In order to alleviate the huge demand for annotated datasets for different tasks, many recent natural language processing datasets have adopted automated pipelines for fast-tracking usable data. However, model training with such datasets poses a chal
lenge because popular optimization objectives are not robust to label noise induced in the annotation generation process. Several noise-robust losses have been proposed and evaluated on tasks in computer vision, but they generally use a single dataset-wise hyperparamter to control the strength of noise resistance. This work proposes novel instance-adaptive training frameworks to change single dataset-wise hyperparameters of noise resistance in such losses to be instance-wise. Such instance-wise noise resistance hyperparameters are predicted by special instance-level label quality predictors, which are trained along with the main classification models. Experiments on noisy and corrupted NLP datasets show that proposed instance-adaptive training frameworks help increase the noise-robustness provided by such losses, promoting the use of the frameworks and associated losses in NLP models trained with noisy data.
The choice of negative examples is important in noise contrastive estimation. Recent works find that hard negatives---highest-scoring incorrect examples under the model---are effective in practice, but they are used without a formal justification. We
develop analytical tools to understand the role of hard negatives. Specifically, we view the contrastive loss as a biased estimator of the gradient of the cross-entropy loss, and show both theoretically and empirically that setting the negative distribution to be the model distribution results in bias reduction. We also derive a general form of the score function that unifies various architectures used in text retrieval. By combining hard negatives with appropriate score functions, we obtain strong results on the challenging task of zero-shot entity linking.
An abundance of methodological work aims to detect hateful and racist language in text. However, these tools are hampered by problems like low annotator agreement and remain largely disconnected from theoretical work on race and racism in the social
sciences. Using annotations of 5188 tweets from 291 annotators, we investigate how annotator perceptions of racism in tweets vary by annotator racial identity and two text features of the tweets: relevant keywords and latent topics identified through structural topic modeling. We provide a descriptive summary of our data and estimate a series of generalized linear models to determine if annotator racial identity and our 12 latent topics, alone or in combination, explain the way racial sentiment was annotated, net of relevant annotator characteristics and tweet features. Our results show that White and non-White annotators exhibit significant differences in ratings when reading tweets with high prevalence of particular, racially-charged topics. We conclude by suggesting how future methodological work can draw on our results and further incorporate social science theory into analyses.
Wireless communication has become more popular over the last decade. Although
there are many advantages to use wireless over fixed line systems, there is a major
disadvantage, which is a high transmission error rate in a temporal and permanent nois
y
environment. There are techniques to reduce effectiveness off transmission errors in
wireless communications. We have discussed in this paper. Many methods and techniques
have been implementing to deal and resist errors (by increasing transmission power,
switching modulation scheme) as well as evaluation the wireless connection performance
taking into account many factors.
This works aims to study the performance of the mobile system via studying the
effects of adaptive modulation technique on the data rate. Fuzzy logic technique is used as
a tool to analyze the data rate of the downlink connection depending on type
of
modulation, coverage quality and the distant between the receiver and the transmitter. For
parameters were considered as input parameters for the fuzzy logic system mainly are:
power of the station, quality of coverage, the type of modulation technique and the signal
to interface noise ratio. One parameter is used as an output of the fuzzy system which is
spectrum efficiency in bit per second per hertz. Analytical study to select the optimal data
rates for different modulation techniques will be considered. Additionally, studying the
input and output memberships of the data rate parameters using fuzzy logic will be
explained. Keywords: Data rate, downlink, signal to noise ratio, modulation type, Fuzzy
logic, spectrum efficiency.
The effect of thermal noise and shot noise on the receiver sensitivity in a digital
optical communication system will be studied and analyzed. Also, the positive and
negative effects of Parameters on the receiver sensitivity will be highlighted. Th
e
programming tools (MATLAB) is used to study the relationship between the receiver
sensitivity and bit rate. It showed that the shot-noise-limited receiver sensitivity is much
better than the thermal -noise-limited receiver sensitivity with best 30dB. The study also
provided an explanation for the preference of optical detector (APD) compared to that in
(PIN) optical detector from sensitivity side due to its internal gain, this can be achieved
when the thermal noise is dominant. However, the preference of optical detector(PIN) will
be better in case of shot noise is dominant.
Analyzed results show that there is a significant improvement in the shot-noiselimited
receiver sensitivity for high quantum efficiency and high wavelength. However, the
sensitivity is deteriorated to get a lower error rate. The thermal -noise-limited receiver
sensitivity showed an improving by increasing the responsivity and the load resistance.
However, it decreases by increasing the number of noise and the error rate.
Speech denoising is a field of engineering that studies techniques used to recover the
original signal from the noisy signal corrupted with different types of noise, such as
broadband noise and narrowband noise, and other types present in environme
nt, but the
spectral subtraction technique consider the most prominent in this area .
In this search we will discuss the parameters impact of the modified spectral
subtraction algorithm and the time window length in the enhancement of speech that
corrupted with broadband noise. We done the study and determine the ideal parameters
values and the ideal window length with different values for the signal -to-noise ratio SNR
for noisy speech and we discuss 18 case for each value. We done
the simulation using MATLAB software and the results were compared based on
improving the value of SNR for each case .
Nursery ward is the safe place for premature after deliverey . It contains the basic
requirements of life to him. but full of noise caused by alarms of devices, conversations of
health team, ringing phones, and close the doors, exceeds what is reco
mmended by the
American Academy of Pediatrics in the Nursery wards 45 dBA (decibel A-weighting).
Although it stimulates the autonomic system leading to physiologic instability like
disorder in the level of oxygenation. We found in this study conducted on 30 premature
with pre and post test, that the level of noise within first hour of non interventions reached
67.2 dBA, level of oxygenation inversely with noise level in it, and reduction of noise level
under 45 dBA at second hour increases the oxygenation significantly. so we suggests to
make reduce noise interventions within daily practices of nursing procedures, and
subsequent research to study effect of noise on the rest of vital premature indicators.
Since Electroencephalogram (EEG) signals have very small magnitude, it's very hard
to capture these signals without having noise (produced by surrounding artifacts) affect the
real EEG signals, so it is necessary to use Filters to remove noise.
Th
is work proposes a design of an electronic circuit using a microcontroller, an
instrumentation amplifier and an operational amplifier able to capture EEG signals, convert
the captured signals from analog state to digital one and send the converted signal (digital
signal) to a group of three digital filters.
This paper gives a design of three digital elliptic filters ready to be used in real time
filtering of EEG signals (which preliminary represents the condition of the brain) making
the software part which complements the hardware part in the EEG signals capturing
system.
Finally we are going to show the way of using the designed electronic circuit with
the three designed digital filters, demonstrate and discuss the results of this work.
We have used Eagle 6.6 software to design and draw the circuit, CodeVision AVR
3.12 software to write the program downloaded on the microcontroller, Mathworks
MATLAB 2014a software to design the three digital filters and Mathworks MATLAB
2014a Simulink tool to make the appropriate experiments and get the results.
This research aims to study the effects of weaknesses on the performance of the existing network such as the impact of Attenuation and crosstalk on the data transfer, and it presents also the improvements of the standards of DSL systems to achieve th
e maximum transition rates in order to have greater speed in the installation of high-speed subscriber lines.
It also offers a new range of services to communication companies through the provision of new xDSL technologies in the near future. Thus, it enables its customers access to broadband services and provides tips for the Syrian Telecommunications Company in their choice of systems.
This research analyses the essence of the techniques and the strengths and weaknesses of the criteria for All Digital Mode digital subscriber line mentioned in Annex A and for higher upstream digital subscriber line mentioned in Annex M. It also includes an analytical study of the damping and crosstalk to show how to consider any criteria as a better choice.
Finally, It includes programmatically represented simulation of the signal which is sent over transmission lines under the influence of the studied factors.
