -0.013619085773825645, -0.025967827066779137, 0.06879919022321701, 0.08410071581602097, 0.014550857245922089, 0.0999143049120903, 0.028349...
777
777
['Pierre Baldi', 'Peter Sadowski', 'Zhiqin Lu']
1612.02734v2
Random backpropagation (RBP) is a variant of the backpropagation algorithm for training neural networks, where the transpose of the forward matrices are replaced by fixed random matrices in the calculation of the weight updates. It is remarkable both because of its effectiveness, in spite of using random matrices to co...
Learning in the Machine: Random Backpropagation and the Deep Learning Channel
2,016
http://arxiv.org/pdf/1612.02734v2
Title Learning Machine Random Backpropagation Deep Learning Channel Summary Random backpropagation RBP variant backpropagation algorithm training neural network transpose forward matrix replaced fixed random matrix calculation weight update remarkable effectiveness spite using random matrix communicate error informatio...
[0.010330100543797016, -0.0180104598402977, -0.01898561604321003, 0.0033646111842244864, -0.006797737441956997, -0.028400084003806114, 0.017507245764136314, -0.012958348728716373, -0.032400455325841904, -0.015578096732497215, 0.00574925122782588, 0.010047715157270432, 0.03639937564730644, 0.04428138956427574, 0.0104406...
778
778
['Klaus Greff', 'Rupesh K. Srivastava', 'Jürgen Schmidhuber']
1612.07771v3
The past year saw the introduction of new architectures such as Highway networks and Residual networks which, for the first time, enabled the training of feedforward networks with dozens to hundreds of layers using simple gradient descent. While depth of representation has been posited as a primary reason for their suc...
Highway and Residual Networks learn Unrolled Iterative Estimation
2,016
http://arxiv.org/pdf/1612.07771v3
Title Highway Residual Networks learn Unrolled Iterative Estimation Summary past year saw introduction new architecture Highway network Residual network first time enabled training feedforward network dozen hundred layer using simple gradient descent depth representation posited primary reason success indication archit...
[-0.025647025555372238, 0.08940903842449188, -0.008388910442590714, 0.05896079167723656, -0.006155331153422594, 0.015333243645727634, 0.01831926964223385, -0.005297107622027397, -0.03869413956999779, 0.006378860678523779, 0.011321782134473324, 0.01754014752805233, 0.0028352364897727966, -0.0004428575630299747, -0.00309...
779
779
['Tamas Madl']
1612.09205v1
Despite of the pain and limited accuracy of blood tests for early recognition of cardiovascular disease, they dominate risk screening and triage. On the other hand, heart rate variability is non-invasive and cheap, but not considered accurate enough for clinical practice. Here, we tackle heart beat interval based class...
Deep neural heart rate variability analysis
2,016
http://arxiv.org/pdf/1612.09205v1
Title Deep neural heart rate variability analysis Summary Despite pain limited accuracy blood test early recognition cardiovascular disease dominate risk screening triage hand heart rate variability noninvasive cheap considered accurate enough clinical practice tackle heart beat interval based classification deep learn...
[-0.0301542766392231, -0.0011170697398483753, -0.015254385769367218, -0.007004126440733671, 0.02874237298965454, -0.004598874598741531, 0.032159462571144104, 0.0011184316826984286, -0.008572089485824108, 0.05171731859445572, 0.04561648145318031, -0.019412219524383545, 0.03170312941074371, 0.07260346412658691, -0.014522...
780
780
['Jianlin Cheng']
0704.1028v1
Ordinal regression is an important type of learning, which has properties of both classification and regression. Here we describe a simple and effective approach to adapt a traditional neural network to learn ordinal categories. Our approach is a generalization of the perceptron method for ordinal regression. On severa...
A neural network approach to ordinal regression
2,007
http://arxiv.org/pdf/0704.1028v1
Title neural network approach ordinal regression Summary Ordinal regression important type learning property classification regression describe simple effective approach adapt traditional neural network learn ordinal category approach generalization perceptron method ordinal regression several benchmark datasets method...
[1.9890934709110297e-05, 0.08679355680942535, -0.03430856764316559, -0.006286998745054007, 0.024378515779972076, -0.0016859039897099137, 0.03369252011179924, 0.021915830671787262, -0.0006870360812172294, -0.01151077076792717, 0.046184342354536057, 0.02408537268638611, 0.009612297639250755, 0.04741339012980461, 0.010345...
781
781
['Keyvan Yahya', 'Pouyan Rafiei Fard']
1007.0546v4
Although the Music Sight Reading process has been studied from the cognitive psychology view points, but the computational learning methods like the Reinforcement Learning have not yet been used to modeling of such processes. In this paper, with regards to essential properties of our specific problem, we consider the v...
Computational Model of Music Sight Reading: A Reinforcement Learning Approach
2,010
http://arxiv.org/pdf/1007.0546v4
Title Computational Model Music Sight Reading Reinforcement Learning Approach Summary Although Music Sight Reading process studied cognitive psychology view point computational learning method like Reinforcement Learning yet used modeling process paper regard essential property specific problem consider value function ...
[-0.015092303976416588, -0.036629535257816315, -0.02216220460832119, 0.012726135551929474, -0.027104927226901054, -0.028467558324337006, 0.0231417715549469, 0.02002018131315708, -0.007233802694827318, 0.039952173829078674, -0.04206623509526253, 0.05796084553003311, 0.015421158634126186, 0.03981016203761101, -0.01185084...
782
782
['Habib Shah', 'Rozaida Ghazali', 'Nazri Mohd Nawi']
1112.4628v1
Nowadays, computer scientists have shown the interest in the study of social insect's behaviour in neural networks area for solving different combinatorial and statistical problems. Chief among these is the Artificial Bee Colony (ABC) algorithm. This paper investigates the use of ABC algorithm that simulates the intell...
Using Artificial Bee Colony Algorithm for MLP Training on Earthquake Time Series Data Prediction
2,011
http://arxiv.org/pdf/1112.4628v1
Title Using Artificial Bee Colony Algorithm MLP Training Earthquake Time Series Data Prediction Summary Nowadays computer scientist shown interest study social insect behaviour neural network area solving different combinatorial statistical problem Chief among Artificial Bee Colony ABC algorithm paper investigates use ...
[-0.022705523297190666, 0.0008704561623744667, -0.029974015429615974, -0.005475393030792475, -0.026523591950535774, -0.024368036538362503, -0.022687533870339394, -0.02391619049012661, 0.04949353635311127, 0.0014225913910195231, 0.03828252851963043, 0.03995503857731819, -0.01710430718958378, -0.003329209517687559, 0.027...
783
783
['Guanjiao Ren', 'Weihai Chen', 'Sakyasingha Dasgupta', 'Christoph Kolodziejski', 'Florentin Wörgötter', 'Poramate Manoonpong']
1407.3269v1
An originally chaotic system can be controlled into various periodic dynamics. When it is implemented into a legged robot's locomotion control as a central pattern generator (CPG), sophisticated gait patterns arise so that the robot can perform various walking behaviors. However, such a single chaotic CPG controller ha...
Multiple chaotic central pattern generators with learning for legged locomotion and malfunction compensation
2,014
http://arxiv.org/pdf/1407.3269v1
Title Multiple chaotic central pattern generator learning legged locomotion malfunction compensation Summary originally chaotic system controlled various periodic dynamic implemented legged robot locomotion control central pattern generator CPG sophisticated gait pattern arise robot perform various walking behavior How...
[0.011876724660396576, -0.0229293555021286, -0.029470782727003098, -0.01031036488711834, 0.024149449542164803, 0.01428877841681242, -0.00956327747553587, -0.004203005228191614, -0.06127917021512985, -0.006363043561577797, -0.040662676095962524, -0.04354863986372948, -0.028669357299804688, -0.008590728044509888, 0.04298...
784
784
['Christopher Clark', 'Amos Storkey']
1412.3409v2
Mastering the game of Go has remained a long standing challenge to the field of AI. Modern computer Go systems rely on processing millions of possible future positions to play well, but intuitively a stronger and more 'humanlike' way to play the game would be to rely on pattern recognition abilities rather then brute f...
Teaching Deep Convolutional Neural Networks to Play Go
2,014
http://arxiv.org/pdf/1412.3409v2
Title Teaching Deep Convolutional Neural Networks Play Go Summary Mastering game Go remained long standing challenge field AI Modern computer Go system rely processing million possible future position play well intuitively stronger humanlike way play game would rely pattern recognition ability rather brute force comput...
[0.006599697284400463, 0.04295464977622032, -0.040187060832977295, 0.0573764406144619, 0.023335492238402367, -0.04744700714945793, 0.028163326904177666, -0.0036823363043367863, -0.064762644469738, -0.00047546590212732553, -0.02353096939623356, 0.016629284247756004, -0.0037019795272499323, 0.0871199443936348, 0.05315583...
785
785
['Kratarth Goel', 'Raunaq Vohra', 'J. K. Sahoo']
1412.7927v1
In this paper, we propose a generic technique to model temporal dependencies and sequences using a combination of a recurrent neural network and a Deep Belief Network. Our technique, RNN-DBN, is an amalgamation of the memory state of the RNN that allows it to provide temporal information and a multi-layer DBN that help...
Polyphonic Music Generation by Modeling Temporal Dependencies Using a RNN-DBN
2,014
http://arxiv.org/pdf/1412.7927v1
Title Polyphonic Music Generation Modeling Temporal Dependencies Using RNNDBN Summary paper propose generic technique model temporal dependency sequence using combination recurrent neural network Deep Belief Network technique RNNDBN amalgamation memory state RNN allows provide temporal information multilayer DBN help h...
[-0.012904885224997997, 0.0404168963432312, -0.0016407378716394305, 0.03014226071536541, -0.04600048065185547, -0.043499480932950974, 0.03317447751760483, -0.027902347967028618, -0.029836535453796387, 0.01135867927223444, 0.012194815091788769, -0.028288602828979492, 0.02435942180454731, 0.03377550467848778, -0.00176268...
786
786
['Arun Nair', 'Praveen Srinivasan', 'Sam Blackwell', 'Cagdas Alcicek', 'Rory Fearon', 'Alessandro De Maria', 'Vedavyas Panneershelvam', 'Mustafa Suleyman', 'Charles Beattie', 'Stig Petersen', 'Shane Legg', 'Volodymyr Mnih', 'Koray Kavukcuoglu', 'David Silver']
1507.04296v2
We present the first massively distributed architecture for deep reinforcement learning. This architecture uses four main components: parallel actors that generate new behaviour; parallel learners that are trained from stored experience; a distributed neural network to represent the value function or behaviour policy; ...
Massively Parallel Methods for Deep Reinforcement Learning
2,015
http://arxiv.org/pdf/1507.04296v2
Title Massively Parallel Methods Deep Reinforcement Learning Summary present first massively distributed architecture deep reinforcement learning architecture us four main component parallel actor generate new behaviour parallel learner trained stored experience distributed neural network represent value function behav...
[0.01950511336326599, 0.04397379606962204, -0.035955168306827545, 0.02042825147509575, 0.0004354389675427228, 0.025455953553318977, 0.01727382279932499, -0.02691436931490898, -0.06118806079030037, 0.022097965702414513, -0.06944891065359116, 0.0007897754549048841, -0.034655291587114334, 0.017840558663010597, -0.01142743...
787
787
['Maxim Borisyak', 'Andrey Ustyuzhanin']
1507.07374v1
The problem of autonomous navigation is one of the basic problems for robotics. Although, in general, it may be challenging when an autonomous vehicle is placed into partially observable domain. In this paper we consider simplistic environment model and introduce a navigation algorithm based on Learning Classifier Syst...
A genetic algorithm for autonomous navigation in partially observable domain
2,015
http://arxiv.org/pdf/1507.07374v1
Title genetic algorithm autonomous navigation partially observable domain Summary problem autonomous navigation one basic problem robotics Although general may challenging autonomous vehicle placed partially observable domain paper consider simplistic environment model introduce navigation algorithm based Learning Clas...
[0.04368383437395096, 0.008930462412536144, -0.019828274846076965, -0.0644092857837677, -0.02506658434867859, 0.009054241701960564, -0.013630657456815243, 0.013399327173829079, 0.042633675038814545, -0.007903601974248886, 0.06910116970539093, 0.034307751804590225, -0.014951305463910103, 0.02001459151506424, -0.01071042...
788
788
['Hao Yi Ong', 'Kevin Chavez', 'Augustus Hong']
1508.04186v2
We propose a distributed deep learning model to successfully learn control policies directly from high-dimensional sensory input using reinforcement learning. The model is based on the deep Q-network, a convolutional neural network trained with a variant of Q-learning. Its input is raw pixels and its output is a value ...
Distributed Deep Q-Learning
2,015
http://arxiv.org/pdf/1508.04186v2
Title Distributed Deep QLearning Summary propose distributed deep learning model successfully learn control policy directly highdimensional sensory input using reinforcement learning model based deep Qnetwork convolutional neural network trained variant Qlearning input raw pixel output value function estimating future ...
[-0.016913609579205513, 0.04123321920633316, -0.0237545408308506, 0.03703076392412186, -0.009450439363718033, -0.019529834389686584, 0.03755539655685425, -0.007258306257426739, -0.03266405686736107, 0.030578289180994034, -0.04520018398761749, 0.02052122727036476, -0.03057701885700226, 0.08222495764493942, 0.02227124571...
789
789
['Gustav Sourek', 'Vojtech Aschenbrenner', 'Filip Zelezny', 'Ondrej Kuzelka']
1508.05128v2
We propose a method combining relational-logic representations with neural network learning. A general lifted architecture, possibly reflecting some background domain knowledge, is described through relational rules which may be handcrafted or learned. The relational rule-set serves as a template for unfolding possibly...
Lifted Relational Neural Networks
2,015
http://arxiv.org/pdf/1508.05128v2
Title Lifted Relational Neural Networks Summary propose method combining relationallogic representation neural network learning general lifted architecture possibly reflecting background domain knowledge described relational rule may handcrafted learned relational ruleset serf template unfolding possibly deep neural ne...
[0.01719510741531849, 0.06408540159463882, -0.04534464329481125, 0.05606064200401306, -0.03824929520487785, -0.011926489882171154, -0.0183150302618742, 0.03267604112625122, -0.009359707124531269, -0.02553440071642399, -0.007110713981091976, 0.004581031855195761, -0.008772253058850765, 0.03764282166957855, -0.0138381076...
790
790
['Matthew Lai']
1509.01549v2
This report presents Giraffe, a chess engine that uses self-play to discover all its domain-specific knowledge, with minimal hand-crafted knowledge given by the programmer. Unlike previous attempts using machine learning only to perform parameter-tuning on hand-crafted evaluation functions, Giraffe's learning system al...
Giraffe: Using Deep Reinforcement Learning to Play Chess
2,015
http://arxiv.org/pdf/1509.01549v2
Title Giraffe Using Deep Reinforcement Learning Play Chess Summary report present Giraffe chess engine us selfplay discover domainspecific knowledge minimal handcrafted knowledge given programmer Unlike previous attempt using machine learning perform parametertuning handcrafted evaluation function Giraffes learning sys...
[0.005608752369880676, 0.017788715660572052, -0.03463148698210716, 0.036784302443265915, -0.0033812965266406536, -0.019422052428126335, -0.005180541425943375, 0.014505052007734776, -0.04271160066127777, -0.012004423886537552, -0.0019216561922803521, 0.0015307688154280186, -0.014542761258780956, 0.07205724716186523, 0.0...
791
791
['Kaisheng Yao', 'Geoffrey Zweig', 'Baolin Peng']
1510.08565v3
In a conversation or a dialogue process, attention and intention play intrinsic roles. This paper proposes a neural network based approach that models the attention and intention processes. It essentially consists of three recurrent networks. The encoder network is a word-level model representing source side sentences....
Attention with Intention for a Neural Network Conversation Model
2,015
http://arxiv.org/pdf/1510.08565v3
Title Attention Intention Neural Network Conversation Model Summary conversation dialogue process attention intention play intrinsic role paper proposes neural network based approach model attention intention process essentially consists three recurrent network encoder network wordlevel model representing source side s...
[0.07418721914291382, 0.028101658448576927, -0.016625864431262016, 0.03261515498161316, -0.005677464418113232, -0.022973161190748215, -0.031870037317276, -0.021354587748646736, -0.006044243928045034, -0.048517849296331406, -0.03193512558937073, 0.004044013097882271, 0.012984806671738625, 0.08076771348714828, 0.00123000...
792
792
['Matthew Hausknecht', 'Peter Stone']
1511.04143v4
Recent work has shown that deep neural networks are capable of approximating both value functions and policies in reinforcement learning domains featuring continuous state and action spaces. However, to the best of our knowledge no previous work has succeeded at using deep neural networks in structured (parameterized) ...
Deep Reinforcement Learning in Parameterized Action Space
2,015
http://arxiv.org/pdf/1511.04143v4
Title Deep Reinforcement Learning Parameterized Action Space Summary Recent work shown deep neural network capable approximating value function policy reinforcement learning domain featuring continuous state action space However best knowledge previous work succeeded using deep neural network structured parameterized c...
[-0.012675650417804718, 0.007990937680006027, -0.026146497577428818, -0.019109152257442474, 0.046831220388412476, -0.029667936265468597, 0.0076109240762889385, -0.06154318153858185, -0.06295673549175262, 0.019057054072618484, -0.030551210045814514, 0.007014790084213018, -0.048812828958034515, 0.08638209104537964, 0.028...
793
793
['Sainbayar Sukhbaatar', 'Arthur Szlam', 'Gabriel Synnaeve', 'Soumith Chintala', 'Rob Fergus']
1511.07401v2
This paper introduces MazeBase: an environment for simple 2D games, designed as a sandbox for machine learning approaches to reasoning and planning. Within it, we create 10 simple games embodying a range of algorithmic tasks (e.g. if-then statements or set negation). A variety of neural models (fully connected, convolu...
MazeBase: A Sandbox for Learning from Games
2,015
http://arxiv.org/pdf/1511.07401v2
Title MazeBase Sandbox Learning Games Summary paper introduces MazeBase environment simple 2D game designed sandbox machine learning approach reasoning planning Within create 10 simple game embodying range algorithmic task eg ifthen statement set negation variety neural model fully connected convolutional network memor...
[0.01864539086818695, -0.02072683721780777, -0.03355177491903305, -0.014497590251266956, -0.017775719985365868, -0.008028334937989712, -0.004133883398026228, -0.010758710093796253, 0.022005965933203697, 0.02145436592400074, -0.01544943731278181, 0.06511269509792328, -0.057799044996500015, 0.05978481471538544, 0.0185393...
794
794
['Juergen Schmidhuber']
1511.09249v1
This paper addresses the general problem of reinforcement learning (RL) in partially observable environments. In 2013, our large RL recurrent neural networks (RNNs) learned from scratch to drive simulated cars from high-dimensional video input. However, real brains are more powerful in many ways. In particular, they le...
On Learning to Think: Algorithmic Information Theory for Novel Combinations of Reinforcement Learning Controllers and Recurrent Neural World Models
2,015
http://arxiv.org/pdf/1511.09249v1
Title Learning Think Algorithmic Information Theory Novel Combinations Reinforcement Learning Controllers Recurrent Neural World Models Summary paper address general problem reinforcement learning RL partially observable environment 2013 large RL recurrent neural network RNNs learned scratch drive simulated car highdim...
[0.040364764630794525, 0.039489295333623886, -0.017605502158403397, 0.01730210706591606, -0.053249262273311615, -0.0025828455109149218, -0.0014882119139656425, -0.010194088332355022, -0.05190899223089218, -0.013389719650149345, -0.0011533999349921942, 0.008059129118919373, -0.007386452052742243, 0.0474996343255043, 0.0...
795
795
['Denis Steckelmacher', 'Peter Vrancx']
1512.05509v1
This paper explores the performance of fitted neural Q iteration for reinforcement learning in several partially observable environments, using three recurrent neural network architectures: Long Short-Term Memory, Gated Recurrent Unit and MUT1, a recurrent neural architecture evolved from a pool of several thousands ca...
An Empirical Comparison of Neural Architectures for Reinforcement Learning in Partially Observable Environments
2,015
http://arxiv.org/pdf/1512.05509v1
Title Empirical Comparison Neural Architectures Reinforcement Learning Partially Observable Environments Summary paper explores performance fitted neural Q iteration reinforcement learning several partially observable environment using three recurrent neural network architecture Long ShortTerm Memory Gated Recurrent Un...
[0.031508006155490875, 0.000662361504510045, -0.008813418447971344, -0.013668354600667953, -0.0028510328847914934, -0.019623108208179474, 0.0009291676105931401, -0.005711337551474571, -0.025212733075022697, 0.019281769171357155, -0.040371086448431015, -0.012081305496394634, -0.04073917865753174, 0.053953949362039566, 0...
796
796
['Cristóbal Esteban', 'Oliver Staeck', 'Yinchong Yang', 'Volker Tresp']
1602.02685v2
In clinical data sets we often find static information (e.g. patient gender, blood type, etc.) combined with sequences of data that are recorded during multiple hospital visits (e.g. medications prescribed, tests performed, etc.). Recurrent Neural Networks (RNNs) have proven to be very successful for modelling sequence...
Predicting Clinical Events by Combining Static and Dynamic Information Using Recurrent Neural Networks
2,016
http://arxiv.org/pdf/1602.02685v2
Title Predicting Clinical Events Combining Static Dynamic Information Using Recurrent Neural Networks Summary clinical data set often find static information eg patient gender blood type etc combined sequence data recorded multiple hospital visit eg medication prescribed test performed etc Recurrent Neural Networks RNN...
[0.013261804357171059, 0.04503307119011879, -0.008894164115190506, -0.035063933581113815, -0.015218491666018963, 0.01147483941167593, -0.010656564496457577, 0.016265563666820526, -0.028623173013329506, 0.0058285086415708065, 0.04173915088176727, -0.06631138175725937, 0.039380863308906555, 0.0908694937825203, 0.00199571...
797
797
['Tim Salimans', 'Diederik P. Kingma']
1602.07868v3
We present weight normalization: a reparameterization of the weight vectors in a neural network that decouples the length of those weight vectors from their direction. By reparameterizing the weights in this way we improve the conditioning of the optimization problem and we speed up convergence of stochastic gradient d...
Weight Normalization: A Simple Reparameterization to Accelerate Training of Deep Neural Networks
2,016
http://arxiv.org/pdf/1602.07868v3
Title Weight Normalization Simple Reparameterization Accelerate Training Deep Neural Networks Summary present weight normalization reparameterization weight vector neural network decouples length weight vector direction reparameterizing weight way improve conditioning optimization problem speed convergence stochastic g...
[-0.020008239895105362, 0.054349448531866074, -0.012255227193236351, 0.04392442852258682, 0.02404322847723961, -0.004250003956258297, 0.035763900727033615, -0.007881603203713894, -0.043586261570453644, 0.043948639184236526, -0.025443583726882935, 0.025280063971877098, 0.0058187455870211124, 0.03038506582379341, -0.0004...
798
798
['Felix Leibfried', 'Daniel Alexander Braun']
1602.08332v2
Bounded rational decision-makers transform sensory input into motor output under limited computational resources. Mathematically, such decision-makers can be modeled as information-theoretic channels with limited transmission rate. Here, we apply this formalism for the first time to multilayer feedforward neural networ...
Bounded Rational Decision-Making in Feedforward Neural Networks
2,016
http://arxiv.org/pdf/1602.08332v2
Title Bounded Rational DecisionMaking Feedforward Neural Networks Summary Bounded rational decisionmakers transform sensory input motor output limited computational resource Mathematically decisionmakers modeled informationtheoretic channel limited transmission rate apply formalism first time multilayer feedforward neu...
[-0.014995863661170006, 0.05422063544392586, -0.01628904789686203, -0.010860799811780453, 0.011774145998060703, -0.008793439716100693, 0.008602979592978954, 0.04339129850268364, 5.646177669405006e-05, 0.008354364894330502, -0.03066168911755085, 0.03283068910241127, 0.016480116173624992, 0.06421598047018051, 0.051957350...
799
799
['Greg Yang']
1602.08671v3
Following the recent trend in explicit neural memory structures, we present a new design of an external memory, wherein memories are stored in an Euclidean key space $\mathbb R^n$. An LSTM controller performs read and write via specialized read and write heads. It can move a head by either providing a new address in th...
Lie Access Neural Turing Machine
2,016
http://arxiv.org/pdf/1602.08671v3
Title Lie Access Neural Turing Machine Summary Following recent trend explicit neural memory structure present new design external memory wherein memory stored Euclidean key space mathbb Rn LSTM controller performs read write via specialized read write head move head either providing new address key space aka random ac...
[0.023067500442266464, -0.03536568954586983, -0.0033056761603802443, 0.035869795829057693, -0.009078841656446457, -0.01579129882156849, 0.026542559266090393, -0.016675584018230438, 0.023853767663240433, -0.026093723252415657, -0.04919576272368431, -0.05345131456851959, -0.006935525219887495, 0.04885188490152359, 0.0583...