Questõesde IF Sul - MG sobre Inglês

AI1 picks up racial and gender biases2 when learning language from text, researchers say. Without any supervision, a machine learning algorithm learns to associate female names more with family words than career words, and black names as being more unpleasant than white names.

For a study published today in Science, researchers tested the bias of a common AI model, and then matched the results against a well-known psychological test that measures bias in humans. The team replicated in the algorithm all the psychological biases they tested, according to a study from co-author Aylin Caliskan, a post-doc at Princeton University. Because machine learning algorithms are so common, influencing everything from translation to scanning names on resumes, this research shows that the biases are pervasive, too. 

An algorithm is a set of instructions that humans write to help computers learn. Think of it like a recipe, says Zachary Lipton, an AI researcher at UC San Diego who was not involved in the study. Because algorithms use existing materials — like books or text on the internet — it’s obvious that AI can pick up biases if the materials themselves are biased. (For example, Google Photos tagged black users as gorillas.) We’ve known for a while, for instance, that language algorithms learn to associate the word “man” with “professor” and the word “woman” with “assistant professor.” But this paper is interesting because it incorporates previous work done in psychology on human biases, Lipton says.

For today’s study, Caliskan’s team created a test that resembles the Implicit Association Test (IAT), which is commonly used in psychology to measure how biased people are (though there has been some controversy over its accuracy). In the IAT, subjects are presented with two images — say, a white man and a black man — and words like “pleasant” or “unpleasant.” The IAT calculates how quickly you match up “white man” and “pleasant” versus “black man” and “pleasant,” and vice versa. The idea is that the longer it takes you to match up two concepts, the more trouble you have associating them.

The test developed by the researchers also calculates bias, but instead of measuring “response time”, it measures the mathematical distance between two words. In other words, if there’s a bigger numerical distance between a black name and the concept of “pleasant” than a white name and “pleasant”, the model’s association between the two isn’t as strong. The further apart the words are, the less the algorithm associates them together.

Caliskan’s team then tested their method on one particular algorithm: Global Vectors for Word Representation (GLoVe) from Stanford University. GLoVe basically crawls the web to find data and learns associations between billions of words. The researchers found that, in GLoVe, female words are more associated with arts than with math or science, and black names are seen as more unpleasant than white names. That doesn’t mean there’s anything wrong with the AI system, per se, or how the AI is learning — there’s something wrong with the material.

Disponível em <http://www.theverge.com/>. Acesso em: 18/04/2017.

AI1 picks up racial and gender biases2 when learning language from text, researchers say. Without any supervision, a machine learning algorithm learns to associate female names more with family words than career words, and black names as being more unpleasant than white names.

For a study published today in Science, researchers tested the bias of a common AI model, and then matched the results against a well-known psychological test that measures bias in humans. The team replicated in the algorithm all the psychological biases they tested, according to a study from co-author Aylin Caliskan, a post-doc at Princeton University. Because machine learning algorithms are so common, influencing everything from translation to scanning names on resumes, this research shows that the biases are pervasive, too. 

An algorithm is a set of instructions that humans write to help computers learn. Think of it like a recipe, says Zachary Lipton, an AI researcher at UC San Diego who was not involved in the study. Because algorithms use existing materials — like books or text on the internet — it’s obvious that AI can pick up biases if the materials themselves are biased. (For example, Google Photos tagged black users as gorillas.) We’ve known for a while, for instance, that language algorithms learn to associate the word “man” with “professor” and the word “woman” with “assistant professor.” But this paper is interesting because it incorporates previous work done in psychology on human biases, Lipton says.

For today’s study, Caliskan’s team created a test that resembles the Implicit Association Test (IAT), which is commonly used in psychology to measure how biased people are (though there has been some controversy over its accuracy). In the IAT, subjects are presented with two images — say, a white man and a black man — and words like “pleasant” or “unpleasant.” The IAT calculates how quickly you match up “white man” and “pleasant” versus “black man” and “pleasant,” and vice versa. The idea is that the longer it takes you to match up two concepts, the more trouble you have associating them.

The test developed by the researchers also calculates bias, but instead of measuring “response time”, it measures the mathematical distance between two words. In other words, if there’s a bigger numerical distance between a black name and the concept of “pleasant” than a white name and “pleasant”, the model’s association between the two isn’t as strong. The further apart the words are, the less the algorithm associates them together.

Caliskan’s team then tested their method on one particular algorithm: Global Vectors for Word Representation (GLoVe) from Stanford University. GLoVe basically crawls the web to find data and learns associations between billions of words. The researchers found that, in GLoVe, female words are more associated with arts than with math or science, and black names are seen as more unpleasant than white names. That doesn’t mean there’s anything wrong with the AI system, per se, or how the AI is learning — there’s something wrong with the material.

Disponível em <http://www.theverge.com/>. Acesso em: 18/04/2017.

AI1 picks up racial and gender biases2 when learning language from text, researchers say. Without any supervision, a machine learning algorithm learns to associate female names more with family words than career words, and black names as being more unpleasant than white names.

For a study published today in Science, researchers tested the bias of a common AI model, and then matched the results against a well-known psychological test that measures bias in humans. The team replicated in the algorithm all the psychological biases they tested, according to a study from co-author Aylin Caliskan, a post-doc at Princeton University. Because machine learning algorithms are so common, influencing everything from translation to scanning names on resumes, this research shows that the biases are pervasive, too. 

An algorithm is a set of instructions that humans write to help computers learn. Think of it like a recipe, says Zachary Lipton, an AI researcher at UC San Diego who was not involved in the study. Because algorithms use existing materials — like books or text on the internet — it’s obvious that AI can pick up biases if the materials themselves are biased. (For example, Google Photos tagged black users as gorillas.) We’ve known for a while, for instance, that language algorithms learn to associate the word “man” with “professor” and the word “woman” with “assistant professor.” But this paper is interesting because it incorporates previous work done in psychology on human biases, Lipton says.

For today’s study, Caliskan’s team created a test that resembles the Implicit Association Test (IAT), which is commonly used in psychology to measure how biased people are (though there has been some controversy over its accuracy). In the IAT, subjects are presented with two images — say, a white man and a black man — and words like “pleasant” or “unpleasant.” The IAT calculates how quickly you match up “white man” and “pleasant” versus “black man” and “pleasant,” and vice versa. The idea is that the longer it takes you to match up two concepts, the more trouble you have associating them.

The test developed by the researchers also calculates bias, but instead of measuring “response time”, it measures the mathematical distance between two words. In other words, if there’s a bigger numerical distance between a black name and the concept of “pleasant” than a white name and “pleasant”, the model’s association between the two isn’t as strong. The further apart the words are, the less the algorithm associates them together.

Caliskan’s team then tested their method on one particular algorithm: Global Vectors for Word Representation (GLoVe) from Stanford University. GLoVe basically crawls the web to find data and learns associations between billions of words. The researchers found that, in GLoVe, female words are more associated with arts than with math or science, and black names are seen as more unpleasant than white names. That doesn’t mean there’s anything wrong with the AI system, per se, or how the AI is learning — there’s something wrong with the material.

Disponível em <http://www.theverge.com/>. Acesso em: 18/04/2017.

Brazilian police have arrested 12 people suspected of planning terrorist acts during next month's Rio Olympics, authorities said.

The group was inspired by ISIS and mostly organized online, Justice Minister Alexandre de Moraes said. He said no specific targets were mentioned, but the Justice Ministry is still investigating the suspects' computers and cell phones to learn more about the possible plans.

De Moraes said the suspects are all Brazilian nationals, and that one minor was mentioned in the conversations.

De Moraes said the group was not an organized cell, calling it "absolutely amateur - with no preparation." The group essentially said, "Let's start training in martial arts, let's start learning how to shoot," the justice minister said.

He noted the group tried to buy a gun online, which no organized cell would do.

Raffaello Pantucci, director of international security studies at the Royal United Services Institute, said there doesn't appear to be evidence of a sophisticated plot.

But Brazil has grappled with a host of threats against the Rio Olympics, now just 11 days away.

This week, Brazil's intelligence agency said it was reviewing all threats after a jihadi messaging channel called for its followers to target the Olympics, which start August 5.

"Many (threats) are discarded and the ones that deserve attention are investigated exhaustively", the agency said.

Earlier this week, a jihadi channel on the messaging app Telegram called for attacks against the games and detailed targets and methods, according to the SITE Intelligence Group.

But Brazil has vowed it will be ready to handle any terror attempt.

A Western diplomat said venues for the games have been "hardened significantly - and I believe the government of Brazil has done what it can to make it very difficult to get into the venues here."

Brazilian forces have been working with French SWAT teams to simulate attack scenarios. In one drill, Brazil special forces and a police dog chase down an armed gunman to thwart a possible attack on Rio's subway system.

"There is not a specific threat," Lt. Gen. Luiz Linhares of Brazil's Ministry of Defense said. "You have to screen for a great (spectrum) of threat."

http://edition.cnn.com/2016/07/21/americas/brazil-olympics-terror-arrests/ 

The National Institute of Economic and Social Research (NIESR) says the UK will go through a "marked economic slowdown" this year and next. It says inflation will also pick up, rising to 3% by the end of next year. "This is the short-term economic consequence of the vote to leave the EU", said Simon Kirby of the NIESR.

Overall the institute forecasts that the UK economy will probably grow by 1.7% this year but will expand by just 1% in 2017. This would see the UK avoid a technical recession, typically defined as two consecutive quarters of economic contraction.

Mr. Kirby argued that the June referendum vote had led to such financial and political uncertainty that this would bear directly on the spending and investment decisions of both businesses and households. "We expect the UK to experience a marked economic slowdown in the second half of this year and throughout 2017," he said. "There is an even chance of a 'technical' recession in the next 18 months, while there is an elevated risk of further deterioration in the near term."

The pick-up in inflation to 3% will mainly be due to the recent fall in the value of the pound, but that should be ignored by the Bank of England, the Institute said. "The Monetary Policy Committee should 'look through' this temporary rise in inflation and ease monetary policy substantially in the coming months", Mr. Kirby said. The institute forecasts that the Bank will reduce interest rates to just 0.1% eventually, after cutting them to 0.25% later this week.

In a separate report, the CBI business lobby group says that the UK's small and medium-sized manufacturers (SMEs) fear they will be hit by a fall in orders in the next three months. Its latest quarterly survey of SMEs says business optimism has fallen at its fastest rate since January 2009, when the UK economy was falling into recession. Now, the culprit is the uncertainty following June's Brexit vote. Despite this, the 472 firms surveyed said that current orders were stable.

Rain Newton-Smith, the CBI's director for economics, said: "The UK's SME manufacturers reported higher production, more staff hired and now expect to sell more of their world-class goods overseas over the next quarter, with a weaker sterling having a hand in this. "But overall they do feel less optimistic and are scaling back some investment plans in machinery and plants".

The CBI's survey is just the latest to suggest that the effect of the June referendum vote may be, in the short term at least, to depress business activity. On Monday the Markit/CIPS manufacturing purchasing managers' index suggested that activity among UK manufacturers in July had shrunk at its fastest pace for three years.

Meanwhile shoppers continue to benefit from falling prices in the UK's shops, stores and supermarkets. According to the latest survey from the trade body the British Retail Consortium, overall prices fell by 1.6% in the year to July. Food was 0.8% cheaper than a year ago and non-food items were 2.2% lower.

Adapted from http://www.bbc.com/news/business-36953247.

The National Institute of Economic and Social Research (NIESR) says the UK will go through a "marked economic slowdown" this year and next. It says inflation will also pick up, rising to 3% by the end of next year. "This is the short-term economic consequence of the vote to leave the EU", said Simon Kirby of the NIESR.

Overall the institute forecasts that the UK economy will probably grow by 1.7% this year but will expand by just 1% in 2017. This would see the UK avoid a technical recession, typically defined as two consecutive quarters of economic contraction.

Mr. Kirby argued that the June referendum vote had led to such financial and political uncertainty that this would bear directly on the spending and investment decisions of both businesses and households. "We expect the UK to experience a marked economic slowdown in the second half of this year and throughout 2017," he said. "There is an even chance of a 'technical' recession in the next 18 months, while there is an elevated risk of further deterioration in the near term."

The pick-up in inflation to 3% will mainly be due to the recent fall in the value of the pound, but that should be ignored by the Bank of England, the Institute said. "The Monetary Policy Committee should 'look through' this temporary rise in inflation and ease monetary policy substantially in the coming months", Mr. Kirby said. The institute forecasts that the Bank will reduce interest rates to just 0.1% eventually, after cutting them to 0.25% later this week.

In a separate report, the CBI business lobby group says that the UK's small and medium-sized manufacturers (SMEs) fear they will be hit by a fall in orders in the next three months. Its latest quarterly survey of SMEs says business optimism has fallen at its fastest rate since January 2009, when the UK economy was falling into recession. Now, the culprit is the uncertainty following June's Brexit vote. Despite this, the 472 firms surveyed said that current orders were stable.

Rain Newton-Smith, the CBI's director for economics, said: "The UK's SME manufacturers reported higher production, more staff hired and now expect to sell more of their world-class goods overseas over the next quarter, with a weaker sterling having a hand in this. "But overall they do feel less optimistic and are scaling back some investment plans in machinery and plants".

The CBI's survey is just the latest to suggest that the effect of the June referendum vote may be, in the short term at least, to depress business activity. On Monday the Markit/CIPS manufacturing purchasing managers' index suggested that activity among UK manufacturers in July had shrunk at its fastest pace for three years.

Meanwhile shoppers continue to benefit from falling prices in the UK's shops, stores and supermarkets. According to the latest survey from the trade body the British Retail Consortium, overall prices fell by 1.6% in the year to July. Food was 0.8% cheaper than a year ago and non-food items were 2.2% lower.

Adapted from http://www.bbc.com/news/business-36953247.

The National Institute of Economic and Social Research (NIESR) says the UK will go through a "marked economic slowdown" this year and next. It says inflation will also pick up, rising to 3% by the end of next year. "This is the short-term economic consequence of the vote to leave the EU", said Simon Kirby of the NIESR.

Overall the institute forecasts that the UK economy will probably grow by 1.7% this year but will expand by just 1% in 2017. This would see the UK avoid a technical recession, typically defined as two consecutive quarters of economic contraction.

Mr. Kirby argued that the June referendum vote had led to such financial and political uncertainty that this would bear directly on the spending and investment decisions of both businesses and households. "We expect the UK to experience a marked economic slowdown in the second half of this year and throughout 2017," he said. "There is an even chance of a 'technical' recession in the next 18 months, while there is an elevated risk of further deterioration in the near term."

The pick-up in inflation to 3% will mainly be due to the recent fall in the value of the pound, but that should be ignored by the Bank of England, the Institute said. "The Monetary Policy Committee should 'look through' this temporary rise in inflation and ease monetary policy substantially in the coming months", Mr. Kirby said. The institute forecasts that the Bank will reduce interest rates to just 0.1% eventually, after cutting them to 0.25% later this week.

In a separate report, the CBI business lobby group says that the UK's small and medium-sized manufacturers (SMEs) fear they will be hit by a fall in orders in the next three months. Its latest quarterly survey of SMEs says business optimism has fallen at its fastest rate since January 2009, when the UK economy was falling into recession. Now, the culprit is the uncertainty following June's Brexit vote. Despite this, the 472 firms surveyed said that current orders were stable.

Rain Newton-Smith, the CBI's director for economics, said: "The UK's SME manufacturers reported higher production, more staff hired and now expect to sell more of their world-class goods overseas over the next quarter, with a weaker sterling having a hand in this. "But overall they do feel less optimistic and are scaling back some investment plans in machinery and plants".

The CBI's survey is just the latest to suggest that the effect of the June referendum vote may be, in the short term at least, to depress business activity. On Monday the Markit/CIPS manufacturing purchasing managers' index suggested that activity among UK manufacturers in July had shrunk at its fastest pace for three years.

Meanwhile shoppers continue to benefit from falling prices in the UK's shops, stores and supermarkets. According to the latest survey from the trade body the British Retail Consortium, overall prices fell by 1.6% in the year to July. Food was 0.8% cheaper than a year ago and non-food items were 2.2% lower.

Adapted from http://www.bbc.com/news/business-36953247.

Brazilian police have arrested 12 people suspected of planning terrorist acts during next month's Rio Olympics, authorities said.

The group was inspired by ISIS and mostly organized online, Justice Minister Alexandre de Moraes said. He said no specific targets were mentioned, but the Justice Ministry is still investigating the suspects' computers and cell phones to learn more about the possible plans.

De Moraes said the suspects are all Brazilian nationals, and that one minor was mentioned in the conversations.

De Moraes said the group was not an organized cell, calling it "absolutely amateur - with no preparation." The group essentially said, "Let's start training in martial arts, let's start learning how to shoot," the justice minister said.

He noted the group tried to buy a gun online, which no organized cell would do.

Raffaello Pantucci, director of international security studies at the Royal United Services Institute, said there doesn't appear to be evidence of a sophisticated plot.

But Brazil has grappled with a host of threats against the Rio Olympics, now just 11 days away.

This week, Brazil's intelligence agency said it was reviewing all threats after a jihadi messaging channel called for its followers to target the Olympics, which start August 5.

"Many (threats) are discarded and the ones that deserve attention are investigated exhaustively", the agency said.

Earlier this week, a jihadi channel on the messaging app Telegram called for attacks against the games and detailed targets and methods, according to the SITE Intelligence Group.

But Brazil has vowed it will be ready to handle any terror attempt.

A Western diplomat said venues for the games have been "hardened significantly - and I believe the government of Brazil has done what it can to make it very difficult to get into the venues here."

Brazilian forces have been working with French SWAT teams to simulate attack scenarios. In one drill, Brazil special forces and a police dog chase down an armed gunman to thwart a possible attack on Rio's subway system.

"There is not a specific threat," Lt. Gen. Luiz Linhares of Brazil's Ministry of Defense said. "You have to screen for a great (spectrum) of threat."

http://edition.cnn.com/2016/07/21/americas/brazil-olympics-terror-arrests/ 

The World Health Organization (WHO) — the agency of the United Nations concerned with international public health — published a report in The Lancet Global Health on Sept. 5 that outlines which nations get the most (and the least) amount of exercise.

The survey of 168 nations ranked the countries by most to least active by measuring the exercise habits of a given percentage of each country’s population and then comparing that percentage with those samples of the other countries involved in the study. The WHO defines enough exercise as at least 75 minutes of vigorous activity or 150 minutes of moderately intense activity per week — or any combination of the two. The WHO analyzed statistics and trends across different economic backgrounds, and between genders.

Overall, there were only four counties in the world where more than 50 percent of the population did not get enough exercise: Kuwait, Iraq, American Samoa, and Saudi Arabia. So these four countries are effectively the “laziest” in the world. Ultimately the nation with the least amount of physical activity was Kuwait, with 67 percent of its adults not exercising enough.

Other nations towards the bottom of this list were the United States, which ranked 143rd out of 168 countries. A whopping 40 percent of the U.S.’s population doesn’t get enough exercise — meaning approximately 130 million Americans can’t manage to get 2.5 hours of moderate activity per week. The U.K. also ranked fairly inactive, with only 35.9 percent of their population getting the proper amount. Other more inactive countries included Brazil, with 47 percent, the Philippines with 39.7 percent, Singapore with 36.5 percent, and India with 34 percent of the population not getting enough exercise.

In Uganda, only about five percent of their sample population did not exercise enough. China also displayed high rates of activity, with just 14.1 percent of their sample population not getting sufficient exercise. Other fairly active countries included Mozambique, with just over five percent, as well as Myanmar, with around 10 percent of their population insufficiently active.

According to their survey results, women tended to get less exercise than men, with an overall eight percent difference between the two. The report states:

“In 159 of 168 countries, prevalence of insufficient physical activity was lower in men than in women, with a difference of at least 10 percentage points in 65 countries, and a difference of more than 20 percentage points in nine countries: Barbados, Bahamas, Saint Lucia, Palau, Iraq, Bangladesh, Trinidad and Tobago, Iran, and Saudi Arabia.” 

The organization also noted some interesting trends across different economic backgrounds. Generally, people in poorer countries tended to be more than twice as active as richer ones. The report explains that this trend could be related to the fact that those with higher incomes tend to have “more sedentary occupations” and that greater access to automobile transportation results in lesser physical activity.

After pooling all of the data together, the WHO found that one in four adults the world over does not get enough exercise — which is a pretty jarring statistic. “This puts more than 1.4 billion adults at risk of developing or exacerbating diseases linked to inactivity, and needs to be urgently addressed,” the report states.

The WHO had previously outlined a goal of lowering global inactivity by 10 percent by the year 2025, but considering the fact that these statistics have not differed much since 2001, the WHO estimates that their goal will likely not be met.

Disponível em: <https://allthatsinteresting.com/worlds-laziest-countries> Acesso em 10 set. 2018

The World Health Organization (WHO) — the agency of the United Nations concerned with international public health — published a report in The Lancet Global Health on Sept. 5 that outlines which nations get the most (and the least) amount of exercise.

The survey of 168 nations ranked the countries by most to least active by measuring the exercise habits of a given percentage of each country’s population and then comparing that percentage with those samples of the other countries involved in the study. The WHO defines enough exercise as at least 75 minutes of vigorous activity or 150 minutes of moderately intense activity per week — or any combination of the two. The WHO analyzed statistics and trends across different economic backgrounds, and between genders.

Overall, there were only four counties in the world where more than 50 percent of the population did not get enough exercise: Kuwait, Iraq, American Samoa, and Saudi Arabia. So these four countries are effectively the “laziest” in the world. Ultimately the nation with the least amount of physical activity was Kuwait, with 67 percent of its adults not exercising enough.

Other nations towards the bottom of this list were the United States, which ranked 143rd out of 168 countries. A whopping 40 percent of the U.S.’s population doesn’t get enough exercise — meaning approximately 130 million Americans can’t manage to get 2.5 hours of moderate activity per week. The U.K. also ranked fairly inactive, with only 35.9 percent of their population getting the proper amount. Other more inactive countries included Brazil, with 47 percent, the Philippines with 39.7 percent, Singapore with 36.5 percent, and India with 34 percent of the population not getting enough exercise.

In Uganda, only about five percent of their sample population did not exercise enough. China also displayed high rates of activity, with just 14.1 percent of their sample population not getting sufficient exercise. Other fairly active countries included Mozambique, with just over five percent, as well as Myanmar, with around 10 percent of their population insufficiently active.

According to their survey results, women tended to get less exercise than men, with an overall eight percent difference between the two. The report states:

“In 159 of 168 countries, prevalence of insufficient physical activity was lower in men than in women, with a difference of at least 10 percentage points in 65 countries, and a difference of more than 20 percentage points in nine countries: Barbados, Bahamas, Saint Lucia, Palau, Iraq, Bangladesh, Trinidad and Tobago, Iran, and Saudi Arabia.” 

The organization also noted some interesting trends across different economic backgrounds. Generally, people in poorer countries tended to be more than twice as active as richer ones. The report explains that this trend could be related to the fact that those with higher incomes tend to have “more sedentary occupations” and that greater access to automobile transportation results in lesser physical activity.

After pooling all of the data together, the WHO found that one in four adults the world over does not get enough exercise — which is a pretty jarring statistic. “This puts more than 1.4 billion adults at risk of developing or exacerbating diseases linked to inactivity, and needs to be urgently addressed,” the report states.

The WHO had previously outlined a goal of lowering global inactivity by 10 percent by the year 2025, but considering the fact that these statistics have not differed much since 2001, the WHO estimates that their goal will likely not be met.

Disponível em: <https://allthatsinteresting.com/worlds-laziest-countries> Acesso em 10 set. 2018

Building was not insured, the museum’s deputy director said, but some pieces survived including the Bendegó meteorite.

As much as 90% of the collection at Brazil’s National Museum was destroyed in a devastating fire on Sunday and – compounding the disaster – the building was not insured, according to the museum’s deputy director.

Some pieces survived, including the famous Bendegó meteorite and a library of 500,000 books – including works dating back to the days of the Portuguese empire – which was kept in a separate annex, Cristiana Serejo told reporters in front of the building’s blackened shell.

But it was still not possible to say how much of the collection had escaped the flames, Serejo said. “It could be 10%, it could be 15, it could be 20,” she said. “We had a very big loss.”

The museum’s Egyptology collection was completely destroyed, Serejo said.

Researchers who were able to enter one area of the building in Rio de Janeiro are starting to catalogue what little is left, said Serejo, who appealed to members of the public to return any items they found.

Asked if the museum was insured, she screwed up her face in mock anguish, and shook her head.

“I hope we learn from this,” she said. “Other public buildings are in the same situation.”

Disponível em: <https://www.theguardian.com/world/2018/sep/04/brazil-national-museum-fire-collection-destroyed-notinsured> Acesso em 07 set. 2018 (Adaptado)

Building was not insured, the museum’s deputy director said, but some pieces survived including the Bendegó meteorite.

As much as 90% of the collection at Brazil’s National Museum was destroyed in a devastating fire on Sunday and – compounding the disaster – the building was not insured, according to the museum’s deputy director.

Some pieces survived, including the famous Bendegó meteorite and a library of 500,000 books – including works dating back to the days of the Portuguese empire – which was kept in a separate annex, Cristiana Serejo told reporters in front of the building’s blackened shell.

But it was still not possible to say how much of the collection had escaped the flames, Serejo said. “It could be 10%, it could be 15, it could be 20,” she said. “We had a very big loss.”

The museum’s Egyptology collection was completely destroyed, Serejo said.

Researchers who were able to enter one area of the building in Rio de Janeiro are starting to catalogue what little is left, said Serejo, who appealed to members of the public to return any items they found.

Asked if the museum was insured, she screwed up her face in mock anguish, and shook her head.

“I hope we learn from this,” she said. “Other public buildings are in the same situation.”

Disponível em: <https://www.theguardian.com/world/2018/sep/04/brazil-national-museum-fire-collection-destroyed-notinsured> Acesso em 07 set. 2018 (Adaptado)