- Rating
- 1
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 2
Read review
Paper summary
The paper proposes a gradient-based swarm method for continuous non-convex optimization. It uses the Gibbs-weighted average of the particles' objective values, , as an ensemble energy and adds Brownian noise to its gradient flow. This energy induces a state-dependent effective learning rate for each particle: some particles perform gradient descent, while others perform gradient ascent.
The theoretical analysis characterizes locally asymptotically stable equilibria and proves an exponential bound on the ensemble minimum under a suitable Polyak–Łojasiewicz condition for the deterministic dynamics. By constructing a one-dimensional channel in which the other particles are frozen, the paper further obtains a quasipotential upper bound of order that is independent of the physical barrier height. The experiments cover double-well and quadruple-well landscapes, Lennard–Jones clusters, the Ackley function, and transformers trained on modular addition. They examine reliability, computational cost, ensemble size, and the contribution of the sign-changing pre-factor.
Critical strengths
- A clear mathematical connection between the ensemble energy, stable equilibria, and escape channels.
Starting from a simple ensemble energy, the paper derives the explicit gradient decomposition and uses it to characterize the combinations of local minima and local maxima that can occur at locally asymptotically stable equilibria. The invariant channel obtained by freezing the other particles further reduces the analysis to an explicit reduced potential and yields a bound independent of the physical barrier height.
- The experiments report meaningful controls, failures, and limitations.
The authors report reliability and cost by difficulty, vary the ensemble size from 5 to 400 particles, and compare signed, positive-part, and absolute-value variants of the pre-factor. The Lennard–Jones experiments apply the same final local relaxation to every method, and the transformer experiments give Adam and SGD the same number of independent initializations as the Softmin ensemble.
Critical weaknesses
- Asymmetric hyperparameter selection weakens the principal reliability and cost comparisons.
The main concern arises in Appendix E.3. Softmin's inverse temperature is ultimately selected through direct validation of actual first-entry time, whereas the initial temperature of Simulated Annealing is selected using a different criterion.
The authors state that increasing makes escape faster in their continuous-diffusion setting. They therefore do not select it by escape time. Instead, they require $ \exp(-h/T_0)\in[0.01,0.05], $ and choose , to avoid making the barrier too easy to cross at the initial temperature.
This restriction is not adequately justified for an optimization comparison whose principal metrics are first-entry time and expected running time. If increasing the temperature improves the baseline under the stated evaluation criterion, that possibility should be included in its parameter selection. Testing similar numbers of candidates does not compensate for optimizing different selection criteria.
This issue directly affects the main conclusions. The authors have already established that SA's failures at high barriers are associated with freezing under the chosen geometric cooling schedule, and these failures substantially increase the cost differences in Figure 1 and Tables 10 and 12. The current experiments therefore do not rule out the possibility that Softmin's advantage would become substantially smaller if the initial temperature and cooling rate were selected under the same objective.
The baseline selection also leaves the interaction mechanism and application-level competitiveness incompletely assessed. Existing Swarm-based Simulated Annealing uses relative objective values to transfer mass between agents and adjusts their temperatures accordingly, making it a directly relevant interacting gradient/noise method. On Lennard–Jones clusters, a comparison with basin-hopping, which the paper itself cites, would provide a more informative assessment of competitiveness on that application. These comparisons would respectively help assess the benefit of particle interaction and the strength of the application results.
- The specific benefit of is not sufficiently isolated; a barrier cap of the same order does not require sign reversal.
The paper emphasizes sign reversal and barrier flattening as its principal mechanism. However, the simpler log-sum-exp energy already discussed in the related work also admits a channel bound of the same order. The following calculation uses the paper's definitions and the same channel construction.
Let $ F_\beta(\mathbf{x})=-\frac1\beta\log\sum_{i=1}^{n}e^{-\beta f(x_i)}, $ and apply the same factor of used in the proposed dynamics. Then
Every pre-factor is positive, so there is no gradient ascent. Along the same channel with the other particles frozen, let . The potential relative to the initial configuration is
For ,
Therefore, for every , $v(h)<\min ${ }.
The same channel-action argument thus gives an upper bound that is independent of and of order . Consequently, the existence of an barrier cap does not by itself distinguish the proposed sign-changing mechanism from Gibbs-weighted diffusion whose drift always follows the descent direction.
This observation does not eliminate the value of . The proposed energy can provide a smaller initial channel-action constant and different post-saddle dynamics, and those differences may be consequential in the small-noise regime. The paper nevertheless needs to establish quantitatively when they translate into an advantage.
The reported ablations reinforce this concern. Appendix E.5.4 reports mean first-entry times of 32.69, 30.70, and 10.82 for the signed, positive-part, and absolute-value variants, respectively, on the quadruple well at under the main setting . The distributions for the three variants substantially overlap on Ackley and Lennard–Jones. The additional ablation does provide some evidence in favor of sign reversal, but it does not explain the advantage under the main settings.
A direct comparison with diffusion driven by , together with a pure-noise control that removes the drift, would assess the specific benefit of the proposed energy more directly than changing only the sign treatment of .
Questions for authors
-
Do the principal conclusions persist when SA is calibrated under the same objective? If the initial temperature and cooling rate are jointly selected using the same reliability, first-entry-time, or ERT criterion used for Softmin, without imposing a low Arrhenius escape factor, do the high-barrier reliability and cost advantages remain?
-
What does add beyond diffusion driven by the log-sum-exp energy? Under comparable parameter-selection ranges, ensemble sizes, noise levels, and computational budgets, can the authors directly compare against drift and include a pure-noise control? In particular, under what conditions do the smaller action constant and altered post-saddle dynamics yield a measurable benefit?
-
Which exit theorem is intended in Corollary 3.5? The point is necessarily an equilibrium, whereas Appendix B excludes equilibria on the boundary. Can the authors state a compatible theorem and its hypotheses, and at least explain how those hypotheses apply to a concrete class of potentials considered in the paper?
-
Are unsuccessful runs stopped early, and what cost does Table 12 record? Table 1 specifies and a budget of steps, while Table 12 reports an SA success rate of approximately . If the remaining runs exhaust the full budget, as stated in the caption, the mean number of per-particle gradient evaluations should be at least approximately whereas the table reports . Early stopping could explain this discrepancy, but the current description is insufficient to establish that.
We thank the reviewer for the detailed review. We checked the calculations in it and agree with most of them.
W1/Q1 The reviewer is right that β and T0 were selected under different criteria. The escape time of Simulated Annealing decreases monotonically in T0, so selecting it by escape time would pick the largest value in any grid; we excluded this with the Arrhenius criterion, but we agree that this does not justify a different objective. We have not re-run Simulated Annealing during the discussion, so we cannot say whether the high-barrier advantage survives: under geometric cooling a larger T0 should shift the barrier at which it freezes rather than remove it, but this has to be checked. In a revised version we will select T0 and the cooling rate jointly by ERT at the same point used for β, without the Arrhenius restriction, extend the sweep until the schedule fails, and report the outcome either way. Until then, our conclusions on Simulated Annealing are conditional on the calibrated schedule.
W2/Q2 The reviewer's calculation is correct: the O(1/β) cap comes from the Gibbs weighting, which log-sum-exp shares, and we will say so. What the sign change adds is a smaller constant. From q(rc)=0 one gets exactly uc=n(rc−1/β), so uc/v∞=(βrc−1)/logn−1n, which is 0.40 at n=2 and decreases to 1/e. The post-saddle structure also changes: for h>rc, Jβ creates a shallow well at Γ(s), but even adding this second crossing the channel cost 2uc−u(h) stays below v(h) for every n≥2. The gain is therefore a constant factor in the bound on the exponent, not a different order in β. Empirically, only the β=2 ablation (App. E.5.4) shows the sign mattering; we will add the drift −nwk∇f(xk) and a pure-noise control on the wells, with the same seeds, ensemble size, noise and budget.
Q3 The reviewer is right: the point Γ(xm) where the pre-factor flips is an index-one saddle of nJβ on ∂D− (for h>rc the Hessian there is diagonal, with entries −nβw1f′(xm)2<0 and n−1nf′′(z−)>0), so the exit theorem recalled in Appendix B does not apply as stated. We will restate Corollary 3.5 with an exit theorem that allows a characteristic boundary, restricted to the exit from D−; we are still checking which result fits our setting and what its exact hypotheses are. For a gradient drift, such a theorem gives limσ→0σlogEτD−=min∂D−nJβ−nJβ(Γ(z−))≤uh. For our double wells, Γ(z−) and Γ(xm) are nondegenerate, while the boundedness of D− and the equilibria on its boundary remain to be checked. The quasipotential bounds of Theorem 3.4 are not affected.
Q4 Yes, unsuccessful runs are stopped early, for every method, when the ensemble collapses (largest per-coordinate standard deviation below 10−6) or the best value does not improve for 5⋅105 steps. All failed Simulated Annealing runs stop by collapse at about 5.1⋅105 steps, which gives the value in Table 12 (0.275×100×512,900≈1.41×107), and all failed Langevin runs stop by stagnation at 5.49⋅105 steps. The caption of Table 12 is wrong: a failed run contributes the steps it ran, not the full budget. We will fix it and describe the rule in App. E.3. Since failed runs are charged less than the full budget, the cost ratios are conservative. Stopping does not change the outcomes: the collapsed Simulated Annealing ensemble is frozen, and for Langevin at the smallest failing barrier Kramers' formula gives about 108 time units for the first of 100 particles to escape, against a full budget of 104.