Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].
Convergence Rate in a Nonlinear Two-Time-Scale Stochastic Approximation with State (Time)-Dependence
Authors: Zixi Chen, Yumin Xu, Ruixun Zhang
AAAI 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We provide two numerical examples to illustrate our theoretical findings in the context of stochastic gradient descent with Polyak Ruppert averaging and stochastic bilevel optimization. |
| Researcher Affiliation | Academia | Zixi Chen1, Yumin Xu1, Ruixun Zhang1,2,3,4* 1School of Mathematical Sciences, Peking University 2Center for Statistical Science, Peking University 3National Engineering Laboratory for Big Data Analysis and Applications, Peking University 4Laboratory for Mathematical Economics and Quantitative Finance, Peking University EMAIL, EMAIL, EMAIL |
| Pseudocode | No | The paper describes algorithms like SGD with Polyak-Ruppert averaging and stochastic bilevel optimization using mathematical equations (1), (3), (5), but does not present them in a structured pseudocode or algorithm block. |
| Open Source Code | No | The paper does not explicitly state that source code is provided for the methodology described, nor does it provide a link to a code repository. It only mentions that supplementary material contains proofs and figures. |
| Open Datasets | No | The numerical experiments utilize synthetic setups, such as minimizing a custom function F(x) = (x2 1+sin x1, ..., x2 5+ sin x5) with normal white noise for SGD with Polyak-Ruppert averaging, and specific functions F(x, y) and G(x, y) for stochastic bilevel optimization. These are not explicitly mentioned as publicly available or open datasets with access information. |
| Dataset Splits | No | The paper uses synthetic problem setups and noise generation for its numerical experiments, such as defining specific functions F(x) and G(x, y) and generating normal white noise. It initializes (x0, y0) = (1, 1) and iteratively updates values, but does not describe any training/test/validation dataset splits. |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., GPU/CPU models, memory, or detailed computer specifications) used for running its experiments. |
| Software Dependencies | No | The paper describes algorithms such as SGD with Polyak-Ruppert averaging and stochastic bilevel optimization, but does not mention any specific software dependencies or their version numbers used for implementation, such as programming languages, libraries, or frameworks. |
| Experiment Setup | Yes | We initialize (x0, y0) = (1, 1), and choose (α, β) to satisfy the same conditions as in Theorems 1-3. The step sizes αk and βk are set as αk = α (k+1+k0)a and βk = β (k+1+k0)b , where k0 is chosen optimally. ... For simplicity of analysis, we set δij δ, Γij Γ, and γk γ. Γ kk Γ as noise parameters. All experiments are repeated 1000 times. |