Anonymous and Copy-Robust Delegations for Liquid Democracy

Liquid democracy with ranked delegations is a novel voting scheme that unites the practicability of representative democracy with the idealistic appeal of direct democracy: Every voter decides between casting their vote on a question at hand or delegating their voting weight to some other, trusted agent. Delegations are transitive, and since voters may end up in a delegation cycle, they are encouraged to indicate not only a single delegate, but a set of potential delegates and a ranking among them. Based on the delegation preferences of all voters, a delegation rule selects one representative per voter. Previous work has revealed a trade-off between two properties of delegation rules called anonymity and copy-robustness. To overcome this issue we study two fractional delegation rules: Mixed Borda branching, which generalizes a rule satisfying copy-robustness, and the random walk rule, which satisfies anonymity. Using the Markov chain tree theorem, we show that the two rules are in fact equivalent, and simultaneously satisfy generalized versions of the two properties. Combining the same theorem with Fulkerson's algorithm, we develop a polynomial-time algorithm for computing the outcome of the studied delegation rule. This algorithm is of independent interest, having applications in semi-supervised learning and graph theory.

Paper

Similar papers

Peer review

Reviewer PkVV7/10 · confidence 4/52023-07-04

Summary

The paper studies fractional allocation rules when each agent indicates a ranking over the agents that agrees to represent her, to overcome impossibility results of deterministic rules. The authors consider two different rules which are shown that are equivalent. They also provide a polynomial time algorithm for finding the outcome of one of the two rules.

Strengths

-Very well-written paper -Interesting model -Natural extension of the axioms to fractional allocations -Technically involved and interesting results -Use of known lemmas in interesting and elegant ways Further notes: Line 73: "casting voter" Line 134: I think you mean "|B|=|V(G)|-1"

Weaknesses

The choice of the top-rank priority axiom, while it is quite natural, seems a bit arbitrary. It would be nice to show what kind of other axioms would lead to the same impossibility result.

Questions

See weaknesses above

Rating

7: Accept: Technically solid paper, with high impact on at least one sub-area, or moderate-to-high impact on more than one areas, with good-to-excellent evaluation, resources, reproducibility, and no unaddressed ethical considerations.

Confidence

4: You are confident in your assessment, but not absolutely certain. It is unlikely, but not impossible, that you did not understand some parts of the submission or that you are unfamiliar with some pieces of related work.

Soundness

4 excellent

Presentation

4 excellent

Contribution

3 good

Limitations

The authors adequately addressed the limitations of their work.

Reviewer rLuU8/10 · confidence 3/52023-07-06

Summary

This paper primarily examines two equivalent fractional delegation rules for Liquid Democracy (transitive delegate voting): Mixed Borda Branching and Random Walk Rule. The main contribution of this paper is the equivalence between these two rules in the generalized setting of fractional delegations. This result is complemented by an axiomatic analysis of the rules w.r.t. notions of anonymity, copy-robustness, confluence, and top-rank priority.

Strengths

The equivalence between the two delegation rules is non-obvious and a significant contribution. Moreover, the proof of equivalence is highly non-trivial. These results in Sections 4-5 are of considerable value. But I am not sure if those alone are enough to carry the paper (see Weaknesses). Although perhaps they are.

Weaknesses

Unless I have misunderstood something, it doesn’t appear that the axiomatic analysis really solves, or resolves, any of the important issues discussed in the introduction. It circumvents them instead in a way that feels cheap. First, allowing fractional delegations clearly means confluence is no longer a constraint. The authors state essentially this in an incredibly oblique way in the proof of Theorem 8, “One can verify that every delegation rule that can be formalized via a Markov chain on the delegation graph (G,c) satisfies confluence.” Confluence requires that if v1 delegates to v2 that the remainder of their delegation path much be consistent. Suppose, for example, that v1 is the only voter delegating to v2, and v2’s delegation is split between v3 and v4. No we “send” ½ of v1’s vote and ½ of v2’s vote to v3, and the remaining ½ + ½ to v4. Of course, this is functionally equivalent to sending v1’s entire vote to v3 and v2’s entire vote to v4, violating confluence. In reality, the axiom of confluence wasn’t generalized, it is made irrelevant, because any delegation that violates confluence can be trivially converted to a fractional one that satisfies its relaxation. Whether we talk in terms of dividing vote tokens or probability distributions over delegations doesn’t matter, it’s the same principle. Similarly, simple path vs. walk makes no difference here. Second, the authors state that the purpose of giving delegations as orderings is to prevent delegation cycles and isolated voters. In footnote 3 on page 4, the authors admit that they handle isolated voters by ignoring them entirely and removing them from the graph. However it is not discussed how this bears on the axioms. Ignoring such voters is the same as assigning them a voting weight of zero, whereas if they cast their vote directly they get a voting weight of at least one (and possibly more if other isolated voters delegate to them). This violates copy-robustness, but this violation does not appear to be mentioned. Lastly, while this is a minor note, the authors should not say that they generalized the axioms, but rather that they relaxed the axioms. The authors generalized the class of delegation rules, and relaxed the axioms required. Edit: Based on the author(s) rebuttal, I have improved my understanding of the paper and its results.I no longer stand by my original criticism.

Questions

Q1: Am I correct in saying that ignoring the isolated voters ultimately violates copy-robustness? Q2: Am I correct that fractional delegation trivializes the axiom of confluence?

Rating

8: Strong Accept: Technically strong paper, with novel ideas, excellent impact on at least one area, or high-to-excellent impact on multiple areas, with excellent evaluation, resources, and reproducibility, and no unaddressed ethical considerations.

Confidence

3: You are fairly confident in your assessment. It is possible that you did not understand some parts of the submission or that you are unfamiliar with some pieces of related work. Math/other details were not carefully checked.

Soundness

4 excellent

Presentation

4 excellent

Contribution

4 excellent

Limitations

No further limitations.

Reviewer eBke7/10 · confidence 4/52023-07-06

Summary

The authors study liquid democracy with fractional (i.e., splittable) delegations. They extend previous work by Brill et al. on delegation rules that satisfy anonymity and copy-robustness, and demonstrate that two delegation rules (mixed Borda branching and the random walk rule) that were previously thought to be different and each satisfy one property are actually the same rule that satisfies (generalized versions of) both properties. Their algorithm for computing the outcome of the combined rule is also the first efficient algorithm for a problem in semi-supervised learning, the directed power watershed.

Strengths

+ The paper is well-written and easy to follow, and the problem they study is well-motivated in the context of liquid democracy. + The algorithm is nontrivial and of independent interest to other communities in computer science. + It's also nice to see that two previously-proposed rules are actually the same; it's extra satisfying that this rule happens to satisfy generalizations of two (really, four) axioms that were thought to be hard to simultaneously satisfy.

Weaknesses

- My main hesitation with this paper is the use of fractional delegations in liquid democracy. One central tenet of LD is the ability to immediately remove a delegation from someone who cast a vote in a way you did not approve of, which becomes much more difficult with splittable delegations. Additionally, if you allow agents to know exactly where portions of their vote ended up, they will potentially be given a lot of information about the rest of the delegation network. - Typos: line 73: casting, line 166: copy-robust, not copy-robustness, line 342: lens

Questions

1. Have you thought about ways of explaining the outcomes of MBB / RWR to users of liquid democracy systems? Are there intuitive ways of showing the flow of splittable votes through the network?

Rating

7: Accept: Technically solid paper, with high impact on at least one sub-area, or moderate-to-high impact on more than one areas, with good-to-excellent evaluation, resources, reproducibility, and no unaddressed ethical considerations.

Confidence

4: You are confident in your assessment, but not absolutely certain. It is unlikely, but not impossible, that you did not understand some parts of the submission or that you are unfamiliar with some pieces of related work.

Soundness

4 excellent

Presentation

4 excellent

Contribution

3 good

Limitations

Yes

Reviewer rGK46/10 · confidence 3/52023-07-07

Summary

The paper studies liquid democracy where voters can delegate their votes to others instead of casting them directly. Within this framework, some voters act as casting voters, while others delegate their votes. Delegation rules determine how casting voters are chosen for each set of delegating voters. However, voters have preferences regarding whom they trust more to represent their votes. When the entire vote must be delegated, it becomes challenging to satisfy all of several desirable axioms simultaneously. Instead, the authors explore fractional/ probabilistic delegations, allowing a vote to be split across casting voters. They demonstrate that by employing the random walk rule, it is possible to recover the possibility of meeting all axioms. Moreover, the authors note that the distribution over branchings, ensuring that every delegating voter is connected to a casting voter, follows a uniform distribution over all minimum-cost branchings. This connection has also been previously observed in the work of Fita Sanmartin et al. To compute the delegation distribution efficiently, the authors propose a polynomial time algorithm that builds upon Fulkerson's algorithm.

Strengths

- Fractional delegations are a natural extension and furthermore they are transparent in the sense that a voter can know what proportion of her vote was transferred to which casting voter. - The result on the possibility of confluence, anonymity and copy-robustness if we move to the realm of fractional delegations is nice to have. - The authors leverage results from graph theory and combinatorial optimization to obtain their results, hence introducing techniques from this literature to their community.

Weaknesses

- Equivalence of Mixed Borda Branching and Random Walk rule: Given an understanding of the Markov chain rule, the equivalence between the Random Walk rule and Mixed Borda Branching seems immediate. The authors' claim of surprise at this equivalence is somewhat perplexing, as the connection should have been anticipated with knowledge of the Markov chain rule. It would have been more understandable if other authors had defined Mixed Borda without recognizing this link, but the authors themselves acknowledge that this interpretation has already been observed by Rita Sanmartin et al. Consequently, the discussion on equivalence could be shortened by referring to Rita Sanmartin et al.'s work rather than presenting it as a surprising result. - Algorithm 2 and Directed Powershed: Unfortunately, I cannot confirm whether the Directed Powershed algorithm is novel to this paper. It appears that Sanmartin et al. discuss an extension of the undirected version to address the directed case (Page 5: "In section 5, we show how the Power Watershed can be generalized to directed graphs by means of the DProbWS.”, Page 8: "In the ProbWS paper [12], it was proven that the Power Watershed [6] is equivalent to applying the ProbWS restricted to the minimum cost spanning forests. This restriction corresponds to the case of a Gibbs distribution of minimal entropy over the forests. In this section, we will prove the analogous result for the DProbWS: When the entropy of the Gibbs distribution over the directed in-forests (3.1) is minimal, then DProbWS is restricted to the minimum cost spanning in-forests (mSF). This permits us to define a natural extension of the Power Watershed to directed graphs. “). So they claim to leverage existing knowledge (Power Watershed: A Unifying Graph-Based Optimization Framework, 2010). However, if the problem remains unsolved, a discussion of the Power Watershed algorithm and its limitations in extending to the directed case would be necessary. It is worth noting that Algorithm 2 has limitations in practicality, with a runtime of O(n^7) (up to log factors), while Power Watershed may offer better efficiency. - The anonymity gained through randomizing over mincost branchings seems not surprising,. It is unclear where the challenges lie in proving that the other properties still hold ( as the authors seem to demonstrate more general versions of the axioms). It would be valuable to have an in-depth discussion of the supervised learning literature, as the problem of liquid democracy appears to be studied under different names with distinct requirements in various fields. Exploring the Power Watershed algorithm and its directed extension, along with providing a comprehensive explanation of the connections between different fields, would be a very useful contribution. This could introduce the supervised learning literature to the computational social choice community, while also allowing the supervised learning community to benefit from the axiomatic analysis conducted in this work. If the extension of power watershed algorithm to the directed from the literature turns out to be faulty or very unclear, I will increase my score.

Questions

- Can you provide an explanation for why Fita Sanmartin et al.'s claim about extending the Power watershed algorithm is incorrect or why it cannot be straightforwardly extended? How does the runtime of your algorithm compare to that of Power Watershed? - What makes the Equivalence of Mixed Borda and Random Walk Rule surprising?

Rating

6: Weak Accept: Technically solid, moderate-to-high impact paper, with no major concerns with respect to evaluation, resources, reproducibility, ethical considerations.

Confidence

3: You are fairly confident in your assessment. It is possible that you did not understand some parts of the submission or that you are unfamiliar with some pieces of related work. Math/other details were not carefully checked.

Soundness

3 good

Presentation

3 good

Contribution

3 good

Limitations

No concerns, limitations are addressed in the appendix.

Reviewer vsDd6/10 · confidence 3/52023-07-13

Summary

This paper studies algorithms for assigning delegation weights in liquid democracy in the setting where participants indicate a set of trusted delegates, along with a ranking describing their preferences over these delegates. Notably, the algorithms it considers permit fractional delegations - i.e., voters have voting weight 1, and they can delegate it fractionally across multiple possible delegates. The paper considers two algorithms for finding a delegation graph (i.e. an assignment of fractional delegations from voters who wish to delegates, to trusted delegates). These two algorithms are: (1) the random walk rule, a Markov-chain based rule proposed in past work; and (2) the mixed Borda branching rule, which is a uniform average over all min-cost branchings, called “borda branchings”, where a branching occurs on a weighted digraph representing to which casting voters each voter is willing to delegate to. The paper makes three main contributions: (1) Provides a polynomial-time algorithm for computing mixed Borda branching. This algorithm is an adaptation of Fulkerson’s algorithm, and relies on two of its canonical properties (as proven by Fulkerson). This algorithm is of independent interest to the direct power watershed problem. (2) Proves the equivalence of rules (1) and (2) above. (3) Shows that rule (1) (and therefore rule (2)) satisfies three axioms: confluence, anonymity, and copy-robustness — a combination of axioms identified in past work that cannot be simultaneously achieved with delegation rules that require voters to delegate their voting weight to a single delegate.

Strengths

- The high-level structure of the paper is clearly laid out - Some aspects of the paper’s analysis (e.g., axiom iii) are practically motivated, and the setting of liquid democracy in general is well-motivated - The authors made an effort to make the paper understandable, with multiple diagrams to aid explanations - The paper speaks to and builds on multiple aspects of the literature (e.g., random walk rule, existing axioms) - The technical results seem to holds, and there is sufficient technical exposition to understand why the results are true.

Weaknesses

1) The potential impact of this paper is not clear entirely to me. What is the main technical challenge with proving the results and what makes it hard? Why are axioms (i) and (ii) important for a rule to satisfy in practice? 2) I found it very hard to understand several phrases in the introduction. In many cases, it seems that the paper is assuming too much specific knowledge of the liquid democracy literature. For example: - Line 47: “(i.e., a digraph with a rank function on the edges)” what is a “rank function on edges”? - Line 51: In the explanation of confluence, what is a “subpath of v1”? - Line 64: “However, since none of the axioms connects the meaning of the ranks to the decisions of the delegation rule” I do not understand what this sentence means. - Line 66: “A single top-rank edge to a casting voter should always be chosen over any other delegation path.” I don’t understand what this sentence means.

Questions

Was it previously known whether the random walk rule satisfies these three axioms simultaneously?

Rating

6: Weak Accept: Technically solid, moderate-to-high impact paper, with no major concerns with respect to evaluation, resources, reproducibility, ethical considerations.

Confidence

3: You are fairly confident in your assessment. It is possible that you did not understand some parts of the submission or that you are unfamiliar with some pieces of related work. Math/other details were not carefully checked.

Soundness

4 excellent

Presentation

3 good

Contribution

2 fair

Limitations

Yes

Reviewer vsDd2023-08-12

Response

I have the read the authors' response and I thank them for the detailed clarifications. I have no further questions.

Reviewer rGK42023-08-15

Thank you for addressing my concerns. Discussing the connection to supervised learning earlier in the paper will contribute to its suitability for NeurIps and this will also be a good place to mention that the equivalence of the two rules also follows from results in [1]. Given that, contrary to my initial impression, [1] does not already solve your problem, I will increase my score.

Reviewer rLuU2023-08-16

Updating my View

First, I thank the authors for their detailed answers to my questions. I can see now that view that my criticism that fractional delegation trivializes the axiom of confluence was incorrect. I retract this criticism. It still appears there is a small problem with isolated voters, but this is an inescapable part of the nature of the problem. There is, in a sense, a violation of the axiom, but not in a way that conflicts with the motivation, and this is certainly not a reason for rejection. My view of the paper is now strongly positive and I argue for acceptance.

Program Chairsdecision2023-09-21

Decision

Accept (spotlight)

© 2026 NYSGPT2525 LLC