这篇论文主要是介绍如何利用swap-dominance efficient voting rules来解决，ethical decision making里面的dilemmas
感觉可以关注点是，如何将这个伦理道德的问题，建模解构成一个AI的问题，然后里面涉及的转化过来的数学语言也是值得去学习的。

abstract:

We present a general approach to automating ethical decisions, drawing on machine learning and computational social choice. In a nutshell, we propose to learn a model of societal preferences, and, when faced with a specific ethical dilemma at runtime, efficiently aggregate those preferences to identify a desirable choice. We provide a concrete algorithm that instantiates our approach; some of its crucial steps are informed by a new theory of swap-dominance efficient voting rules. Finally, we implement and evaluate a system for ethical decision making in the autonomous vehicle domain, using preference data collected from 1.3 million people through the Moral Machine website.

首先是这里的问题:

(列车问题: 一辆疾驰的车，突然刹车失灵了，然后只能往左或者往右打方向盘，往左打的话，会让3个行人不幸，往右打的话，会让一个运动员和他的狗失去生命)
在列车问题里面最主要的问题，是没有ground truth，什么是对，什么是错，已经被争论的了几千年了。

main idea:

We submit that decision making can, in fact, be automated, even in the absence of such ground-truth principles, by aggregating people’s opinions on ethical dilemmas.
就是说没有对错的标准，根据人们对这个道德困境观点进行聚集，然后在自动的去做判断。
然后有一个大佬，之前给出一个算法：computational social choice，给出了一些intuition。然后，这篇论文是在原来的基础上更进一步。

approach:

I data collection:

Datacollection:Ask human voters to compare pairs of alternatives (say a few dozen per voter). In the autonomous vehicle domain, an alternative is determined by a vector of features such as the number of victims and their gender, age, health — even species!

obtained from 1,303,778 voters, through the website Moral Machine

II Learning:

Use the pairwise comparisons to learn a model of the preferences of each voter over all possible alternatives.

III Summarization:

Combine the individual models into a single model, which approximately captures the collec- tive preferences of all voters over all possible alternatives.
这里的multi-model感觉就像10-701里面讲的boosting

IV Aggregation:

At runtime, when encountering an ethical dilemma involving a specific subset of alternatives, use the summary model to deduce the preferences of all voters over this particular subset, and apply a voting rule to aggregate these preferences into a collective decision.
In the autonomous vehicle domain, the selected alternative is the outcome that society (as represented by the voters whose preferences were elicited in Step I) views as the least catastrophic among the grim options the vehicle currently faces. Note that this step is only applied when all other options have been exhausted, i.e., all technical ways of avoiding the dilemma in the first place have failed, and all legal constraints that may dictate what to do have also failed.

I-IV: 涉及的理论是 random utility models (for alternatives to generate rankings over the alterna- tives)
在这个情况下，作者需要的是infinite的选择set以及有限的encountered set，然后这个过程叫做permutation processes

central question:
This means we can apply a voting rule in order to aggregate the preferences — but which voting rule should we apply?
And how can we compute the outcome efficiently?
solution:

new pipeline:
learning a permutation process for each voter (Step II); summarizing these individual processes into a single permutation process that satisfies the required swap-dominance property (Step III); and using any swap- dominance efficient voting rule, which is computationally efficient given the swap-dominance property (Step IV)
theory 部分