ChaseDream

标题: 【揽瓜阁3.0】Day4 2020.07.23【社会科学-科技】 [打印本页]

作者: 妥妥 时间: 2020-7-22 22:24
标题: 【揽瓜阁3.0】Day4 2020.07.23【社会科学-科技】

揽瓜阁俱乐部第三期

Day4 2020.07.23

【社会科学-科技】

AI can predict which criminals may break laws again better than humans

（710字精读必做篇）

Computer algorithms can outperform people at predicting which criminals will get arrested again, a new study finds.

Risk-assessment algorithms that forecast future crimes often help judges and parole boards decide who stays behind bars. But these systems have come under fire for exhibiting racial biases , and some research has given reason to doubt that algorithms are any better at predicting arrests than humans are. One 2018 study that pitted human volunteers against the risk-assessment tool COMPAS found that people predicted criminal reoffence about as well as the software.

The new set of experiments confirms that humans predict repeat offenders about as well as algorithms when the people are given immediate feedback on the accuracy of their predications and when they are shown limited information about each criminal. But people are worse than computers when individuals don’t get feedback, or if they are shown more detailed criminal profiles.

In reality, judges and parole boards don’t get instant feedback either, and they usually have a lot of information to work with in making their decisions. So the study’s findings suggest that, under realistic prediction conditions, algorithms outmatch people at forecasting recidivism, researchers report online February 14 in Science Advances.

Computational social scientist Sharad Goel of Stanford University and colleagues started by mimicking the setup of the 2018 study. Online volunteers read short descriptions of 50 criminals — including features like sex, age and number of past arrests — and guessed whether each person was likely to be arrested for another crime within two years. After each round, volunteers were told whether they guessed correctly. As seen in 2018, people rivaled COMPAS’s performance: accurate about 65 percent of the time.

But in a slightly different version of this human vs. computer competition, Goel’s team found that COMPAS had an edge over people who did not receive feedback. In this experiment, participants had to predict which of 50 criminals would be arrested for violentcrimes, rather than just any crime.

With feedback, humans performed this task with 83 percent accuracy — close to COMPAS’ 89 percent. But without feedback, human accuracy fell to about 60 percent. That’s because people overestimated the risk of criminals committing violent crimes, despite being told that only 11 percent of the criminals in the dataset fell into this camp, the researchers say. The study did not investigate whether factors such as racial or economic biases contributed to that trend.

In a third variation of the experiment, risk-assessment algorithms showed an upper hand when given more detailed criminal profiles. This time, volunteers faced off against a risk-assessment tool dubbed LSI-R. That software could consider 10 more risk factors than COMPAS, including substance abuse, level of education and employment status. LSI-R and human volunteers rated criminals on a scale from very unlikely to very likely to reoffend.

When shown criminal profiles that included only a few risk factors, volunteers performed on par with LSI-R. But when shown more detailed criminal descriptions, LSI-R won out. The criminals with highest risk of getting arrested again, as ranked by people, included 57 percent of actual repeat offenders, whereas LSI-R’s list of most probable arrestees contained about 62 percent of actual reoffenders in the pool. In a similar task that involved predicting which criminals would not only get arrested, but re-incarcerated, humans’ highest-risk list contained 58 percent of actual reoffenders, compared with LSI-R’s 74 percent.

Computer scientist Hany Farid of the University of California, Berkeley, who worked on the 2018 study, is not surprised that algorithms eked out an advantage when volunteers didn’t get feedback and had more information to juggle. But just because algorithms outmatch untrained volunteers doesn’t mean their forecasts should automatically be trusted to make criminal justice decisions, he says.

Eighty percent accuracy might sound good, Farid says, but “you’ve got to ask yourself, if you’re wrong 20 percent of the time, are you willing to tolerate that?”

Since neither humans nor algorithms show amazing accuracy at predicting whether someone will commit a crime two years down the line, “should we be using [those forecasts] as a metric to determine whether somebody goes free?” Farid says. “My argument is no.”

Perhaps other questions, like how likely someone is to get a job or jump bail, should factor more heavily into criminal justice decisions, he suggests.

Source: Science News

【社会科学-科技】

Lab-Grown Human Mini Brains Show Brainy Activity

（432字 2分54秒精听必做篇）

先做精听再核对原文哦~

[attach]251572[/attach]

It’s not easy to study the early development of the human brain.

“The brain is very inaccessible, especially the early fetal stages. It’s just not ethical to study normal, healthy human brains.”

University of California, San Diego, biologist Alysson Muotri. He says researchers have instead relied on animal models.

“But the human brain is so much different from other species that we’re desperate to have, really, a human model so we can study the human brain.”

Now Muotri’s team may have that model, in the form of small globules of brain cells they’ve created in the lab. These pea-sized structures develop from stem cells that are bathed in a culture of nutrients, along with proteins that control gene activation. As the little structures grow, their constituents also specialize into different types of brain cells.

“And they will form connections, and these connections will form functional synapses that will, later on, turn into networks.”

After two months, the mini brains even begin to emit brain waves.

“And you can record every week to see how the activity has changed. And when they reach about six months of age, we see a growth exponentially in the number of connections and synapses that they can make.”

And at around 10 months, their brain activity compares to that of premature human infants.

“They’re pretty much following the same trajectory as the human brain does.”

That could make the mini brains very useful for understanding how our brains become wired early on. And they could also provide insights into the development of neurological conditions such as autism and epilepsy.

“These very early stages are exactly when some neurological conditions appear. And we have the possibility to help millions of people with neurological conditions.”

But Muotri also cautions that as the technology moves forward, ethical questions will start to emerge.

“Someone might ask, ‘Are they conscious or are they self-aware? Can they feel pain?’ I think we are in a gray zone, where this technology could evolve to something more complex. And then I think the ethical question would be, ‘What’s the moral status of these miniaturized brains?’”

Muotri says that same question has formed the basis for the rules and regulations governing the use of animals in the lab, which can serve as a model to guide the mini brain research. The findings are in the journal Cell Stem Cell.

In addition to shedding light on neurological development, mini brains could also help reveal how the human brain evolved and play a role in improving algorithms for artificial intelligence. These pea-sized brains may produce some big insights.

Source: Scientific American

【笔记格式要求】

精读笔记格式要求：

1.总结文章中心大意

2.总结分论点或每段段落大意

3.摘抄印象深刻或者觉得优美的句子

4.总结文章中的生词

5.记录阅读时间、总结时间、总时间

精听笔记格式要求：

1.逐句听写整篇文章

2.对照原文修改听写稿，标记出错原因

3.总结文章中心大意

4.总结精听过程中的生词

5.记录听写时间、总结时间、总时间

这里也给大家两点学习小建议哦~

精读：如遇到读不懂的复杂句，建议找出句子主干，分析句子成分，也可以尝试翻译句子来帮助理解~

精听：建议每句不要反复纠结听，如果听 5 遍都没听出来，那就跳过，等完成后再回听总结原因，时间宝贵，不要过于执着哦~

作者: 妥妥 时间: 2020-7-22 22:24

揽瓜阁俱乐部，自「language」一词谐音而来，是一个为帮助大家提升英语语言能力而建立的学习小团队。在这里，我们将定时发布涵盖各类话题的外刊语料，供大家练习听、读。同时还设置了严格的打卡机制，督促大家克服懒惰坚持学习。

同时我们也招募volunteer协助维护团队，确保学习活动顺利开展~大家一起营造积极向上的学习氛围~

想要提升英语能力的小伙伴，快快添加微信（theTOEFL）报名加入吧，大家一起观尽天下新鲜事，览遍四海热议瓜～

作者: Garfieldsu 时间: 2020-7-23 00:26
day 4 打卡

作者: shuzijun 时间: 2020-7-23 09:11
Day 4 打卡。
今天精听完毕iPad突然卡了，整篇文章消失，因此没有计算整理时间，只有精听时间。

作者: lucin 时间: 2020-7-23 10:10
Day4

作者: Dannibiubiubiu 时间: 2020-7-23 10:29
day4

作者: 龙驾马 时间: 2020-7-23 11:12
阅读笔记

中心大意

科学家正在利用人工智能对服刑犯人在假释或释放后，会否再次犯罪进行评测。研究发现，如果人类在未收到犯人的有效反馈情况下，做出的判断不如人工智能准确。

第一部分（总论 1-4段）人工智能对于服刑犯人是否会再次犯罪的预测准确率，可能要比人类做的更好，尤其是在法官和假释委员会没有获得足够的反馈信息的情况下（而实际上他们也经常不能收到足够的信息）

第二部分（展开 5-9段）描述了相关的实验细节，将计算机的预测，与人类在有收到反馈信息和没有收到反馈信息的情况下做对比，结果显示计算机的表现不输于人类。

第三部分（局限10-13段）人工智能80%的准确率，其实也意味着还有20%的错误概率，如果让人工智能的预测结果来替代人类的判断，至少目前来看是不合适的。

句子摘抄

The study did not investigate whether factors such as racial or economic biases contributed to that trend.

In a third variation of the experiment, risk-assessment algorithms showed an upper hand when given more detailed criminal profiles.

Since neither humans nor algorithms show amazing accuracy at predicting whether someone will commit a crime two years down the line, “should we be using [those forecasts] as a metric to determine whether somebody goes free?”

生词摘抄

parole n.假释
mimicking n.模仿
algorithms n.算法
rival v.竞争
recidivism n.再犯，累犯
violentcrime n.暴力犯罪

用时
阅读时间10分钟总结时间 30分钟共计40分钟

听力笔记

[attach]251863[/attach][attach]251864[/attach][attach]251865[/attach][attach]251866[/attach]

中心意思

科学家正在实验室中培训小型的脑细胞以进行对大脑的研究，这些吸猫可以增长成一个与人脑细胞接近的神经网络。这一成果对科学家研究人脑在早期是如何发育的非常有用，还有助于对癫痫症的治疗。当然，这些神经细胞是否会有疼痛感，是否会有意识，也成了一个值得注意的科研伦理问题。

生词摘抄

globule n.球状体，小滴
synaps n.（神经）突触
autism n.孤独症
epilepsy n.癫痫症
regulation n.规范，管理
neurological a.神经学的

用时
听写 30分钟总结 40分钟总计70分钟

作者: ppxstar 时间: 2020-7-23 18:08
Day4打卡

作者: ChloeSolo 时间: 2020-7-23 18:47
Day 4

作者: Amaranth- 时间: 2020-7-23 20:26
交作业！

作者: 殺G成功 时间: 2020-7-23 21:05
Day 4打卡

作者: IvyPei 时间: 2020-7-23 21:06
Day4打卡！ [url=]Day4阅读.pdf[/url][url=]Day 4听力.pdf[/url]

作者: Carlisler 时间: 2020-7-23 21:17
Day 4 打卡
[attach]251880[/attach][attach]251878[/attach][attach]251882[/attach][attach]251879[/attach][attach]251881[/attach][attach]251877[/attach]

作者: micahlan 时间: 2020-7-23 21:23
Day4打卡

[attach]251887[/attach]

作者: xxnmm 时间: 2020-7-23 21:23
D4

作者: LostSanta 时间: 2020-7-23 21:26
Language Club Day 4

作者: Dovis 时间: 2020-7-23 21:45
揽瓜阁 Day 4
Dovis 精听 Lab-Grown Human Mini Brains Show Brainy Activit
一文章大意
研究人员此前创建创建迷你大脑细胞模型，用于研究早期人类大脑的构造和神经系统发育。这项研究阐明神经系统的发展，帮助揭示人脑如何进化，并在改善人工智能算法中发挥作用，但也会带来一些伦理问题。
二生词摘录
fetal 胎儿的; 胎儿般的
pea-sized 豌豆大的
globules (液体或熔化了的固体的) 小滴，小球体
synapses (神经元的) 突触
exponentially 以指数方式
epilepsy 癫痫; 羊痫风; 羊角风
miniaturized 使微型化; 使成为缩
三用时记录
听写 31min 总结 17min 共计48.8min

作者: chealsea 时间: 2020-7-23 21:45
Day 4打卡

作者: 啊哒 时间: 2020-7-23 22:18
打卡

作者: wanglu1994143 时间: 2020-7-23 22:24
打卡

作者: Dovis 时间: 2020-7-23 22:25

精读 AI can predict which criminals may break laws again better than humans

一文章大意

科学家针对服刑犯人在假释或释放后，会否再次犯罪进行评测。通过人类志愿者和电脑算法分别进行预测，并对预测准确性进行了分析。发现人类在没有反馈和信息背景情况下，准确性不及电脑算法。

二段落总结

一项新的研究发现，电脑算法在预测哪些罪犯将再次被捕时可以胜过人类。
预测未来犯罪的风险评估算法因表现出种族偏见而遭受质疑。
这项新研究是关于人类如何预测再次犯罪中准确性与信息反馈之间的联系。
该研究结果表明，在现实的预测条件下，计算机算法在预测累犯性方面不及人类。
研究表明，在给出足够的罪犯信息之后，人类预测再次犯罪率的准确性可以和算法媲美。
在另一项实验版本发现，电脑算法在没有给出重犯反馈和背景信息的罪犯预测上具有优势。
在没有反馈的情况下，人类预测的准确度下降到60%。
在第三次变量实验中，如果给出更详细的犯罪概况，则风险评估算法将占上风。
当显示犯罪概况仅包含一些风险因素时，人类志愿者的表现与LSI-R相当。但是当显示更多详细的犯罪描述时，LSI-R胜了。
伯克利大学教授肯定了以上实验结果。
即使有80%的准确性，但也代表了20%的容错率。因此人类和算法的预测数据作为判断罪犯是否假释的判断仍然是不合适的。

三生词摘录

outperform (效益上) 超过，胜过

parole 假释; 有条件的释放; 言语

pitted 表面有小点(或小洞)的; 坑坑洼洼的; 去核的;

recidivism 惯犯; 但累犯; 再犯

mimick 模仿，摹拟

incarcerated 监禁; 关押; 禁闭

四句子摘抄

But these systems have come under fire for exhibiting racial biases , and some research has given reason to doubt that algorithms are any better at predicting arrests than humans are.

That’s because people overestimated the risk of criminals committing violent crimes, despite being told that only 11 percent of the criminals in the dataset fell into this camp

五用时记录

阅读10min 总结22.3min 共计33min

作者: rebeccaaaaaa 时间: 2020-7-23 22:51
DAY4

作者: cgvmc 时间: 2020-7-23 22:55
DAY4

作者: YLEmily 时间: 2020-7-23 22:56
DAY 4

作者: ling0602 时间: 2020-7-23 22:58
day4

作者: はな 时间: 2020-7-23 22:59
day4

作者: cynwis 时间: 2020-7-23 22:59
Day 4

作者: 猫小痕77 时间: 2020-7-23 23:06
day4

作者: ste蔡 时间: 2020-7-23 23:55
day4

作者: davidylxu 时间: 2020-7-24 00:02
day4

作者: Lingli_Here 时间: 2020-7-24 00:07
DAY 4:

精读笔记格式要求：

1. Ai technique helped human being to do the prediction of crime. The different model have different performance. And even with the more accuracy one,  we haven't plan to use that to do the criminal decisions

2. Computer algorithms can outperform people at predicting which criminals will get rested again.
2-1. Risk-assessment algorithms compared with human being.
2-2. Computational social scientist SG has found without feedback, humans performed better. Without feedback COMPAS perform better.
2-3. LSI-R and human volunteers compete again. LSI-R won out
2-4. Algorithms outmatch untrained volunteers doesn't mean their forecast should be trusted to justice decisions.

3&4:
parole boards 假释木板
reoffend 再次犯罪
eked out 弥补。。。不足

5:
10:00/ 25min/ 35min

精听笔记格式要求：

2&4
small globules 小水珠 of brain
These pea-sized structures develop from stem 梗 cells that are bathed in a culture of nutrients 营养, along with proteins 蛋白质 that control gene activation.
synapses 突触
trajectory 轨迹
neurological 神经系统
moral status 道德
miniaturize 使微型
autism and epilepsy 自我中心和癫痫
premature human infants  过早的人类婴儿
exponentially 呈几何级数
emit 散发 brain waves

3. Mini brains could also help reveal how the human brain evolved and play a role in improving algorithms for artificial intelligence.

5.
3min/ 20min/ 23min

作者: 木辛青 时间: 2020-7-24 08:24
打卡

欢迎光临 ChaseDream (https://forum.chasedream.com/)