最近需要做一个关于自动驾驶汽车的英语演讲部分，因为是小组性质的，我负责自动驾驶汽车算法部分，着重强调了机器学习算法在其中发挥的作用和解决的问题。在此我假设自己是一位斯坦福大学的研究员来表述这段演讲。
Hello, everyone! Iam from Stanford University, major in intelligence algorithm. The algorithm usedin autonomous vehicles is important but complex. In autonomous cars, one of themain tasks of algorithms is to continuously render the surrounding environmentand predict the changes that may be caused to these environments. These tasks aredivided into 3 subtasks:
-
Objectdetection
-
Object identificationor object classified identification
-
Object locationand motion prediction
大家好！我来自斯坦福大学，主修智能算法方向，自动驾驶汽车所使用的算法很重要但是也很复杂。在自动驾驶汽车中，算法的主要任务之一是连续渲染周围环境，并预测可能对这些环境造成的变化。这些任务分为3个子任务：
-
物体检测
-
物体识别或物体分类识别
-
物体定位与运动预测
Then machinelearning algorithm could do these works well. A regression algorithm such aslinear regression or neural network regression can be used for object location as well as objectdetection or motion prediction.
机器学习算法能够很好地完成这些工作。回归算法比如线性回归或者神经网络回归能够用于物体定位以及物体检测或运动预测。
Before classifyingobjects, pattern recognition is an important step in the data sets which are obtainedby the sensors in the Advanced Driver Assistance System, and support vectormachine can manage the identification tasks by decision plane separating.
在物体分类之前，在高级辅助驾驶系统中由传感器收集到的数据集中，模式识别是很重要的一步。支持向量机能够通过决策平面分离来处理分类任务。
As a saying goes “Thereis no free lunch in the world” means that many algorithms may get the samedestination, but we should choose the better plan in the specific problem.
就像“天下没有免费的午餐”说的那样，许多算法都能够达到同一个目标，但是我们应该选择在特定情况下的那个更好的方案。
网友评论