ecc工作原理:改善所有用户的反馈

I’m Meg Risdal, the new Product Manager for Public Q&A. My team works on the public platform and the community tools that help power it (in contrast to our private Q&A products, Stack Overflow for Teams). Recently, we’ve shipped a new tag synonyms dashboard, a redesign of the mod flags dashboard, and Custom Filters. And now, over the coming months, we will be focused on a small number of new initiatives based on the strategy Sara Chipps, our Director of Public Q&A, shared in her blog post “What a very bad day at work taught me about building Stack Overflow’s community”.

Before I dive into specifics about what we’re working on, I’ll tell you a bit more about myself. I came to Stack Overflow specifically to work on public Q&A. I’m passionate about open access to high quality knowledge that helps developers successfully learn and do their jobs. I come to Stack Overflow from Kaggle, a Google company and online data science community, so I appreciate how special it is to be able to learn and work so closely with users in developing products.

As someone who’s worked on a public community product, I’ve always valued transparency with users. For this reason, I’m thrilled to tell you more about what we’re currently focused on and why.

我们正在努力

Last year, we kicked off an initiative to make Stack Overflow more welcoming. We launched a new Code of Conduct, an improved question asking experience, and improvements to the moderator dashboard. We spoke to many of our users about their personal experiences using Stack Overflow, and through these conversations, we heard that the limitations of the Q&A system continue to create the very environment we set out to discourage. People who need help with coding problems feel attacked when their questions are closed or downvoted, while those curating site content feel blamed for doing what the system has asked them to do.

It’s simply not enough to ask people to be nice or change their behavior when the software that underlies everyone’s interactions doesn’t facilitate this. That’s why we’re looking at ways to revitalize the Q&A system, starting with the way questions are handled. We’re working on changes to reduce friction among all users: making embarrassing, public feedback less harsh and more helpful for everyone, reducing the many burdens of our outdated commenting system on questions, and applying what we’ve learned from our new Ask Question Wizard to help all users ask successful questions.

改善反馈

As Sara noted in her recent blog post, it’s hard not to take feedback personally when it’s piled on, no matter how constructive it is. Unfortunately, this overwhelming pile-on is exactly what users who ask an imperfect question are confronted with. And asking questions that other people can readily answer on the first try is hard enough as is. Not only is it a learned skill to research and ask a well formulated programming question as a novice (yes, I approve of putting “Googling” on your resume), there’s lots of intimidating written and unwritten rules to grapple with. This is why we’re prioritizing improving ways our system facilitates feedback.

The first thing we’re tackling is post notices. If you’ve come across a duplicate question or closed question, you’ve probably seen a post notice. These are the pale yellow informational banners that sometimes appear on questions. For people who ask questions today, if your question is closed, feedback that is directed toward you privately is shared publicly with anyone who views your question. Plus, the names of people who voted to close the question are highlighted publicly, too, setting them up for attack when they’re just trying to curate content according to the system.

Here’s what our holistic redesign of all post notices will prioritize:

  • 向帖子作者提供更好的私人反馈
  • 不要让那些负责内容管理的用户当场
  • 为绝大多数公众观众提供可行的,可理解的信息
The graphic below compares an example of a current “on hold” post notice and what it could look like if we provide tailored feedback for authors versus the public view:
A comparison of a current post notice with an example new design.
While working on building post notices, we’re also doing discovery on ways to make our question close workflows and review queues better facilitate feedback and content curation for seasoned moderators, technology experts, and new question-askers alike. We’ve already heard lots of ideas from our moderators and curators who use the review queues to know that there are immediate improvements we can consider making. This will supplement the qualitative and quantitative research we do to identify near-term improvements. Finally, we will be undertaking this work with an eye towards long-term changes to ensure sustainability.

“他们告诉我不要阅读评论……”

On Stack Overflow, comments were originally intended to be used to facilitate clarifying questions in order to improve the quality of the question (both as a long-term artifact and so it can be answered accurately). However, it probably comes as a surprise to no one that the comments p on Stack Overflow doesn’t always serve this purpose in practice. Comments can be distracting, outright harmful, pure spam, and everything in between.

We encourage people to flag content that doesn’t belong, and our moderators do incredible work to review everything. But there’s a lot of opportunity to reduce the burden for moderators and users who flag inappropriate content in ways that make using Stack Overflow more pleasant for everyone.

For example, some of the things we’re exploring right now:

  • 更好地区分“答案”和“评论”操作(由于大量有用的“未回答”标志指示界面有问题)。
  • 如何减少无用评论的数量 首先可能会被删除。
There’s a lot we could do to make commenting and interacting with questions more useful and intuitive for everyone. If this is something that interests you, share your feedback with us! Join our research list (you’ll need a Stack Overflow account) for opportunities to participate in UX research.

改善了所有人的提问方式

Several months ago, we launched the Ask Question Wizard on Stack Overflow. The new experience defaults users with low reputation to use a guided mode to formulate their questions. So far, we’re pleased with how it’s helping users ask better questions and as a result have a better experience on the site. In fact, we’re so pleased that our next iteration is focused on incorporating aspects of the guided mode into the question-asking form for everyone.

We’re still in early stages, but some of the changes we’re working on so far include:

  • 初次提问者的前期指导
  • 设置对提出问题后会发生什么的期望
  • 在起草问题时改进了“如何提出建议”的指导
  • 通过将数十条验证消息整合到单个“审阅”界面中,可以更轻松地提高问题质量。

We’re excited to begin working on these changes because it will make it far easier for us to extend the improvements across the rest of the Stack Exchange network sites as well as for our Stack Overflow for Teams private Q&A products. It will also give us an opportunity to create a better user experience for everyone. More and more, we hope to look for ways to simplify our platform to seamlessly and intuitively meet the needs of all users who come to Stack Overflow to find and share knowledge.

Dr. Julia Silge, our data scientist at Stack Overflow, will be sharing more insights from how the Ask Question Wizard is doing very soon. Stay tuned!

下一步是什么

We know that millions and millions of users come to Stack Overflow each month and rely on and contribute to the site and community in so many unique ways. Stack Overflow wouldn’t be what it is without those contributions. This is why I’m excited for our team to focus on ways we can make it easier for everyone to participate.

We will share more updates on the blog as our work progresses. Subscribe to the blog to keep up with the latest on Stack Overflow.
from:https://stackoverflow.blog/2019/08/20/upcoming-on-stack-overflow/
1、资源项目源码均已通过严格测试验证,保证能够正常运行; 2、项目问题、技术讨论,可以给博主私信或留言,博主看到后会第一时间与您进行沟通; 3、本项目比较适合计算机领域相关的毕业设计课题、课程作业等使用,尤其对于人工智能、计算机科学与技术等相关专业,更为适合; 4、下载使用后,可先查看README.md或论文文件(如有),本项目仅用作交流学习参考,请切勿用于商业用途。 5、资源来自互联网采集,如有侵权,私聊博主删除。 6、可私信博主看论文后选择购买源代码。 1、资源项目源码均已通过严格测试验证,保证能够正常运行; 2、项目问题、技术讨论,可以给博主私信或留言,博主看到后会第一时间与您进行沟通; 3、本项目比较适合计算机领域相关的毕业设计课题、课程作业等使用,尤其对于人工智能、计算机科学与技术等相关专业,更为适合; 4、下载使用后,可先查看README.md或论文文件(如有),本项目仅用作交流学习参考,请切勿用于商业用途。 5、资源来自互联网采集,如有侵权,私聊博主删除。 6、可私信博主看论文后选择购买源代码。 1、资源项目源码均已通过严格测试验证,保证能够正常运行; 2、项目问题、技术讨论,可以给博主私信或留言,博主看到后会第一时间与您进行沟通; 3、本项目比较适合计算机领域相关的毕业设计课题、课程作业等使用,尤其对于人工智能、计算机科学与技术等相关专业,更为适合; 4、下载使用后,可先查看README.md或论文文件(如有),本项目仅用作交流学习参考,请切勿用于商业用途。 5、资源来自互联网采集,如有侵权,私聊博主删除。 6、可私信博主看论文后选择购买源代码。
应用背景为变电站电力巡检,基于YOLO v4算法模型对常见电力巡检目标进行检测,并充分利用Ascend310提供的DVPP等硬件支持能力来完成流媒体的传输、处理等任务,并对系统性能做出一定的优化。.zip深度学习是机器学习的一个子领域,它基于人工神经网络的研究,特别是利用多层次的神经网络来进行学习和模式识别。深度学习模型能够学习数据的高层次特征,这些特征对于图像和语音识别、自然语言处理、医学图像分析等应用至关重要。以下是深度学习的一些关键概念和组成部分: 1. **神经网络(Neural Networks)**:深度学习的基础是人工神经网络,它是由多个层组成的网络结构,包括输入层、隐藏层和输出层。每个层由多个神经元组成,神经元之间通过权重连接。 2. **前馈神经网络(Feedforward Neural Networks)**:这是最常见的神经网络类型,信息从输入层流向隐藏层,最终到达输出层。 3. **卷积神经网络(Convolutional Neural Networks, CNNs)**:这种网络特别适合处理具有网格结构的数据,如图像。它们使用卷积层来提取图像的特征。 4. **循环神经网络(Recurrent Neural Networks, RNNs)**:这种网络能够处理序列数据,如时间序列或自然语言,因为它们具有记忆功能,能够捕捉数据中的时间依赖性。 5. **长短期记忆网络(Long Short-Term Memory, LSTM)**:LSTM 是一种特殊的 RNN,它能够学习长期依赖关系,非常适合复杂的序列预测任务。 6. **生成对抗网络(Generative Adversarial Networks, GANs)**:由两个网络组成,一个生成器和一个判别器,它们相互竞争,生成器生成数据,判别器评估数据的真实性。 7. **深度学习框架**:如 TensorFlow、Keras、PyTorch 等,这些框架提供了构建、训练和部署深度学习模型的工具和库。 8. **激活函数(Activation Functions)**:如 ReLU、Sigmoid、Tanh 等,它们在神经网络中用于添加非线性,使得网络能够学习复杂的函数。 9. **损失函数(Loss Functions)**:用于评估模型的预测与真实值之间的差异,常见的损失函数包括均方误差(MSE)、交叉熵(Cross-Entropy)等。 10. **优化算法(Optimization Algorithms)**:如梯度下降(Gradient Descent)、随机梯度下降(SGD)、Adam 等,用于更新网络权重,以最小化损失函数。 11. **正则化(Regularization)**:技术如 Dropout、L1/L2 正则化等,用于防止模型过拟合。 12. **迁移学习(Transfer Learning)**:利用在一个任务上训练好的模型来提高另一个相关任务的性能。 深度学习在许多领域都取得了显著的成就,但它也面临着一些挑战,如对大量数据的依赖、模型的解释性差、计算资源消耗大等。研究人员正在不断探索新的方法来解决这些问题。
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