Leila Arras, Franziska Horn, Grégoire Montavon, Klaus-Robert Müller, Wojciech Samek
Comments: 19 pages, 7 figures
Subjects: Computation and Language (cs.CL); Information Retrieval (cs.IR); Learning (cs.LG); Machine Learning (stat.ML)
Text documents can be described by a number of abstract concepts such as
semantic category, writing style, or sentiment. Machine learning (ML) models
have been trained to automatically map documents to these abstract concepts,
allowing to annotate very large text collections, more than could be processed
by a human in a lifetime. Besides predicting the text’s category very
accurately, it is also highly desirable to understand how and why the
categorization process takes place. In this paper, we demonstrate that such
understanding can be achieved by tracing the classification decision back to
individual words using layer-wise relevance propagation (LRP), a recently
developed technique for explaining predictions of complex non-linear
classifiers. We train two word-based ML models, a convolutional neural network
(CNN) and a bag-of-words SVM classifier, on a topic categorization task and
adapt the LRP method to decompose the predictions of these models onto words.
Resulting scores indicate how much individual words contribute to the overall
classification decision. This enables one to distill relevant information from
text documents without an explicit semantic information extraction step. We
further use the word-wise relevance scores for generating novel vector-based
document representations which capture semantic information. Based on these
document vectors, we introduce a measure of model explanatory power and show
that, although the SVM and CNN models perform similarly in terms of
classification accuracy, the latter exhibits a higher level of explainability
which makes it more comprehensible for humans and potentially more useful for
other applications.
Yann N. Dauphin, Angela Fan, Michael Auli, David Grangier
Subjects: Computation and Language (cs.CL)
The pre-dominant approach to language modeling to date is based on recurrent
neural networks. In this paper we present a convolutional approach to language
modeling. We introduce a novel gating mechanism that eases gradient propagation
and which performs better than the LSTM-style gating of (Oord et al, 2016)
despite being simpler. We achieve a new state of the art on WikiText-103 as
well as a new best single-GPU result on the Google Billion Word benchmark. In
settings where latency is important, our model achieves an order of magnitude
speed-up compared to a recurrent baseline since computation can be parallelized
over time. To our knowledge, this is the first time a non-recurrent approach
outperforms strong recurrent models on these tasks.
Kamal Sarkar
Comments: This work is awarded the first prize in the NLP tool contest on “POS Tagging for Code-Mixed Indian Social Media Text”, held in conjunction with the 13th International Conference on Natural Language Processing 2016(ICON 2016), Indian Institute of Technology (BHU), India
Subjects: Computation and Language (cs.CL)
In this work, we describe a conditional random fields (CRF) based system for
Part-Of- Speech (POS) tagging of code-mixed Indian social media text as part of
our participation in the tool contest on POS tagging for codemixed Indian
social media text, held in conjunction with the 2016 International Conference
on Natural Language Processing, IIT(BHU), India. We participated only in
constrained mode contest for all three language pairs, Bengali-English,
Hindi-English and Telegu-English. Our system achieves the overall average F1
score of 79.99, which is the highest overall average F1 score among all 16
systems participated in constrained mode contest.
Lei Shu, Bing Liu, Hu Xu, Annice Kim
Comments: 10 pages
Subjects: Computation and Language (cs.CL); Learning (cs.LG)
One of the key tasks of sentiment analysis of product reviews is to extract
product aspects or features that users have expressed opinions on. In this
work, we focus on using supervised sequence labeling as the base approach to
performing the task. Although several extraction methods using sequence
labeling methods such as Conditional Random Fields (CRF) and Hidden Markov
Models (HMM) have been proposed, we show that this supervised approach can be
significantly improved by exploiting the idea of concept sharing across
multiple domains. For example, “screen” is an aspect in iPhone, but not only
iPhone has a screen, many electronic devices have screens too. When “screen”
appears in a review of a new domain (or product), it is likely to be an aspect
too. Knowing this information enables us to do much better extraction in the
new domain. This paper proposes a novel extraction method exploiting this idea
in the context of supervised sequence labeling. Experimental results show that
it produces markedly better results than without using the past information.
Leila Arras, Franziska Horn, Grégoire Montavon, Klaus-Robert Müller, Wojciech Samek
Comments: 19 pages, 7 figures
Subjects: Computation and Language (cs.CL); Information Retrieval (cs.IR); Learning (cs.LG); Machine Learning (stat.ML)
Text documents can be described by a number of abstract concepts such as
semantic category, writing style, or sentiment. Machine learning (ML) models
have been trained to automatically map documents to these abstract concepts,
allowing to annotate very large text collections, more than could be processed
by a human in a lifetime. Besides predicting the text’s category very
accurately, it is also highly desirable to understand how and why the
categorization process takes place. In this paper, we demonstrate that such
understanding can be achieved by tracing the classification decision back to
individual words using layer-wise relevance propagation (LRP), a recently
developed technique for explaining predictions of complex non-linear
classifiers. We train two word-based ML models, a convolutional neural network
(CNN) and a bag-of-words SVM classifier, on a topic categorization task and
adapt the LRP method to decompose the predictions of these models onto words.
Resulting scores indicate how much individual words contribute to the overall
classification decision. This enables one to distill relevant information from
text documents without an explicit semantic information extraction step. We
further use the word-wise relevance scores for generating novel vector-based
document representations which capture semantic information. Based on these
document vectors, we introduce a measure of model explanatory power and show
that, although the SVM and CNN models perform similarly in terms of
classification accuracy, the latter exhibits a higher level of explainability
which makes it more comprehensible for humans and potentially more useful for
other applications.
Nan Ding, Sebastian Goodman, Fei Sha, Radu Soricut
Comments: 11 pages
Subjects: Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV)
We introduce a new multi-modal task for computer systems, posed as a combined
vision-language comprehension challenge: identifying the most suitable text
describing a scene, given several similar options. Accomplishing the task
entails demonstrating comprehension beyond just recognizing “keywords” (or
key-phrases) and their corresponding visual concepts. Instead, it requires an
alignment between the representations of the two modalities that achieves a
visually-grounded “understanding” of various linguistic elements and their
dependencies. This new task also admits an easy-to-compute and well-studied
metric: the accuracy in detecting the true target among the decoys.
The paper makes several contributions: an effective and extensible mechanism
for generating decoys from (human-created) image captions; an instance of
applying this mechanism, yielding a large-scale machine comprehension dataset
(based on the COCO images and captions) that we make publicly available; human
evaluation results on this dataset, informing a performance upper-bound; and
several baseline and competitive learning approaches that illustrate the
utility of the proposed task and dataset in advancing both image and language
comprehension. We also show that, in a multi-task learning setting, the
performance on the proposed task is positively correlated with the end-to-end
task of image captioning.
Amanda Bienz, William D. Gropp, Luke N. Olson
Subjects: Distributed, Parallel, and Cluster Computing (cs.DC); Mathematical Software (cs.MS)
This paper introduces a method to reduce communication that is injected into
the network during a sparse matrix-vector multiply by reorganizing messages on
each node. This results in a reduction of the inter-node communication,
replaced by less-costly intra-node communication, which reduces both the number
and size of messages that are injected into the network.
Kuniaki Saito, Yusuke Mukuta, Yoshitaka Ushiku, Tatsuya Harada
Subjects: Learning (cs.LG); Machine Learning (stat.ML)
Obtaining common representations from different modalities is important in
that they are interchangeable with each other in a classification problem. For
example, we can train a classifier on image features in the common
representations and apply it to the testing of the text features in the
representations. Existing multi-modal representation learning methods mainly
aim to extract rich information from paired samples and train a classifier by
the corresponding labels; however, collecting paired samples and their labels
simultaneously involves high labor costs. Addressing paired modal samples
without their labels and single modal data with their labels independently is
much easier than addressing labeled multi-modal data. To obtain the common
representations under such a situation, we propose to make the distributions
over different modalities similar in the learned representations, namely
modality-invariant representations. In particular, we propose a novel algorithm
for modality-invariant representation learning, named Deep Modality Invariant
Adversarial Network (DeMIAN), which utilizes the idea of Domain Adaptation
(DA). Using the modality-invariant representations learned by DeMIAN, we
achieved better classification accuracy than with the state-of-the-art methods,
especially for some benchmark datasets of zero-shot learning.
Marcell V-Chanlatte, Jyotirmoy V. Deshmukh, Xiaoqing Jin, Sanjit Seshia
Subjects: Learning (cs.LG); Logic in Computer Science (cs.LO)
To effectively analyze and design cyberphysical systems (CPS), designers
today have to combat the data deluge problem, i.e., the burden of processing
intractably large amounts of data produced by complex models and experiments.
In this work, we utilize monotonic Parametric Signal Temporal Logic (PSTL) to
design features for unsupervised classification of time series data. This
enables using off-the-shelf machine learning tools to automatically cluster
similar traces with respect to a given PSTL formula. We demonstrate how this
technique produces simple and interpetable formulas that are amenable to
analysis and understanding using a few representative examples. We illustrate
this with a number of case studies related to automotive engine testing,
highway traffic analysis, and auto-grading massively open online courses.
Onur Atan, William R. Zame, Qiaojun Feng, Mihaela van der Schaar
Subjects: Machine Learning (stat.ML); Learning (cs.LG)
This paper proposes a novel approach for constructing effective personalized
policies when the observed data lacks counter-factual information, is biased
and possesses many features. The approach is applicable in a wide variety of
settings from healthcare to advertising to education to finance. These settings
have in common that the decision maker can observe, for each previous instance,
an array of features of the instance, the action taken in that instance, and
the reward realized — but not the rewards of actions that were not taken: the
counterfactual information. Learning in such settings is made even more
difficult because the observed data is typically biased by the existing policy
(that generated the data) and because the array of features that might affect
the reward in a particular instance — and hence should be taken into account
in deciding on an action in each particular instance — is often vast. The
approach presented here estimates propensity scores for the observed data,
infers counterfactuals, identifies a (relatively small) number of features that
are (most) relevant for each possible action and instance, and prescribes a
policy to be followed. Comparison of the proposed algorithm against the
state-of-art algorithm on actual datasets demonstrates that the proposed
algorithm achieves a significant improvement in performance.
Jesse H. Krijthe
Comments: Presented at RRPR 2016: 1st Workshop on Reproducible Research in Pattern Recognition
Subjects: Machine Learning (stat.ML); Learning (cs.LG)
In this paper, we introduce a package for semi-supervised learning research
in the R programming language called RSSL. We cover the purpose of the package,
the methods it includes and comment on their use and implementation. We then
show, using several code examples, how the package can be used to replicate
well-known results from the semi-supervised learning literature.
Lei Shu, Bing Liu, Hu Xu, Annice Kim
Comments: 10 pages
Subjects: Computation and Language (cs.CL); Learning (cs.LG)
One of the key tasks of sentiment analysis of product reviews is to extract
product aspects or features that users have expressed opinions on. In this
work, we focus on using supervised sequence labeling as the base approach to
performing the task. Although several extraction methods using sequence
labeling methods such as Conditional Random Fields (CRF) and Hidden Markov
Models (HMM) have been proposed, we show that this supervised approach can be
significantly improved by exploiting the idea of concept sharing across
multiple domains. For example, “screen” is an aspect in iPhone, but not only
iPhone has a screen, many electronic devices have screens too. When “screen”
appears in a review of a new domain (or product), it is likely to be an aspect
too. Knowing this information enables us to do much better extraction in the
new domain. This paper proposes a novel extraction method exploiting this idea
in the context of supervised sequence labeling. Experimental results show that
it produces markedly better results than without using the past information.
C.J.C. Burges, T. Hart, Z. Yang, S. Cucerzan, R.W. White, A. Pastusiak, J. Lewis
Subjects: Artificial Intelligence (cs.AI); Learning (cs.LG)
Modern statistical machine learning (SML) methods share a major limitation
with the early approaches to AI: there is no scalable way to adapt them to new
domains. Human learning solves this in part by leveraging a rich, shared,
updateable world model. Such scalability requires modularity: updating part of
the world model should not impact unrelated parts. We have argued that such
modularity will require both “correctability” (so that errors can be corrected
without introducing new errors) and “interpretability” (so that we can
understand what components need correcting).
To achieve this, one could attempt to adapt state of the art SML systems to
be interpretable and correctable; or one could see how far the simplest
possible interpretable, correctable learning methods can take us, and try to
control the limitations of SML methods by applying them only where needed. Here
we focus on the latter approach and we investigate two main ideas: “Teacher
Assisted Learning”, which leverages crowd sourcing to learn language; and
“Factored Dialog Learning”, which factors the process of application
development into roles where the language competencies needed are isolated,
enabling non-experts to quickly create new applications.
We test these ideas in an “Automated Personal Assistant” (APA) setting, with
two scenarios: that of detecting user intent from a user-APA dialog; and that
of creating a class of event reminder applications, where a non-expert
“teacher” can then create specific apps. For the intent detection task, we use
a dataset of a thousand labeled utterances from user dialogs with Cortana, and
we show that our approach matches state of the art SML methods, but in addition
provides full transparency: the whole (editable) model can be summarized on one
human-readable page. For the reminder app task, we ran small user studies to
verify the efficacy of the approach.
Tong Wu, Prudhvi Gurram, Raghuveer M. Rao, Waheed U. Bajwa
Comments: Submitted for journal publication
Subjects: Machine Learning (stat.ML); Learning (cs.LG)
Representation of human actions as a sequence of human body movements or
action attributes enables the development of models for human activity
recognition and summarization. We present an extension of the low-rank
representation (LRR) model, termed the clustering-aware structure-constrained
low-rank representation (CS-LRR) model, for unsupervised learning of human
action attributes from video data. Our model is based on the union-of-subspaces
(UoS) framework, and integrates spectral clustering into the LRR optimization
problem for better subspace clustering results. We lay out an efficient linear
alternating direction method to solve the CS-LRR optimization problem. We also
introduce a hierarchical subspace clustering approach, termed hierarchical
CS-LRR, to learn the attributes without the need for a priori specification of
their number. By visualizing and labeling these action attributes, the
hierarchical model can be used to semantically summarize long video sequences
of human actions at multiple resolutions. A human action or activity can also
be uniquely represented as a sequence of transitions from one action attribute
to another, which can then be used for human action recognition. We demonstrate
the effectiveness of the proposed model for semantic summarization and action
recognition through comprehensive experiments on five real-world human action
datasets.
Leila Arras, Franziska Horn, Grégoire Montavon, Klaus-Robert Müller, Wojciech Samek
Comments: 19 pages, 7 figures
Subjects: Computation and Language (cs.CL); Information Retrieval (cs.IR); Learning (cs.LG); Machine Learning (stat.ML)
Text documents can be described by a number of abstract concepts such as
semantic category, writing style, or sentiment. Machine learning (ML) models
have been trained to automatically map documents to these abstract concepts,
allowing to annotate very large text collections, more than could be processed
by a human in a lifetime. Besides predicting the text’s category very
accurately, it is also highly desirable to understand how and why the
categorization process takes place. In this paper, we demonstrate that such
understanding can be achieved by tracing the classification decision back to
individual words using layer-wise relevance propagation (LRP), a recently
developed technique for explaining predictions of complex non-linear
classifiers. We train two word-based ML models, a convolutional neural network
(CNN) and a bag-of-words SVM classifier, on a topic categorization task and
adapt the LRP method to decompose the predictions of these models onto words.
Resulting scores indicate how much individual words contribute to the overall
classification decision. This enables one to distill relevant information from
text documents without an explicit semantic information extraction step. We
further use the word-wise relevance scores for generating novel vector-based
document representations which capture semantic information. Based on these
document vectors, we introduce a measure of model explanatory power and show
that, although the SVM and CNN models perform similarly in terms of
classification accuracy, the latter exhibits a higher level of explainability
which makes it more comprehensible for humans and potentially more useful for
other applications.
Ashish Shrivastava, Tomas Pfister, Oncel Tuzel, Josh Susskind, Wenda Wang, Russ Webb
Comments: Submitted for review to a conference on Nov 15, 2016
Subjects: Computer Vision and Pattern Recognition (cs.CV); Learning (cs.LG)
With recent progress in graphics, it has become more tractable to train
models on synthetic images, potentially avoiding the need for expensive
annotations. However, learning from synthetic images may not achieve the
desired performance due to a gap between synthetic and real image
distributions. To reduce this gap, we propose Simulated+Unsupervised (S+U)
learning, where the task is to learn a model to improve the realism of a
simulator’s output using unlabeled real data, while preserving the annotation
information from the simulator. We develop a method for S+U learning that uses
an adversarial network similar to Generative Adversarial Networks (GANs), but
with synthetic images as inputs instead of random vectors. We make several key
modifications to the standard GAN algorithm to preserve annotations, avoid
artifacts and stabilize training: (i) a ‘self-regularization’ term, (ii) a
local adversarial loss, and (iii) updating the discriminator using a history of
refined images. We show that this enables generation of highly realistic
images, which we demonstrate both qualitatively and with a user study. We
quantitatively evaluate the generated images by training models for gaze
estimation and hand pose estimation. We show a significant improvement over
using synthetic images, and achieve state-of-the-art results on the MPIIGaze
dataset without any labeled real data.
Peyman Siyari, Marwan Krunz, Diep N. Nguyen
Comments: 36 pages, 8 figures
Subjects: Information Theory (cs.IT); Computer Science and Game Theory (cs.GT)
We consider joint optimization of artificial noise (AN) and information
signals in a MIMO wiretap interference network, wherein the transmission of
each link may be overheard by several MIMO-capable eavesdroppers. Each
information signal is accompanied with AN, generated by the same user to
confuse nearby eavesdroppers. Using a noncooperative game, a distributed
optimization mechanism is proposed to maximize the secrecy rate of each link.
The decision variables here are the covariance matrices for the information
signals and ANs. However, the nonconvexity of each link’s optimization problem
(i.e., best response) makes conventional convex games inapplicable, even to
find whether a Nash Equilibrium (NE) exists. To tackle this issue, we analyze
the proposed game using a relaxed equilibrium concept, called quasi-Nash
equilibrium (QNE). Under a constraint qualification condition for each player’s
problem, the set of QNEs includes the NE of the proposed game. We also derive
the conditions for the existence and uniqueness of the resulting QNE. It turns
out that the uniqueness conditions are too restrictive, and do not always hold
in typical network scenarios. Thus, the proposed game often has multiple QNEs,
and convergence to a QNE is not always guaranteed. To overcome these issues, we
modify the utility functions of the players by adding several specific terms to
each utility function. The modified game converges to a QNE even when multiple
QNEs exist. Furthermore, players have the ability to select a desired QNE that
optimizes a given social objective (e.g., sum-rate or secrecy sum-rate).
Depending on the chosen objective, the amount of signaling overhead as well as
the performance of resulting QNE can be controlled. Simulations show that due
to the QNE selection mechanism, we can achieve a significant improvement in
terms of secrecy sum-rate and power efficiency.
Roohollah Ghavamirad, Hossein Babashah, Mohammad Ali Sebt
Subjects: Information Theory (cs.IT); Data Analysis, Statistics and Probability (physics.data-an)
This paper will design non-linear frequency modulation (NLFM) signal for
Chebyshev, Kaiser, Taylor, and raised-cosine power spectral densities (PSDs).
Then, the variation of peak sidelobe level with regard to mainlobe width for
these four different window functions are analyzed. It has been demonstrated
that reduction of sidelobe level in NLFM signal can lead to increase in
mainlobe width of autocorrelation function. Furthermore, the results of power
spectral density obtained from the simulation and the desired PSD are compared.
Finally, error percentage between simulated PSD and desired PSD for different
peak sidelobe level are illustrated. The stationary phase concept is the
possible source for this error.
Francisco J. Martin-Vega, Beatriz Soret, Mari Carmen Aguayo-Torres, Istvan Z. Kovacs, Gerardo Gomez
Comments: 15 pages and 16 figures. This paper have been submitted for possible publication in IEEE Transactions on Vehicular Technology
Subjects: Information Theory (cs.IT)
Delivery of broadcast messages among vehicles for safety purposes, which is
known as one of the key ingredients of Intelligent Transportation Systems
(ITS), requires an efficient Medium Access Control (MAC) that provides low
average delay and high reliability. To this end, a Geo-Location Based Access
(GLOC) for vehicles has been proposed for Vehicle-to-Vehicle (V2V)
communications, aiming at maximizing the distance of co-channel transmitters
while preserving a low latency when accessing the resources. In this paper we
analyze, with the aid of stochastic geometry, the delivery of periodic and
non-periodic broadcast messages with GLOC, taking into account path loss and
fading as well as the random locations of transmitting vehicles. Analytical
results include the average interference, average Binary Rate (BR), capture
probability, i.e., the probability of successful message transmission, and
Energy Efficiency (EE). Mathematical analysis reveals interesting insights
about the system performance, which are validated thought extensive Monte Carlo
simulations. In particular, it is shown that the capture probability is an
increasing function with exponential dependence with respect to the transmit
power and it is demonstrated that an arbitrary high capture probability can be
achieved, as long as the number of access resources is high enough. Finally, to
facilitate the system-level design of GLOC, the optimum transmit power is
derived, which leads to a maximal EE subject to a given constraint in the
capture probability.
Ulrich Michel, Martin Kliesch, Richard Kueng, David Gross
Comments: Simplified proof of the main theorem in [arXiv:1511.01513] and a converse statement. 3+1 pages
Subjects: Information Theory (cs.IT); Quantum Physics (quant-ph)
The diamond norm plays an important role in quantum information and operator
theory. Recently, it has also been proposed as a reguralizer for low-rank
matrix recovery. The norm constants that bound the diamond norm in terms of the
nuclear norm (also known as trace norm) are explicitly known. This note
provides a simple characterization of all operators saturating the upper and
the lower bound.
Zuling Chang, Martianus Frederic Ezerman, San Ling, Huaxiong Wang
Comments: An extended abstract containing preliminary results was presented at SETA 2016
Subjects: Information Theory (cs.IT)
We determine the cycle structure of linear feedback shift register with
arbitrary monic characteristic polynomial over any finite field. For each
cycle, a method to find a state and a new way to represent the state are
proposed.
Mohammad Ali Sedaghat, Ali Bereyhi, Ralf Mueller
Subjects: Information Theory (cs.IT)
We analyze a general class of nonlinear precoders called Least Square Error
(LSE) precoders in multiuser multiple-input multiple-output broadcast channels
using the replica method from statistical physics. In LSE precoders, signal on
each antenna at base station is limited to be in a predefined set. This
predefined set is used to model several hardware constraints such a peak power,
constant envelope, discrete constellation constraints. Both Replica Symmetry
(RS) and one step Replica Symmetry Breaking (1-RSB) assumptions are applied.
For the cases of peak power constrained and constant envelope signals on the
transmit antennas, it is shown that the RS assumption provides a good
prediction. It is shown that the LSE precoder designed for Peak to Average
Power Ratio (PAPR) of (3{
m dB}) performs as well as the known Regularized
Zero Forcing (RZF) precoder with high PAPRs. Moreover, it is shown that
constant envelope LSE precoders achieve the same performance as the RZF
precoder with about (20\%) more number of transmit antennas. For PSK signals,
the RS assumption gives an optimistic prediction as the inverse load factor,
defined as the number of transmit antennas divided by the number of users,
increases. Thus, the 1-RSB assumption is applied which gives a better
prediction than the RS assumption.
Jing Qian, Feifei Gao, Gongpu Wang, Shi Jin, Hongbo Zhu
Comments: 30 pages, 11 figures
Subjects: Information Theory (cs.IT)
We study a novel communication mechanism, ambient backscatter, that utilizes
radio frequency (RF) signals transmitted from an ambient source as both energy
supply and information carrier to enable communications between low-power
devices. Different from existing non-coherent schemes, we here design the
semi-coherent detection, where channel parameters can be obtained from unknown
data symbols and a few pilot symbols. We first derive the optimal detector for
the complex Gaussian ambient RF signal from likelihood ratio test and compute
the corresponding closed-form bit error rate (BER). To release the requirement
for prior knowledge of the ambient RF signal, we next design a suboptimal
energy detector with ambient RF signals being either the complex Gaussian or
the phase shift keying (PSK). The corresponding detection thresholds, the
analytical BER, and the outage probability are also obtained in closed-form.
Interestingly, the complex Gaussian source would cause an error floor while the
PSK source does not, which brings nontrivial indication of constellation design
as opposed to the popular Gaussian-embedded literatures. Simulations are
provided to corroborate the theoretical studies.
Jie Xu, Lingjie Duan, Rui Zhang
Comments: To appear in IEEE Wireless Communications
Subjects: Information Theory (cs.IT); Cryptography and Security (cs.CR)
Conventional wireless security assumes wireless communications are rightful
and aims to protect them against malicious eavesdropping and jamming attacks.
However, emerging infrastructure-free mobile communication networks are likely
to be illegally used (e.g., by criminals or terrorists) but difficult to be
monitored, thus imposing new challenges on the public security. To tackle this
issue, this article presents a paradigm shift of wireless security to the
surveillance and intervention of infrastructure-free suspicious and malicious
wireless communications, by exploiting legitimate eavesdropping and jamming
jointly. In particular, {emph{proactive eavesdropping}} (via jamming) is
proposed to intercept and decode information from suspicious communication
links for the purpose of inferring their intentions and deciding further
measures against them. {emph{Cognitive jamming}} (via eavesdropping) is also
proposed so as to disrupt, disable, and even spoof the targeted malicious
wireless communications to achieve various intervention tasks.
Jiun-Hung Yu, Hans-Andrea Loeliger
Subjects: Information Theory (cs.IT)
This paper introduces the simultaneous partial-inverse problem for
polynomials and develops its application to decoding interleaved Reed-Solomon
codes and subfield-evaluation codes beyond half the minimum distance. The
simultaneous partial-inverse problem has a unique solution (up to a scale
factor), which can be computed by an efficient new algorithm, for which we also
offer some variations. Decoding interleaved Reed-Solomon codes and
subfield-evaluation codes (beyond half the minimum distance) can be reduced to
the simultaneous partial-inverse problem, and pertinent decoding algorithms are
obtained by easy adaptions of the simultaneous partial-inverse algorithms. The
resulting unique-decoding algorithms are new and efficient, and they have
state-of-the-art decoding capability.
Ingo Roth, Martin Kliesch, Jens Eisert, Gerhard Wunder
Comments: 9+4 pages, 7 figures
Subjects: Information Theory (cs.IT); Quantum Physics (quant-ph)
We examine and propose a solution to the problem of recovering a block sparse
signal with sparse blocks from linear measurements. Such problems naturally
emerge in the context of mobile communication, in settings motivated by
desiderata of a 5G framework. We introduce a new variant of the Hard
Thresholding Pursuit (HTP) algorithm referred to as HiHTP. For the specific
class of sparsity structures, HiHTP performs significantly better in numerical
experiments compared to HTP. We provide both a proof of convergence and a
recovery guarantee for noisy Gaussian measurements that exhibit an improved
asymptotic scaling in terms of the sampling complexity in comparison with the
usual HTP algorithm.
Ahmed El Shafie, Naofal Al-Dhahir, Ridha Hamila
Comments: Presented in Globecom 2015
Subjects: Networking and Internet Architecture (cs.NI); Information Theory (cs.IT)
We investigate joint information and energy cooperative schemes in a
slotted-time cognitive radio network with a primary transmitter-receiver pair
and a set of secondary transmitter-receiver pairs. The primary transmitter is
assumed to be an energy-harvesting node. We propose a three-stage cooperative
transmission protocol. During the first stage, the primary user releases a
portion of its time slot to the secondary nodes to send their data and to power
the energy-harvesting primary transmitter from the secondary radio-frequency
signals. During the second stage, the primary transmitter sends its data to its
destination and to the secondary nodes. During the third stage, the secondary
nodes amplify and forward the primary data. We propose five different schemes
for secondary access and powering the primary transmitter. We derive
closed-form expressions for the primary and secondary rates for all the
proposed schemes. Two of the proposed schemes use distributed beamforming to
power the primary transmitter. We design a sparsity-aware relay-selection
scheme based on the compressive sensing principles. Our numerical results
demonstrate the gains of our proposed schemes for both the primary and
secondary systems.
Ahmed El Shafie, Naofal Al-Dhahir
Comments: Published in IEEE Communications Letters this http URL
Subjects: Networking and Internet Architecture (cs.NI); Information Theory (cs.IT)
We propose a simple yet efficient scheme for a set of energy-harvesting
sensors to establish secure communication with a common destination (a master
node). An eavesdropper attempts to decode the data sent from the sensors to
their common destination. We assume a single modulation scheme that can be
implemented efficiently for energy-limited applications. We design a
multiple-access scheme for the sensors under secrecy and limited-energy
constraints. In a given time slot, each energy-harvesting sensor chooses
between sending its packet or remaining idle. The destination assigns a set of
data time slots to each sensor. The optimization problem is formulated to
maximize the secrecy sum-throughput.