Добавил:
ivanov666
Опубликованный материал нарушает ваши авторские права? Сообщите нам.
Вуз:
Предмет:
Файл:Текст как средство интерактивного общения. Учебное пособие
.pdf
61
Social facilitation theory states that a person's performance is impacted if other
people watch them. The mere presence of an audience improves the performance of
simple tasks, especially those that require stamina. "The studies have been relatively
clear so far, but the results are more heterogeneous when it comes to more complex
coordinative tasks," explains Amelie Heinrich from the Institute of Sports Science at
MLU. In general, it is assumed that performance tends to deteriorate when there is an
audience.
Heinrich is a sports psychology expert who coaches Germany's junior biathlon
squad. In her new study she took advantage of the special situation in sport caused by
the coronavirus. "The pandemic offers a unique opportunity to study an audience's
influence outside of experimental conditions in the real world," says Heinrich. She
compared the running times and shooting successes of male and female biathletes
from the 2018/2019 season with their performances in the 2020 season in the sprint
and mass start events. "The men's results were as expected: they ran faster with an
audience present, but performed more poorly in shooting," says Heinrich. While
cross-country skiing mainly requires stamina, shooting is a coordinative task.
"Interestingly, it was the other way around for women." They ran slower in the
presence of spectators, but on average, it took them an entire second less to make
their shot and, at least in the sprint, their scoring performance was five per cent
higher. The researchers believe the results are not only due to a fluctuation in the
athletes' performance. The study had a good basis of evidence, with 83 (sprint) and
34 (mass start) World Cup biathletes, and the same tendency was shown for both
disciplines.
"To our knowledge, this is the first time that a study was able to show a
different effect of the audience on men and women," says Professor Oliver Stoll,
head of the sports psychology section at MLU. Most of the previous studies on the
topic have been conducted with men mostly. "Our study raises questions about the
generalisability of the social facilitation theory and indicates there might be a
previously unknown difference between men and women," says Heinrich. She says,
this should be investigated more thoroughly in further studies for other sports that
also contain both stamina-related and coordination-related elements.
So far, the researchers can only speculate about the reasons for the possible
gender-specific performance differences in response to audiences or the lack of. "It is
possible that gender-specific stereotypes play a role," says Heinrich. For example,
men are considered to be physically stronger -- a stereotype that could be reinforced
by the presence of spectators. Some studies also show that women react more
sensitively to feedback. In any case, according to Heinrich, the findings show once
again that gender should be taken into account in psychological studies as a possible
influencing factor.
62
Unit 10
The mind's eye of a neural network system
In the background of image recognition software that can ID our friends on
social media and wildflowers in our yard are neural networks, a type of artificial
intelligence inspired by how own our brains process data. While neural networks
sprint through data, their architecture makes it difficult to trace the origin of errors
that are obvious to humans -- like confusing a Converse high-top with an ankle boot -
- limiting their use in more vital work like health care image analysis or research. A
new tool developed at Purdue University makes finding those errors as simple as
spotting mountaintops from an airplane.
"In a sense, if a neural network were able to speak, we're showing you what it
would be trying to say," said David Gleich, a Purdue professor of computer science in
the College of Science who developed the tool, which is featured in a paper published
in Nature Machine Intelligence. "The tool we've developed helps you find places
where the network is saying, 'Hey, I need more information to do what you've asked.'
I would advise people to use this tool on any high-stakes neural network decision
scenarios or image prediction task."
Code for the tool is available on GitHub, as are use case demonstrations.
Gleich collaborated on the research with Tamal K. Dey, also a Purdue professor of
computer science, and Meng Liu, a former Purdue graduate student who earned a
doctorate in computer science.
In testing their approach, Gleich's team caught neural networks mistaking the
identity of images in databases of everything from chest X-rays and gene sequences
to apparel. In one example, a neural network repeatedly mislabeled images of cars
from the Imagenette database as cassette players. The reason? The pictures were
drawn from online sales listings and included tags for the cars' stereo equipment.
Neural network image recognition systems are essentially algorithms that
process data in a way that mimics the weighted firing pattern of neurons as an image
is analyzed and identified. A system is trained to its task -- such as identifying an
animal, a garment or a tumor -- with a "training set" of images that includes data on
each pixel, tagging and other information, and the identity of the image as classified
within a particular category. Using the training set, the network learns, or "extracts,"
the information it needs in order to match the input values with the category. This
information, a string of numbers called an embedded vector, is used to calculate the
probability that the image belongs to each of the possible categories. Generally
speaking, the correct identity of the image is within the category with the highest
probability.
But the embedded vectors and probabilities don't correlate to a decisionmaking process that humans would recognize. Feed in 100,000 numbers representing
the known data, and the network produces an embedded vector of 128 numbers that
don't correspond to physical features, although they do make it possible for the
network to classify the image. In other words, you can't open the hood on the
63
algorithms of a trained system and follow along. Between the input values and the
predicted identity of the image is a proverbial "black box" of unrecognizable numbers
across multiple layers.
"The problem with neural networks is that we can't see inside the machine to
understand how it's making decisions, so how can we know if a neural network is
making a characteristic mistake?" Gleich said.
Rather than trying to trace the decision-making path of any single image
through the network, Gleich's approach makes it possible to visualize the relationship
that the computer sees among all the images in an entire database. Think of it like a
bird's-eye view of all the images as the neural network has organized them.
The relationship among the images (like network's prediction of the identity
classification of each of the images in the database) is based on the embedded vectors
and probabilities the network generates. To boost the resolution of the view and find
places where the network can't distinguish between two different classifications,
Gleich's team first developed a method of splitting and overlapping the classifications
to identify where images have a high probability of belonging to more than one
classification.
The team then maps the relationships onto a Reeb graph, a tool taken from the
field of topological data analysis. On the graph, each group of images the network
thinks are related is represented by a single dot. Dots are color coded by
classification. The closer the dots, the more similar the network considers groups to
be, and most areas of the graph show clusters of dots in a single color. But groups of
images with a high probability of belonging to more than one classification will be
represented by two differently colored overlapping dots. With a single glance, areas
where the network cannot distinguish between two classifications appear as a cluster
of dots in one color, accompanied by a smattering of overlapping dots in a second
color. Zooming in on the overlapping dots will show an area of confusion, like the
picture of the car that's been labeled both car and cassette player.
"What we're doing is taking these complicated sets of information coming out
of the network and giving people an 'in' into how the network sees the data at a
macroscopic level," Gleich said. "The Reeb map represents the important things, the
big groups and how they relate to each other, and that makes it possible to see the
errors."
"Topological Structure of Complex Predictions" was produced with the support
of the National Science Foundation and the U.S. Department of Energy.
1. Text Comprehension
1.1 Answer the following questions:
1. What are neural networks?
2. Where was a new tool for finding errors developed?
3. Where was a new tool for finding errors published?
4. What does the tool help to find?
64
5. How did Gleich’s team test their approach?
6. What is an embedded vector?
7. What is the embedded vector used for?
8. What is the relationship among the images based on?
1.2 Define if the following sentences True/False
1. The architecture of neural networks makes it easy to trace the origin of errors
that are obvious to humans.
2. A new tool developed at Purdue University makes finding those errors as easy
as spotting mountaintops from an airplane.
3. Code for the tool is available at any platform.
4. In one example, a neural network repeatedly mislabeled images of fridges from
the Imagenette database as cassette players.
5. The system is trained to identify different images within a particular category.
6. Using the training set, the network extracts the information it needs in order to
match the input values with the category.
7. The network produces an embedded vector of 130 numbers that don't
correspond to physical features.
8. Between the input values and the predicted identity of the image is an
embedded vector of unrecognizable numbers across multiple layers.
2. Vocabulary Comprehension
2.1 Give the Russian equivalents for the following word combinations:
1. Image recognition software
2. Professor of computer science
3. Nature Machine Intelligence
4. Neural network decision scenarios
5. Image prediction task
6. Case demonstrations
7. Decision-making process
8. Unrecognizable numbers
9. Topological data analysis
10. Training set
2.2 Fill in the sentences with the appropriate word or word combination given
below:
Research, algorithms, embedded vector, identity, network, errors, case
demonstrations, zooming, relationship, databases.
1. The new tool helps to find ………… easier than it was before.
65
2. Code for the tool is available on GitHub, as are use ………………………..
3. Gleich collaborated on the ………. with Tamal K. Dey, also a Purdue
professor of computer science.
4. Gleich's team caught neural networks mistaking the identity of images in
…………. of everything from chest X-rays and gene sequences to apparel.
5. Neural network image recognition systems are essentially ………..that process
data in a way that mimics the weighted firing pattern of neurons as an image is
analyzed and identified.
6. This information, a string of numbers called an ………………, is used to
calculate the probability that the image belongs to each of the possible
categories.
7. Generally speaking, the correct …………. of the image is within the category
with the highest probability.
8. The relationship among the images is based on the embedded vectors and
probabilities the ……….. generates.
9. ……………………. in on the overlapping dots will show an area of confusion.
10. Gleich's approach makes it possible to visualize the ………………. that the
computer sees among all the images in an entire database.
3. Grammar Comprehension
3.1 Define the tense and voice in the underlined predicates:
1. The tool we've developed helps you find places where the network is saying
something.
2. Gleich collaborated on the research with Tamal K. Dey, also a Purdue
professor of computer science.
3. The pictures were drawn from online sales listings and included tags for the
cars' stereo equipment.
4. A system is trained to its task -- such as identifying an animal, a garment or a
tumor.
5. The embedded vectors and probabilities don't correlate to a decision-making
process that humans would recognize.
6. On the graph, each group of images the network thinks are related is
represented by a single dot.
7. Gleich's team first developed a method of splitting and overlapping the
classifications to identify where images have a high probability of belonging to
more than one classification.
8. Topological Structure of Complex Predictions" was produced with the support
of the National Science Foundation and the U.S. Department of Energy.
3.2 Define the part of speech and its function of the underlined words in the
following sentences:
66
1. A new tool developed at Purdue University makes finding those errors as
simple as spotting mountaintops from an airplane.
2. In testing their approach, Gleich's team caught neural networks mistaking the
identity of images in databases of everything from chest X-rays and gene
sequences to apparel.
3. A system is trained to its task -- such as identifying an animal, a garment or a
tumor -- with a "training set" of images that includes data on each pixel,
tagging and other information, and the identity of the image as classified
within a particular category.
4. Using the training set, the network learns, or "extracts," the information it
needs in order to match the input values with the category.
5. Feed in 100,000 numbers representing the known data, and the network
produces an embedded vector of 128 numbers that don't correspond to physical
features.
6. The team then maps the relationships onto a Reeb graph, a tool taken from the
field of topological data analysis.
7. "What we're doing is taking these complicated sets of information coming out
of the network and giving people an 'in' into how the network sees the data at a
macroscopic level," Gleich said.
8. With a single glance, areas where the network cannot distinguish between two
classifications appear as a cluster of dots in one color, accompanied by a
smattering of overlapping dots in a second color.
4. Discussion Part
Here are given the phrases which may be useful for discussion. Study them before
doing the task
Во время дискуссии люди часто пользуются определенными устойчивыми
фразами, выражающими согласие или несогласие с точкой зрения говорящего,
вежливыми клише для прерывания собеседника или, наоборот, для того, чтобы
показать, что его точка зрения представляет интерес. Эти фразы отличаются в
зависимости от формата беседы. Для начала рассмотрим формальный стиль
общения.
Формальное согласие, как правило, такими фразами:
• I completely (absolutely, totally) agree with you
• I couldn’t agree more
• Exactly…, absolutely…
• There is nothing more to add to this
• This is perfectly true…
67
Эти фразы используются, если вы абсолютно согласны и вам больше нечего
добавить, однако в случае, если вы хотите добавить еще что-то, то можно
сказать:
• Well, I agree with you on the whole, but …
• I agree in principle with you that…; however…
• I can agree with you to a certain extent but …
• You definitely have the point here but I’d like to add that …
• I take your point, however it seems to me that …
• It is certainly reasonable, however …
Если же беседа носит неформальный характер, то свое согласие можно
выразить следующими фразами:
• I’m with you on this point
• I couldn’t agree more
• Yes, absolutely (of course)
• Sure
• There’s no doubt about it
• You bet!
• I think so too
• I couldn’t have put it better myself
• Great minds think alike
• You took the words right out of my mouth.
Последние три фразы имеют еще и функцию похвалы, чтобы вдохновить
собеседника на дальнейшие рассуждения.
Несогласие в формальном языке, как правило, носит нейтральный характер,
чтобы не обидеть собеседника и не привести беседу к конфликту:
• Do you really think so?
• I can’t say I share your view on this…
• I feel I must disagree…
• I respect your opinion of course, but on the other hand…
• I wouldn’t say that, really.
• Well, taking your point into consideration, I therefore must admit that …
• Taking your point I still can’t help feeling that…
• I’m afraid, I disagree with you …
• I’m afraid I don’t see it this way …
• To tell you the truth I have a different opinion.
68
В неформальном общении присутствует большая эмоциональность, поэтому,
если вы абсолютно не согласны с утверждением своих друзей, вы можете
сказать:
• Rubbish! Nonsense!
• You can’t be serious! You must be joking (kidding)!
• No, no, it’s not right.
• I disagree with you completely.
• Come off it!
Для более мягкого несогласия используются фразы:
• I’m not sure you’re right
• I’m not sure about that ….
• I agree up to a point but …
• You could be right but …
• But I thought …
• Yes, but …
• That’s not how I see it ….
• That’s another pair of shoes…
Для того, чтобы показать, что мы заинтересованы в том, о чем говорит наш
собеседник, можно вставлять в ваш разговор такие фразы и слова, как:
• Really?
• Amazing!
• How interesting!
• That’s a good idea…
• Is it right?
• Unbelievable!
И менее формальные:
• Cool!
• Awesome!
• Wow! Great!
• You’re kidding! (если вас удивило какое-то высказывание)
Для ведения разговора иногда просто необходимо перебить собеседника, чтобы
дополнить что-то или высказать свою точку зрения. Это можно сделать
следующим образом:
• Sorry to interrupt you but ….
• Sorry for the interruption but …
• Pardon me / Excuse me …
• I hate to interrupt you but …
• Yes, but if I can interrupt you…
Для более неформального варианта подойдут фразы:
• Hang on!
69
• Hang on a minute!
• One moment!
• What?
Разговорные клише необходимы для общения, так как большинство ситуаций
требуют использования клишированных фраз, а не отдельных слов, связанных
между собой грамматикой. Именно поэтому обогащение своего словарного
запаса устойчивыми фразами позволит вам чувствовать себя непринужденно в
любой ситуации.
The task – study the necessary information in order to get prepared for the following
discussion:
There are some topics:
“Russian technological development will be increased with the help of the youth”
“People will always prefer the books in the printed format”
“Mobile phones will totally replace TV and radio”
70
Список литературы
1. Сase Study: Методические указания и задания по организации
самостоятельной работы студентов I – II курсов экономических
специальностей / Сост. О.В. Гофман, Т.М. Кобенко. – Нижневартовск:
Изд-во Нижневартовского гуманитарного университета, 2008. – 28 с.
2. Федорова М.А. От академического письма – к научному выступлению.
Английский язык: учеб. пособие / М.А. Федорова. – М.: ФЛИНТА: Наука,
2015. – 168 с.
3. Webwise [Электронный ресурс] URL: https://performdigi.com/report-
writing-examples/ (дата обращения: 12.09.2023).
4. Webwise [Электронный ресурс] URL:
https://enghelp4students.wordpress.com/2022/11/23/how_to_write_an_abstract
/ (дата обращения: 09.10.2023).
5. Webwise [Электронный ресурс] URL: https://xn--
80aegcaa6cbngm5a6c1ci.xn--p1ai/kak-napisat-annotaciyu-na-anglijskom/
(дата обращения: 13.09.2023).
6. Webwise [Электронный ресурс] URL: https://sibac.info/blog/kak-pravilno-
pisat-annotaciyu-na-angliyskom-yazyke (дата обращения: 24.08.2023).
7. Webwise [Электронный ресурс] URL:
https://www.sciencedaily.com/releases/2021/06/210622123303.htm (дата
обращения: 15.11.2023).
8. Webwise [Электронный ресурс] URL:
https://www.learnathome.ru/blog/usefull-phrases (дата обращения:
22.11.2023).
9. Webwise [Электронный ресурс] URL:
https://www.sciencedaily.com/releases/2023/09/230911191001.htm (дата
обращения: 26.10.2023).
10. Webwise [Электронный ресурс] URL:
https://www.sciencedaily.com/releases/2023/10/231009191623.htm (дата
обращения: 27.10.2023).
11. Webwise [Электронный ресурс] URL: https://spectrum.ieee.org/helicopters-
go-electric (дата обращения: 04.11.2023).
12. Webwise [Электронный ресурс] URL: https://spectrum.ieee.org/road-safety-
5g (дата обращения: 05.11.2023).
13. Webwise [электронный ресурс] URL: https://proza.ru/2007/01/15-51 (дата
обращения: 25.10.2023).
14. Webwise [Электронный ресурс] URL: http://Language Acquisition News --
ScienceDaily (дата обращения: 25.10.2023).
15. Webwise [Электронный ресурс] URL:
https://www.sciencedaily.com/releases/2023/11/231113192036.htm (дата
обращения: 25.10.2023).
Соседние файлы в предмете [НЕСОРТИРОВАННОЕ]