Machine Learning: An Overview Pt.1
机器学习(ML)是一个吸引了大量兴趣的新兴领域, but is not well understood. To improve this understanding, 这篇博文以“常见问题”的形式概述了机器学习的原理和应用.
What is machine learning?
机器学习是人工智能(AI)的一种应用，计算机在没有明确编程的情况下人工地“学习”和“解决问题”. 机器学习的主要好处是提供更大的计算和预测能力. 机器学习能够处理大量数据，而不会有人为错误的风险. 机器学习还可以在人类使用传统方法难以完成的范围内进行预测.
One example is IBM’s Watson for Oncology which uses ML to “study” a patient’s medical record. 基于各种背景资料和医疗记录, 机器学习算法识别医生需要考虑的关键信息, and offers potential treatment options.
How does machine learning work?
机器学习通过获取大量数据并使用算法对其进行分类来“学习”, look for patterns, and find relationships. 然后，机器学习将这些“知识”应用于用户指示的任务. 预测只是用户可以要求ML执行的任务之一.
ML算法代表一个过程或一组要遵循的规则，具有高度的灵活性. 这意味着它们可以应用于各种各样的数据集、主题和问题. However, 以使机器学习算法提供足够的用户价值, the algorithm may require more data or sources to “study”. This process can be iterative and time-consuming, 在考虑使用机器学习时必须考虑哪些因素.
(Further reading: A Tour of Machine Learning Algorithms)
What are some examples of machine learning in practice?
Vehicle classification at toll gates:
- 这个例子是由Michael sciocca和系统开发团队领导的IBI Group项目.
- At toll gates in India, 机器学习算法分析激光扫描仪和感应回路数据，根据高度对车辆进行分类, axle, and chassis characteristics to charge the correct vehicle toll.
- 与之前的分类方法相比，ML将收费分配准确率从84%提高到97%.
Detecting infrastructure defects based on drone footage:
- Drones capture thermal camera footage of concrete bridges. 然后将其输入ML算法，该算法用于开发地下分层的热图.
- 机器学习被认为是一种有效的、非破坏性的检查桥面的方法.
- Further reading: Concrete Bridge Deck Maintenance
Click here to view larger image.
Analyzing traffic conditions based on drone footage:
- 移动传感平台用于识别和分类无人机拍摄的车辆.
- Vehicle position, 速度和加速度用于执行与交通计数相关的各种任务, travel time, speed, delay analysis, lane utilization, emissions, and more.
- 多亏了机器学习，至少96%的车辆被正确识别和分类.
- Further reading: Data From Sky
利用x射线图像预测癌细胞的存在:
- 而低剂量计算机断层扫描(LDCT)和计算机断层扫描(CT)扫描提供比x射线图像更多的信息, access to this technology is limited in rural areas.
- 机器学习被用来制作热图，以筛查胸部x光片中肺部癌细胞的存在.
- 该算法在大约75%的情况下能够识别出癌细胞的存在.
Further reading: 基于深度学习方法的胸部x线图像肺癌自动预测
What are some of the barriers to adopting machine learning?
由于它仍然是一个新兴领域，因此在考虑ML时必须谨慎. Barriers to adoption include:
- Management of large quantities of data
- Unknown technological capabilities
- Potential impacts on employment
- Data privacy
- Deciding on the type of ML algorithm to be used
This article is based on the findings of “Machine Learning: An Overview“, an internal Pocket R&D report by Andrew Wong, Anthony Lui and Cameron Berko. Pocket R&D是IBI的战术微观研究计划，它利用我们人才库的知识来告知我们如何定义未来的城市.
