text
stringlengths
27
153
label
stringlengths
27
153
id
int64
0
40k
each the only two. and In one contrast, concern vehicle segmentation of detection semantic
In contrast, vehicle detection and semantic segmentation each only concern one of the two.
35,700
propose tackle this to boundary-aware a We with multi-task semantic problem network. learning
We propose to tackle this problem with a semantic boundary-aware multi-task learning network.
35,701
reconstructed can material fiber The model. reinforced using a structure be composite of a tracking-by-detection
The structure of a fiber reinforced composite material can be reconstructed using a tracking-by-detection model.
35,702
accuracy, the heavily performance. on detection rely recall Tracking-by-detection algorithms especially
Tracking-by-detection algorithms rely heavily on detection accuracy, especially the recall performance.
35,703
significant tracking advanced over state-of-the-art with algorithms fiber show together performance. Experiments improvements detection
Experiments show significant improvements over state-of-the-art fiber detection algorithms together with advanced tracking performance.
35,704
powerful are feedforward for tool functional (DFNNs) a neural approximation. Deep networks
Deep feedforward neural networks (DFNNs) are a powerful tool for functional approximation.
35,705
learning our also in method. We perform to how deep variable describe selection
We also describe how to perform variable selection in our deep learning method.
35,706
experiments our that SISR method validate methods. outperforms state-of-the-art Extensive the existing
Extensive experiments validate that our SISR method outperforms the existing state-of-the-art methods.
35,707
model perspective assume focal a distortion. unknown We and radial unknown length with camera
We assume a perspective camera model with unknown focal length and unknown radial distortion.
35,708
one modelled is distortion The using the parameter. radial model division with
The radial distortion is modelled using the division model with one parameter.
35,709
This a results solver. numerically in fast stable very and
This results in a very fast and numerically stable solver.
35,710
achieve with dynamic routing fewer generalization capsules model more helps parameters. many The capacity
The dynamic routing helps capsules achieve more generalization capacity with many fewer model parameters.
35,711
with methods multiple parallel benchmarks. The achieve other leading a models in proposed performance
The proposed models achieve a parallel performance with other leading methods in multiple benchmarks.
35,712
neural to stimuli. Artificial behavior in activation internal also have response networks
Artificial neural networks also have internal activation behavior in response to stimuli.
35,713
identifies Further, both tasks. high with performance on networks HMS
Further, HMS identifies networks with high performance on both tasks.
35,714
transmit to The explanations knowledge humans. main of to objective is
The main objective of explanations is to transmit knowledge to humans.
35,715
construct for proposes This machine explanations made models. informative predictions learning to work from
This work proposes to construct informative explanations for predictions made from machine learning models.
35,716
human ranked These to understanding. that explanations informative improves filtered produce and concepts are
These concepts are filtered and ranked to produce informative explanations that improves human understanding.
35,717
to are developed experiments the illustrate Numerical provided theory.
Numerical experiments are provided to illustrate the developed theory.
35,718
manifold. Definite Specifically, Symmetric a it is (SPD) Positive
Specifically, it is a Symmetric Positive Definite (SPD) manifold.
35,719
to input being SPD layer we the rectifying the However singular. matrices exploit from prevent
However we exploit the rectifying layer to prevent the input SPD matrices from being singular.
35,720
with objective innovative a introduce function. the also We non-linear an of network learning proposed
We also introduce a non-linear learning of the proposed network with an innovative objective function.
35,721
version of We developed a also named deep have RieMNet, as DRieMNet.
We have also developed a deep version of RieMNet, named as DRieMNet.
35,722
the results our the method experimental superiority Extensive of over state-of-the-art. show
Extensive experimental results show the superiority of our method over the state-of-the-art.
35,723
next is to of KT to question. response The the objective a predict student's
The objective of KT is to predict a student's response to the next question.
35,724
propose address regularization to a these novel then problems. We
We then propose a novel regularization to address these problems.
35,725
method to us proposed DLKT. The allows achieve trustworthy
The proposed method allows us to achieve trustworthy DLKT.
35,726
Such in cortex. the observed mechanisms activity can patterns models spatiotemporal of search underlying facilitate
Such models can facilitate the search of mechanisms underlying observed spatiotemporal activity patterns in cortex.
35,727
spiking (LFPs) the With from local forward activity. models, can be simulated field potentials computed
With forward models, local field potentials (LFPs) can be computed from the simulated spiking activity.
35,728
probabilities distance-dependent the densities, introduces and connection and neuron upscaling delays. The preserves
The upscaling preserves the neuron densities, and introduces distance-dependent connection probabilities and delays.
35,729
disentangled a crucial task. yet representations interpretable is challenging Learning
Learning interpretable disentangled representations is a crucial yet challenging task.
35,730
a of takes the form Our dual autoencoder model structure.
Our model takes the form of a dual autoencoder structure.
35,731
dual be semi-supervised effective. setting, This out very to for swap mechanism, turns tailored
This dual swap mechanism, tailored for semi-supervised setting, turns out to be very effective.
35,732
biometric introduced. security these privacy and address been concerns has cancelable To
To address these security and privacy concerns cancelable biometric has been introduced.
35,733
nearest ridge-based each sector and neighbor are computed. in The minutiae identified, are features
The nearest neighbor minutiae in each sector are identified, and ridge-based features are computed.
35,734
In stimulating experiments, consolidation be presynaptic neurons. synaptic by repeatedly induced can
In experiments, synaptic consolidation can be induced by repeatedly stimulating presynaptic neurons.
35,735
- tool. method gold LCModel the is standard commercial quantification a The fitting current
The current gold standard quantification method is the LCModel - a commercial fitting tool.
35,736
involve simulated Experiments testing brain and in-vivo on and spectra. training human
Experiments involve training and testing on simulated and in-vivo human brain spectra.
35,737
improvement standard consistent classification image are reported datasets. experiments over demonstrate which a Preliminary
Preliminary experiments are reported which demonstrate a consistent improvement over standard image classification datasets.
35,738
tasks is extremely Applying in reinforcement physical-world challenging. learning
Applying reinforcement learning in physical-world tasks is extremely challenging.
35,739
disclose that results recovers The faithfully the Taobao of environment. important Virtual real properties
The results disclose that Virtual Taobao faithfully recovers important properties of the real environment.
35,740
is boundary between relationship estimation What landmarks localisation? the and
What is the relationship between boundary estimation and landmarks localisation?
35,741
other Our easily can method information datasets. from also integrate
Our method can also easily integrate information from other datasets.
35,742
our selection and demonstrate various We in technique segmentation of keypoint effectiveness the applications.
We demonstrate the effectiveness of our technique in various segmentation and keypoint selection applications.
35,743
an information enrich acquisitions Multi-contrast for available of MRI magnitude the anatomy diagnosis. of
Multi-contrast MRI acquisitions of an anatomy enrich the magnitude of information available for diagnosis.
35,744
factor. excessive associated be additional contrasts with may scan times Yet, limiting a
Yet, excessive scan times associated with additional contrasts may be a limiting factor.
35,745
diminished at higher factors acceleration decreases sampling particularly density performance towards reconstruction, In high-spatial-frequencies. with
In reconstruction, performance decreases towards higher acceleration factors with diminished sampling density particularly at high-spatial-frequencies.
35,746
The improve multi-contrast of MRI quality proposed method and can the efficiency exams. scan help
The proposed method can help improve the quality and scan efficiency of multi-contrast MRI exams.
35,747
correlated one-self to protect Physiological preparations biological danger. are with from changes these
Physiological changes are correlated with these biological preparations to protect one-self from danger.
35,748
extend core We idea key the with contributions. two
We extend the core idea with two key contributions.
35,749
the show qualitative and results both segmentations of we inpaintings. Moreover,
Moreover, we show qualitative results of both the segmentations and inpaintings.
35,750
summarization. This the addresses paper problem video of
This paper addresses the problem of video summarization.
35,751
In summarization paper, this a problem. labeling sequence we formulate as video
In this paper, we formulate video summarization as a sequence labeling problem.
35,752
effectiveness Extensive benchmark the two of our on and models. analysis datasets demonstrate experiments
Extensive experiments and analysis on two benchmark datasets demonstrate the effectiveness of our models.
35,753
combining results. the our global prediction local features, we By the final and obtain
By combining the global and the local features, we obtain our final prediction results.
35,754
we After convolution layers, hashing and pooling. deploy for block-wise binary the histogram
After the convolution layers, we deploy binary hashing and block-wise histogram for pooling.
35,755
is identification use for biometrics. known The as physical and behavioral human of characteristics
The use of physical and behavioral characteristics for human identification is known as biometrics.
35,756
is traits biometrics fingerprint Among used. most many the widely the available, the
Among the many biometrics traits available, the fingerprint is the most widely used.
35,757
also have due In addition, recognition focused on multibiometrics researches to many advantages. its recent
In addition, recent researches have also focused on multibiometrics recognition due to its many advantages.
35,758
properly to to A understand kitchen the needs environment cooking cooking robot any activities. continue
A kitchen robot properly needs to understand the cooking environment to continue any cooking activities.
35,759
state not so But like researched detection detection. as object's has well far object been
But object's state detection has not been researched well so far as like object detection.
35,760
of The model optimizers. terms layers, is robustly and analyzed different in
The model is analyzed robustly in terms of different layers, and optimizers.
35,761
a results show qualitative progression state-of-the-art. clear model delivers our quantitative over and Both
Both quantitative and qualitative results show our model delivers a clear progression over state-of-the-art.
35,762
of preventable irreversible the world. in blindness cause but the leading Glaucoma is
Glaucoma is the leading cause of irreversible but preventable blindness in the world.
35,763
the in that observed recently glaucoma been ocular hemodynamics. has It changes induces
It has been recently observed that glaucoma induces changes in the ocular hemodynamics.
35,764
characterizing approach first this propose hemodynamics. those paper for In changes using we computational a
In this paper we propose a first approach for characterizing those changes using computational hemodynamics.
35,765
is from on patient-specific simulation images. The extracted fundus arterial trees performed
The simulation is performed on patient-specific arterial trees extracted from fundus images.
35,766
computation. weight involved score MCW The the technique for matcher achieves in each the optimal
The MCW technique achieves the optimal weight for each matcher involved in the score computation.
35,767
DS to applied to final Further, decision. the induced the theory scores is output
Further, DS theory is applied to the induced scores to output the final decision.
35,768
each A of study the performance shrinkage the evaluate priors. to conducted of is simulation
A simulation study is conducted to evaluate the performance of each of the shrinkage priors.
35,769
illustration empirical data. pan-European regional we use an For growth economic
For an empirical illustration we use pan-European regional economic growth data.
35,770
yet to e.g. well-understood knowledge, not ontological However, what is extent it
However, it is not yet well-understood to what extent ontological knowledge, e.g.
35,771
be (existential) in given principled set a of can a as way. embedded rules,
given as a set of (existential) rules, can be embedded in a principled way.
35,772
convex relations model a are in modelled which regions. Second, we as study
Second, we study a model in which relations are modelled as convex regions.
35,773
previously method improvements methods. the considerably ensures than proposed This accuracy discrimination
This method ensures considerably discrimination accuracy improvements than the previously proposed methods.
35,774
regions. find Recent object instances while categorizing candidate detectors
Recent object detectors find instances while categorizing candidate regions.
35,775
find in We meta-measures. improvements almost large all the
We find large improvements in almost all the meta-measures.
35,776
Experimental evaluations method current to the the state-of-the-art. outperforms in that comparison indicate
Experimental evaluations indicate that the method outperforms in comparison to the current state-of-the-art.
35,777
shown favorable trackers (SVM) machines based performance. have vector Recently, part-based support and
Recently, part-based and support vector machines (SVM) based trackers have shown favorable performance.
35,778
training limit updating applications. the their and process online real-time time-consuming Nonetheless,
Nonetheless, the time-consuming online training and updating process limit their real-time applications.
35,779
In work, recursion to extend GLMB propagate posterior. multi-object analytic we this the the
In this work, we extend the analytic GLMB recursion to propagate the multi-object posterior.
35,780
on mathematical models mass-action based focus deterministic we Here on kinetics.
Here we focus on mathematical models based on deterministic mass-action kinetics.
35,781
positive equations These models of systems coupled nonlinear orthant. the on differential are
These models are systems of coupled nonlinear differential equations on the positive orthant.
35,782
be Recent accurate. surprisingly fast, can low-bitwidth that compact has neural work networks shown
Recent work has shown that fast, compact low-bitwidth neural networks can be surprisingly accurate.
35,783
to experiments We verify effectiveness extensive of the method. the perform proposed
We perform extensive experiments to verify the effectiveness of the proposed method.
35,784
SBU and RGB+D dataset. Kinect proposed Interaction NTU evaluate We on method our
We evaluate our proposed method on NTU RGB+D and SBU Kinect Interaction dataset.
35,785
method interactions. actions of improves addition, the accuracy our in significantly the two-person In
In addition, our method significantly improves the accuracy of the actions in two-person interactions.
35,786
by signals measured and EEG Rating Scale scores. (HDRS) were Depression The responses Hamilton ketamine
The ketamine responses were measured by EEG signals and Hamilton Depression Rating Scale (HDRS) scores.
35,787
EEG patterns may advance. ketamine The for baseline prefrontal recognising account effects at in
The prefrontal EEG patterns at baseline may account for recognising ketamine effects in advance.
35,788
irregular their because trees mostly twisted are of branches. While other shape in
While other trees are irregular in shape mostly because of their twisted branches.
35,789
The other self-shadows than leaves. palms ordinary have large
The palms have large self-shadows other than ordinary leaves.
35,790
in shown palms in of the satellite Then image. field the any it number counts
Then it counts the number of palms in any field shown in the satellite image.
35,791
Dataset). (UCSD Dataset on two Avenue approach our benchmarks We test and popular
We test our approach on two popular benchmarks (UCSD Dataset and Avenue Dataset).
35,792
of potential networks zero-shot great task neural convolutional shown have for learning. the Graph recently
Graph convolutional neural networks have recently shown great potential for the task of zero-shot learning.
35,793
connections node's descendants. on its are These and based relationship a ancestors to added
These connections are added based on a node's relationship to its ancestors and descendants.
35,794
that constraints in specified research in operations the model. be require problems Many
Many problems in operations research require that constraints be specified in the model.
35,795
is the Determining laborsome a right task. hard and constraints
Determining the right constraints is a hard and laborsome task.
35,796
propose process machine learning an principles. to approach We automate this artificial intelligence and using
We propose an approach to automate this process using artificial intelligence and machine learning principles.
35,797
how adjustment to the However, degree never been determine has of discussed.
However, how to determine the degree of adjustment has never been discussed.
35,798
thus testing structure. for market unique market organization a the The environment provides
The market thus provides a unique testing environment for the market organization structure.
35,799