Diversified Recommendations for Agents with Adaptive Preferences

Authors: William Brown, Arpit Agarwal

NeurIPS 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical We formalize this problem as an adversarial bandit task. At each step, the Recommender presents a menu of k (out of n) items to the Agent, who selects one item in the menu according to their unknown preference model... We define a class of preference models which are locally learnable... For this class, we give an algorithm for the Recommender which obtains O(T 3/4) regret... We also give a set of negative results justifying our assumptions, in the form of a runtime lower bound for non-local learning and linear regret lower bounds for alternate benchmarks.
Researcher Affiliation Academia Arpit Agarwal Department of Computer Science Columbia University New York, NY 10027 arpit.agarwal@columbia.edu William Brown Department of Computer Science Columbia University New York, NY 10027 w.brown@columbia.edu
Pseudocode Yes Algorithm 1 (Robust Contracting FKM). Input: sequence of contracting convex decision sets K1, . . . KT containing 0, perturbation vectors ξ1, . . . , ξT where ξt ϵ, parameters δ, η. Set x1 = 0 for t = 1 to T do... Algorithm 2 A no-regret recommendation algorithm for adaptive agents. Input: Item set [n], menu size k, Agent with λ-dispersed memory model M for λ k2 n , where M belongs to an S-locally learnable class M, diversity constraint Hc, horizon T, G-Lipschitz linear losses ρi, . . . , ρT . Let tpad = Θ(1/ϵ3)...
Open Source Code No The paper does not contain any statements about releasing source code for the described methodology or links to code repositories.
Open Datasets No This is a theoretical paper presenting algorithms and regret bounds, not empirical results based on datasets. Therefore, no information on public datasets or their access is provided.
Dataset Splits No This is a theoretical paper presenting algorithms and regret bounds, not empirical results based on datasets. Therefore, no dataset split information for training, validation, or testing is provided.
Hardware Specification No This is a theoretical paper that does not report on empirical experiments requiring specific hardware. Therefore, no hardware specifications are mentioned.
Software Dependencies No This is a theoretical paper that does not report on empirical experiments requiring specific software dependencies with version numbers. Therefore, no such details are provided.
Experiment Setup No This is a theoretical paper that does not report on empirical experiments. Therefore, it does not include specific experimental setup details such as hyperparameters or training configurations.