Notice: This website is continuously updated as new conference proceedings and journal volumes are published. Consequently, the figures presented here may differ from those reported in our publications.
A brief overview of our methodology is provided on the Methods page, along with references to publications documenting our approach and links to the code and data used in each study.
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.
Improved Regret Bounds for Tracking Experts with Memory
Authors: James Robinson, Mark Herbster
NeurIPS 2021 | Venue PDF | LLM Run Details | Input Tokens: 18,494 Total number of tokens sent to the LLM as input for this paper's analysis. | Output Tokens: 5,720 Total number of tokens produced by the LLM (including reasoning/thinking tokens) for this paper's analysis.
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Theoretical | We give a linear-time algorithm that improves on the best known regret bounds [27]. This algorithm incorporates a relative entropy projection step. This projection is advantageous over previous weight-sharing approaches in that weight updates may come with implicit costs as in for example portfolio optimization. We give an algorithm to compute this projection step in linear time, which may be of independent interest. In this paper we present an O(n)-time per trial projection-based algorithm for which we prove the best known regret bound for tracking experts with memory. |
| Researcher Affiliation | Academia | James Robinson Department of Computer Science University College London London United Kingdom EMAIL Mark Herbster Department of Computer Science University College London London United Kingdom EMAIL |
| Pseudocode | Yes | Algorithms 1&2 Po DS-θ / Share-θ Input: n > 0, η = 1 c > 0, α [0, 1], θ [0, 1] 1: init: w1 1 n Po DS-θ & Share-θ 2: for t 1 to T do 3: receive xt Dn 4: predict ˆyt = pred(wt, xt) 5: receive yt Y 6: for i 1 to n do 7: wt i wt ie ηℓt i Pn j=1 wt je ηℓt j 8: wt+1 P( wt; C(βt)) (15) 9: βt+1 (1 θ)βt + θα wt 8: wt+1 (1 α) wt + αvt (16) 9: vt+1 (1 θ)vt + θ wt (17) Algorithm 3 P(w; C(β)) in O(n) time Input: w ri n; β (0, 1)n s.t. β 1 1 Output: w = P(w; C(β)) 1: init: W [n]; r w 1 β; Sw 0; Sβ 0 2: while W = do 3: φ median({ri : i W}) 4: L {i W : ri < φ} 5: Lβ P i L βi; Lw P i L wi 6: M {i W : ri = φ} 7: Mβ P i M βi; Mw P i M wi 8: H {i W : ri > φ} 9: λ 1 Sβ Lβ 1 Sw Lw 10: if φλ < 1 then 11: Sw Sw + Lw + Mw 12: Sβ Sβ + Lβ + Mβ 13: if H = then 14: φ min({ri : ri > φ, i [n]}) 15: W H 16: else 17: W L 18: λ 1 Sβ 1 Sw 19: i : 1, . . . , n : w i βi ri < φ λwi ri φ |
| Open Source Code | No | The paper does not provide an unambiguous statement or link indicating the release of open-source code for the methodology described. |
| Open Datasets | No | The paper focuses on theoretical analysis and mathematical proofs of regret bounds, and does not describe the use of any specific dataset for training or experimentation. |
| Dataset Splits | No | The paper focuses on theoretical analysis and does not describe the use of any specific dataset splits (e.g., training, validation, test) for experimental reproduction. |
| Hardware Specification | No | The paper focuses on theoretical contributions and does not mention any specific hardware used for running experiments. |
| Software Dependencies | No | The paper focuses on theoretical contributions and does not specify any software dependencies with version numbers for experimental reproducibility. |
| Experiment Setup | No | The paper focuses on theoretical contributions and does not provide specific experimental setup details such as hyperparameter values or training configurations. |