Coin Betting and Parameter-Free Online Learning
Authors: Francesco Orabona, David Pal
NeurIPS 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We have also run an empirical evaluation to show that the theoretical difference between classic online learning algorithms and parameter-free ones is real and not just theoretical. In Figure 1, we have used three regression datasets4, and solved the OCO problem through OLO. In all the three cases, we have used the absolute loss and normalized the input vectors to have L2 norm equal to 1. From the empirical results, it is clear that the optimal learning rate is completely data-dependent, yet parameter-free algorithms have performance very close to the unknown optimal tuning of the learning rate. Moreover, the KT-based Algorithm 1 seems to dominate all the other similar algorithms. |
| Researcher Affiliation | Collaboration | Francesco Orabona Stony Brook University, Stony Brook, NY francesco@orabona.com David P al Yahoo Research, New York, NY dpal@yahoo-inc.com |
| Pseudocode | Yes | Algorithm 1 Algorithm for OLO over Hilbert space H based on KT potential; Algorithm 2 Algorithm for Learning with Expert Advice based on δ-shifted KT potential |
| Open Source Code | No | The paper does not provide a direct link to the source code for the methodology described, nor does it explicitly state that the code is publicly released. |
| Open Datasets | Yes | Datasets available at https://www.csie.ntu.edu.tw/~cjlin/libsvmtools/datasets/. |
| Dataset Splits | No | The paper mentions using datasets for empirical evaluation but does not specify the training, validation, or test splits (e.g., percentages or exact counts) needed to reproduce the data partitioning. |
| Hardware Specification | No | The paper describes running empirical evaluations but does not specify any hardware details (e.g., CPU/GPU models, memory) used for these experiments. |
| Software Dependencies | No | The paper does not specify any software dependencies with version numbers (e.g., programming languages, libraries, or specific solvers). |
| Experiment Setup | No | The paper refers to general parameters like "learning rate" when discussing other algorithms and describes dataset generation for LEA, but it does not provide specific hyperparameter values or detailed system-level training configurations for its own algorithms (Algorithm 1 and 2) used in the experiments. |