- Rating
- 3
- Confidence
- 2
- Informativeness
- 2
- Presentation
- 3
- Soundness
- 2
Read review
Paper summary
In this paper, the authors consider residual layers and time-conditioned flow via the full path of intermediate transformations. Under the several assumptions, they provide the upper and lower bounds for the prefix information rate.
Critical strengths
The strength of this paper is to show that, when , the full transformation path and its endpoint have same worst-case leading information power, while reconstructing the full vector field requires more information. The claim obtained theoretically is very interesting. The experiments provide good empirical support for the theoretical results.
Critical weaknesses
The authors claim the full transformation path and its endpoint have same worst-case leading information power when . However, since , the condition covers the broad range. The restriction to might be restrictive. Also, the analysis seems to be restricted as Euler grids. I am wondering if the rate might changed when using other discretization, such as Runge-Kutta or adaptive solvers.
Questions for authors
-
Is there an intrinsic difficulty in extending the analysis beyond ? Also, if such a extension is possible, could we expect the same claim that the full transformation path and its endpoint have same worst-case leading information power when .
-
Could authors discuss the other discretization, such as Runge-Kutta or adaptive solvers ? I do not ask the authors to add details, but just a brief discussion on this possibility.
2 comments
Thank you very much for the clarification. I would like to keep my current score as weak acceptance.
Thank you for your questions on higher regularity and other solvers. They helped us extend the analysis and clarify the solver comparison.
Higher spatial regularity. New Appendix K proves matching path information bounds for every fixed integer s>2. For fields supported in [−1,1]d with uniformly bounded spatial Cs−1,1 norms and time-Lipschitz values, the endpoint power is d/s. On dyadic L≥Cd,sε−1/2, the all-prefix path power is max{d/s,(1+d/s)/2}, up to logarithms. Thus endpoint and path have the same power when d≥s. The proof uses inverse-stable, quantized local Taylor approximations. One time-conditioned ReLU network realizes the code; its full graph and per-call work have polylog(L/ε) overhead.
Runge--Kutta and adaptive solvers. New Appendix G.3 shows that one fixed graph defines a continuous flow within ε of every compatible teacher flow. The graph is built from L0≳ε−1 full layer functions of a field in our class, with s∈{1,2} and exact graph arithmetic. Compatible teacher fields are in this class and agree with all source slices. If a Runge--Kutta or adaptive solver has uniform global node error δ on this graph, its error against the teacher flow at those nodes is at most ε+δ. Our Ω(ε−1/2) lower bound counts linear-output segments with original-grid knots.