Pricing a Low-regret Seller
Authors: Hoda Heidari, Mohammad Mahdian, Umar Syed, Sergei Vassilvitskii, Sadra Yazdanbod
ICML 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | 7. Simulations In this section we empirically evaluate the performance of Algorithm 1 and 2, and compare them with a baseline. ... Table 1. Regret values after T = 106 steps ALGORITHM NOT SELECTED EXTRA PAYMENT REGRET ALGORITHM 1 61110 32040 74817 ALGORITHM 2 8585 9227 15236 BASELINE 840149 -908 587196 ... Figure 1 illustrates the total regret of each algorithm as a function of time in the logarithmic scale. |
| Researcher Affiliation | Collaboration | Hoda Heidari HODA@CIS.UPENN.EDU Mohammad Mahdian MAHDIAN@GOOGLE.COM Umar Syed USYED@GOOGLE.COM Sergei Vassilvitskii SERGEIV@GOOGLE.COM Sadra Yazdanbod YAZDANBOD@GATECH.EDU |
| Pseudocode | Yes | Algorithm 1 Binary Search Pricing Algorithm (lines 1-15) and Algorithm 2 Heuristic Pricing Algorithm (lines 1-10) are presented in structured blocks with clear steps. |
| Open Source Code | No | The paper does not contain any statements about providing source code, a link to a repository, or mentioning code in supplementary materials. |
| Open Datasets | No | The paper describes a simulation setup where the price of the outside option 'comes from a uniform distribution on [0, 2µ] where µ = 0.3'. It does not use or provide access information for a publicly available or open dataset. |
| Dataset Splits | No | The paper details a simulation setup where data is generated and parameters are chosen, but it does not specify train/validation/test splits from a fixed dataset. The text mentions 'We take T = 10^6 and run both Algorithms 1, 2 with a range of values for their free parameters' but not explicit data partitioning. |
| Hardware Specification | No | The paper does not provide any specific details about the hardware (e.g., GPU models, CPU types, memory) used to run the simulations or experiments. |
| Software Dependencies | No | The paper mentions using 'the algorithm EXP3.P' for the seller and baseline, citing a related work. However, it does not provide specific version numbers for EXP3.P or any other software components or libraries used in their own implementation. |
| Experiment Setup | Yes | Simulation setup The simulation setup is as follows: we assume the price p B t of the outside option comes from a uniform distribution on [0, 2µ] where µ = 0.3. ... We take T = 106 and run both Algorithms 1, 2 with a range of values for their free parameters (i.e. the function f and the value θ for Algorithm 1, and the values α and β for Algorithm 2). ... For Algorithm 1, we use the functional form f(k) = a log(T)2βk ... A grid search over the ranges a [0.5, 2.5], β [1, 2.5], and θ [0.1, 0.3] reveals that the values a = 2, β = 1.5, and θ = 0.2 result in the lowest regret. ... For Algorithm 2, a grid search over the range 0 < α < β 1 finds that the combination α = 0.1 and β = 0.5 results in the lowest regret. |