ReSync: Riemannian Subgradient-based Robust Rotation Synchronization

This work presents ReSync, a Riemannian subgradient-based algorithm for solving the robust rotation synchronization problem, which arises in various engineering applications. ReSync solves a least-unsquared minimization formulation over the rotation group, which is nonsmooth and nonconvex, and aims at recovering the underlying rotations directly. We provide strong theoretical guarantees for ReSync under the random corruption setting. Specifically, we first show that the initialization procedure of ReSync yields a proper initial point that lies in a local region around the ground-truth rotations. We next establish the weak sharpness property of the aforementioned formulation and then utilize this property to derive the local linear convergence of ReSync to the ground-truth rotations. By combining these guarantees, we conclude that ReSync converges linearly to the ground-truth rotations under appropriate conditions. Experiment results demonstrate the effectiveness of ReSync.

Paper

References (48)

Scroll for more · 36 remaining

Similar papers

Peer review

Reviewer aU9r7/10 · confidence 3/52023-07-01

Summary

The paper concerns synchronization of observed rotations with incomplete and corrupted observations. The authors construct the method ReSync that is a gradient-based algorithm for solving the problem. The paper describes the context, prior results on the synchronization problem, and the algorithm, and presents a thorough convergence analysis together with experimental evaluation.

Strengths

- well-written and clearly presented paper - the presented method deals with an important problem - the presentation of the algorithm is followed by a thorough convergence analysis - the method performs well in the experimental validation

Weaknesses

- constructing a gradient descent algorithm is not a big contribution in itself. However, I believe the geometric setting and the connection to the theoretical analysis, which is not trivial, makes the contribution important

Questions

no questions

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

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

yes

Reviewer grLa7/10 · confidence 3/52023-07-03

Summary

This paper presents a theoretical study for robust rotation synchronization with a least-unsquared minimization formulation over the rotation group. In particular, this paper proposes a two-step algorithm called ReSync, where the first step uses spectral initialization to generate an initial guess and the second step performs Riemannian Subgradient descent from the initial guess. The paper proves that, under suitable conditions of the random corruption model, this algorithm converges linearly to the groundtruth rotations. The paper presents numerical experiments that verify the correctness of the theorem and compares the performance of ReSync with other state-of-the-art algorithms.

Strengths

- The theoretical contribution of this paper advances the previous state of the art. - Paper is well written and easy to follow, despite being a theory paper.

Weaknesses

- I am curious if similar guarantees could be made in the case where the inlier measurements are corrupted by small (and bounded) noise? Could you guarantee the algorithm converges to a solution that has bounded error from the groundtruth rotations?

Questions

See Weaknesses.

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

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

N/A

Reviewer W8A37/10 · confidence 4/52023-07-09

Summary

This work proposes to solve the rotation synchronization problem using Riemannian subgradient method with spectral initialization. The proposed formuation is sum of absolute deviations, which is robust to outliers. Exact recovery guarantees are provided under uniform corruption model (the graph is Erdos Renyi and probability of corruption 1-p). Numerical results show competitive performance of the proposed method compared to other state-of-the-art methods.

Strengths

1. The theoretical result in the noiseless case (corruption only) is quite strong. That is, it shows linearly convergence to the ground truth rotations whenever n>1/(p^7q^2) up to a log factor, where p is the probability of being a clean edge (conditioned on being an edge), and q is the probability of being an edge. 2. The numerical experiments show advantages over previous state-of-the-art methods in the presense of both corruption and noise. 3. The proofs look correct. Overall, I enjoyed reading the paper.

Weaknesses

1. This is not necessarily a weakness, but it would be even nicer if the authors could comment on the stability of your algorithm to noise (would it be possible to show an approximate recovery in this case)?

Questions

I wonder how sensitive is your method to initialization? For example, given random initialization, what is the typical behavior of your algorithm in numerical experiments and what about the theoretical results?

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

3 good

Presentation

3 good

Contribution

3 good

Limitations

Yes

Reviewer m1vD6/10 · confidence 4/52023-07-29

Summary

The paper proposes a Riemannian subgradient based algorithm for the robust rotation synchronization (RRS) problem. RRS involves recovering the absolute rotations of objects from the possibly corrupted/noisy relative rotations between pairs of objects. The problem setting involves two ratios: q \in [0,1] denotes the observation ratio and p \in [0,1] denotes the true observation ratio. The paper pose the problem as a (non-convex and non-smooth) least-unsquared minimization formulation over the rotation group. The proposed method ReSync has a spectral relaxation based initialization procedure, which is followed by Riemannian subgradient iterations. The main contribution of the paper is to show that under random corruption model (RCM) setting: (a) the proposed initialization (X^0) can be relatively close to true solution (X^*), depending on p and q, and (b) given the initialization guarantee, the Riemannian subgradient descent show local linear rate of convergence. Overall, the paper show that ReSync converges linearly to the ground-truth rotations when p^7 q^2 = \Omega(log n / n). Towards the end of the draft, the paper has few experimental results that compare the proposed algorithm against state-of-the-art.

Strengths

The paper presents an interesting approach for recovering ground truth (X^*) in RRS problem. The key theoretical guarantees for ReSync comes from a) an initialization procedure SpectrIn, which ensures that the initialization X^0 is close to X^*, b) weak sharpness property of the least unsquared formulation, which is being solved via ReSync, and c) local linear convergence analysis for ReSync based on initialization and weak sharpness property. I have, however, not verified the correctness of the theoretical results.

Weaknesses

Concerns regarding theory: 1. The paper assumes missing observations as zero matrix, which does not lie on the SO(d) manifold. Hence, only if (i,j) belongs to available observation, Y_{ij} \in SO(d). The paper does not provide any justification for this choice. A more suitable choice seems to be Identity matrix as it lies on SO(d). 2. In line 178-180, it is stated that E[Y_ij] = pq X_i^*(X_j^*)^\top for all (i,j). This does not seem correct as it is not clear how outlier points O_ij \in SO(d) are handled while computing this expectation. 3. While the paper cites and discusses its differences with [27] in lines 171-175, it seems that [27] should be discussed in more detail. While [27] focuses on orthogonal group with additive Gaussian noise and permutation group with outliers, it should be noted that permutation group is a special subset of orthogonal group. Interestingly, [27] states that "though it is not analyzed in our manuscript, the proof technique for the permutation group synchronization under uniform corruption could be directly modified to tackle this O(d) synchronization under uniform multiplicative corruption" (in the paragraph before Section 3.2). The proof of leave-one-out technique seems to be adopted from [27]. Hence, while [27] has been cited in Section 3.1 of the paper, it does not seem to be the main contribution and could been discussed in the supplementary material. Overall, [27] deserves more discussion, especially w.r.t. the above quoted statement, and in this regard contribution of the paper should be clearly highlighted. 4. A discussion on computational cost of the proposed algorithm is missing. Concerns regarding experiments: 1. The paper shows only a few empirical results on synthetic datasets. While this gives some insights on how the algorithm works in lab environment, performance on real-world setting gives an idea on how the algorithm will perform when used in real applications. If space was a factor, the paper could have moved some of the proofs/proof-outlines to the supplementary section. 2. While the paper discusses [40] and states that it "introduces a least-unsquared formulation and applies the SDR method to tackle it", the paper should have mentioned more directly that the main formulation (2), which papers tries to solve, was originally proposed in [40]. Hence, while the theoretical results of [40] are in q=1 setting, the paper should have empirically compared with [40] as well. Similarly, paper [30] should also be compared empirically. 3. Experiments are done in two settings: with and without additive noise. Without additive noise setting has been theoretically analyzed in the paper. One question is that in this setting, DESC method seems to be better or similar to ReSync. Any insights as to why it does better (where it does)?

Questions

Please look in the weakness section.

Rating

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

1 poor

Presentation

2 fair

Contribution

2 fair

Limitations

not applicable

Reviewer aU9r2023-08-12

Thanks for the reply. My scoring has not changed.

Authorsrebuttal2023-08-18

Thank you for your response

Thank you very much for reading our rebuttal and for your response. We will be closely following the Reviewer-Author discussion period in case the reviewer has any additional concerns or questions.

Reviewer grLa2023-08-18

Thanks for the response, I maintain my original score.

Authorsrebuttal2023-08-18

Thank you for your response

Thank you very much for reading our rebuttal and for your response. We will be closely following the Reviewer-Author discussion period in case the reviewer has any additional concerns or questions.

Reviewer W8A32023-08-22

I thanks authors for the response and it addressed all my questions. After reading all the reviews and comments, it does not change my opinion that this is a solid theoretical paper. Therefore, I prefer not to change my score.

Authorsrebuttal2023-08-22

Thank you for your response

Thank you so much for reading all the reviews and our rebuttals and for your response.

Reviewer m1vD2023-08-22

Response to rebuttal

I thank the authors for the detailed response. Regarding point A: The following point is not clear. It would be nice if the authors could elaborate on this. >In theory, this setting ensures that the leading eigenvectors of $\mathbb{E}[\mathbf{Y}]$ are $\mathbf{X^{*}}$. Regarding point B: It would be nice if the authors could explain why independently and uniformly distributed outliers on SO(d) will have zero mean. >We have used the fact $\mathbb{E}[\mathbf{O}_{ij}]=0$ since outliers are assumed to be independently and uniformly distributed on SO(d) in the RCM. The authors have answered my other questions. I have accordingly changed my score

Program Chairsdecision2023-09-21

Decision

Accept (poster)

© 2026 NYSGPT2525 LLC