L5 U2 做决定

Module 1 比较选择

1. 使用 as...as

比较选择

您或许熟悉两种比较方法。一种是 than 搭配 more + 形容词,还有一种是像 bigger (更大)和 better (更好)这样的单词。
Bangkok is more expensive than New Delhi.
Spain is a better place for the conference.

我们也可以使用 as + 形容词 + as 的结构来进行比较。
New Delhi isn't as expensive as Bangkok.
New York isn't as far as Los Angeles.

这一短语常用于否定,它有多种构成方式。
Berlin isn't as exciting as London.
Berlin's not as exciting as London.

Paris is more expensive than Berlin.

2. 介绍观点

介绍观点的有用短语

我们在介绍观点时,有许多词汇和短语可用于告诉他人我们要说的事情。在开始新的话题,So 就非常有用。
So, the conference is next month.
So, we're having a meeting later today.

要提出观点,我们可以使用像 For me 这样的短语。
For me, Rio is a better location.

在表示同意或部分同意时,我们也可以使用 True 和 Yes, but 这种有用的词汇和短语。
A: London's not as hot as Paris in the summer. 伦敦夏天没有巴黎热。
B: True.

A: London's not as hot as Paris in the summer. 伦敦夏天没有巴黎热。
B: Yes, but Paris isn't as expensive.

像 Well 和 Hmm 等词可以给我们时间思考接下来要说的话。
A: Which place is better? 哪个地方更好?
B: Well, Moscow is very nice.

A: Which location is cheaper? 哪个地方更便宜?
B: Hmm. I'm not sure.

Yes, but London isn't as near as Tokyo.
True. Maybe Rio is better.
Rio isn't as expensive as New York.
London is not as exciting as New York.
For me, I prefer New York.

Module 2 做决定

1. 最高级

the nearest
the cheapest
the best
the farthest
the most expensive
最高级
我们知道您掌握了使用 as … as 和 more than 比较两个事物的方法。当您想表达一个事物比所有其他事物都好时,可以使用最高级。
Tokyo is the best choice.
Rio is the cheapest city.

有的比较级和最高级构成方式规则,分别为 + er 和 the + 形容词 + est。
cheap – cheaper – the cheapest
far – farther – the farthest
expensive – more expensive – the most expensive

其他的比较级和最高级构成方式不规则,必须背诵识记。
good – better – the best

2. 表示异议

表示异议的不同方法
表达异议时,我们应该礼貌地提出。例如,我们可以提出疑问,而不是直接加以陈述。
A: I think LA's the best city. 我认为洛杉矶是最好的城市。
B: Really? For me, it's not as exciting as New York!
A: Paris is so boring. 巴黎非常无聊。
B: Do you really think so?

表示异议前,我们可以使用一个短语,提醒对方即将要说的话。
A: I think Rio is the best location. 我认为里约热内卢是最好的地方。
B: To be honest, I think Tokyo's better.
A: I love New York. 我非常喜欢纽约。
B: To be honest with you, I'm not a fan.
坦白跟你说,我不喜欢纽约。

OK. Barcelona, then.好,那就巴塞罗那。
To be honest, New York isn't the best choice.
Rio is the cheapest, but Paris is the nearest.
Really? I don't think Paris is the nearest.
I think Milan is the best option.
The most exciting city is Barcelona.

3. 理解主要思想

听取主要思想

When you are listening to someone, the most important thing to understand is the person's main idea or decision. Look for the phrases you've just learned to indicate what the speaker's meaning is. For example, when you hear the phrase to be honest, you should expect disagreement afterward. When you hear I think that, you should listen for the speaker's opinion afterward. Listen to the speaker below. What venue does the speaker think is best?

Our discussion today was excellent! Thanks for all of your ideas. I've been thinking over what everyone said, and I've made my decision. To be honest, New York City is just too expensive. We don't have the budget this year. Paris is too far. Sorry, Marta! I think that Atlanta is the best place for our conference. It's the closest and much less expensive.
即使你不理解很多单词,你能理解最后的选择是Atlanta。

Module 3 讨论细节

1. 会议语言

projector 投影仪a machine that extends an image or film onto a screen or wall.
We need meeting rooms with projectors.

presentation 展示an activity that involves someone giving information to a group of people, often by showing or describing
We'll have three presentations this year at the conference.

facilities 设施locations or things that are built for a specific function
What facilities does the hotel have?

venue 场所 a location where an event happens
This is a great venue for the conference.

team building 团队建设 an activity designed to encourage teamwork and improve the relationships of coworkers
Let's have a team-building event on the last day.

department 部门 a section of a company
Each department should give a presentation.

meeting 会议 a gathering of a group of people for a specific purpose
On day one, let's have some meetings.

2. 建议和推荐

提出建议和推荐
在组织会议等活动时,动词 recommend(推荐)和 suggest(建议)非常有用。
A: So, first, about the venue. What would you suggest? 那先说说场地。你有什么建议?
B: I'd recommend the Lowatt Hotel. It's the best in Rio, and it has the most
modern conference facilities, too.

Module 4 总结决定

1. 阅读具体信息

寻找关键词
阅读英语看起来似乎很难,但好消息是您不需要每个词都认识,只要认识重要的词就可以了。如果您想知道今年在哪里开会,哪些关键词有助于理解这个句子?
As I mentioned in Monday's meeting, we're going to Moscow for the conference this year!

您或许只需要 going、Moscow 和 conference 三个关键词。如果您知道自己要找什么信息,就用不着每个单词都看了。

2. 报告并指代决定

提及说过的事情
在会上做出决定后,使用类似短语回顾您说过的事情。这些短语在商务写作中尤为有用。
As I mentioned in Monday's meeting, we're going to Moscow for the conference this year!
如我在周一会上所提到的,我们今年打算去莫斯科开会!
As discussed in the meeting, there will be presentations from all departments.
如会上所讨论的,所有部门都会做演示。

宣布决定
使用 decide、choose 和 agree 等动词宣布已经做出的重要决定。

We've decided to book three rooms.
我们决定预订三个房间。
We've chosen to have presentations on the first day.
我们决定第一天做演示。
We agreed to have the conference on the 5th.
我们决定把会议安排在五号。
如您看到的,所有这些动词都可接 to 后再接动词。

As I mentioned, we're going to Moscow.
We've chosen to have three presentations.
We've decided to book three rooms.
We've agreed to go on May 5th.
As discussed, there will be presentations.

  • 0
    点赞
  • 0
    收藏
    觉得还不错? 一键收藏
  • 0
    评论
1 目标检测的定义 目标检测(Object Detection)的任务是找出图像中所有感兴趣的目标(物体),确定它们的类别和位置,是计算机视觉领域的核心问题之一。由于各类物体有不同的外观、形状和姿态,加上成像时光照、遮挡等因素的干扰,目标检测一直是计算机视觉领域最具有挑战性的问题。 目标检测任务可分为两个关键的子任务,目标定位和目标分类。首先检测图像中目标的位置(目标定位),然后给出每个目标的具体类别(目标分类)。输出结果是一个边界框(称为Bounding-box,一般形式为(x1,y1,x2,y2),表示框的左上角坐标和右下角坐标),一个置信度分数(Confidence Score),表示边界框中是否包含检测对象的概率和各个类别的概率(首先得到类别概率,经过Softmax可得到类别标签)。 1.1 Two stage方法 目前主流的基于深度学习的目标检测算法主要分为两类:Two stage和One stage。Two stage方法将目标检测过程分为两个阶段。第一个阶段是 Region Proposal 生成阶段,主要用于生成潜在的目标候选框(Bounding-box proposals)。这个阶段通常使用卷积神经网络(CNN)从输入图像中提取特征,然后通过一些技巧(如选择性搜索)来生成候选框。第二个阶段是分类和位置精修阶段,将第一个阶段生成的候选框输入到另一个 CNN 中进行分类,并根据分类结果对候选框的位置进行微调。Two stage 方法的优点是准确度较高,缺点是速度相对较慢。 常见Tow stage目标检测算法有:R-CNN系列、SPPNet等。 1.2 One stage方法 One stage方法直接利用模型提取特征值,并利用这些特征值进行目标的分类和定位,不需要生成Region Proposal。这种方法的优点是速度快,因为省略了Region Proposal生成的过程。One stage方法的缺点是准确度相对较低,因为它没有对潜在的目标进行预先筛选。 常见的One stage目标检测算法有:YOLO系列、SSD系列和RetinaNet等。 2 常见名词解释 2.1 NMS(Non-Maximum Suppression) 目标检测模型一般会给出目标的多个预测边界框,对成百上千的预测边界框都进行调整肯定是不可行的,需要对这些结果先进行一个大体的挑选。NMS称为非极大值抑制,作用是从众多预测边界框中挑选出最具代表性的结果,这样可以加快算法效率,其主要流程如下: 设定一个置信度分数阈值,将置信度分数小于阈值的直接过滤掉 将剩下框的置信度分数从大到小排序,选中值最大的框 遍历其余的框,如果和当前框的重叠面积(IOU)大于设定的阈值(一般为0.7),就将框删除(超过设定阈值,认为两个框的里面的物体属于同一个类别) 从未处理的框中继续选一个置信度分数最大的,重复上述过程,直至所有框处理完毕 2.2 IoU(Intersection over Union) 定义了两个边界框的重叠度,当预测边界框和真实边界框差异很小时,或重叠度很大时,表示模型产生的预测边界框很准确。边界框A、B的IOU计算公式为: 2.3 mAP(mean Average Precision) mAP即均值平均精度,是评估目标检测模型效果的最重要指标,这个值介于0到1之间,且越大越好。mAP是AP(Average Precision)的平均值,那么首先需要了解AP的概念。想要了解AP的概念,还要首先了解目标检测中Precision和Recall的概念。 首先我们设置置信度阈值(Confidence Threshold)和IoU阈值(一般设置为0.5,也会衡量0.75以及0.9的mAP值): 当一个预测边界框被认为是True Positive(TP)时,需要同时满足下面三个条件: Confidence Score > Confidence Threshold 预测类别匹配真实值(Ground truth)的类别 预测边界框的IoU大于设定的IoU阈值 不满足条件2或条件3,则认为是False Positive(FP)。当对应同一个真值有多个预测结果时,只有最高置信度分数的预测结果被认为是True Positive,其余被认为是False Positive。 Precision和Recall的概念如下图所示: Precision表示TP与预测边界框数量的比值 Recall表示TP与真实边界框数量的比值 改变不同的置信度阈值,可以获得多组Precision和Recall,Recall放X轴,Precision放Y轴,可以画出一个Precision-Recall曲线,简称P-R
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值