# Multitask Online Learning: Listen to the Neighborhood Buzz

arXiv: https://arxiv.org/abs/2310.17385 · HTML: https://mlsanity.com/paper/2310.17385/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 12 of its 40 references that are in the map.
- First posted on arXiv 26 Oct 2023.

## Builds on (most cited first, 12 of 12)

- [Federated Multi-Task Learning](https://mlsanity.com/paper/1705.10467/) (2017)
- [Online Learning: A Modern Introduction Using Convex Optimization](https://mlsanity.com/paper/1912.13213/) (2019)
- [Delay and Cooperation in Nonstochastic Bandits](https://mlsanity.com/paper/1602.04741/) (2016)
- [Personalized and Private Peer-to-Peer Machine Learning](https://mlsanity.com/paper/1705.08435/) (2017)
- [The Price of Differential Privacy For Online Learning](https://mlsanity.com/paper/1701.07953/) (2017)
- [Cooperative Online Learning: Keeping your Neighbors Updated](https://mlsanity.com/paper/1901.08082/) (2019)
- [Multi-Agent Online Optimization with Delays: Asynchronicity, Adaptivity, and Optimism](https://mlsanity.com/paper/2012.11579/) (2020)
- [Individual Regret in Cooperative Nonstochastic Multi-Armed Bandits](https://mlsanity.com/paper/1907.03346/) (2019)
- [Doubly Adversarial Federated Bandits](https://mlsanity.com/paper/2301.09223/) (2023)
- [An Efficient Algorithm for Cooperative Semi-Bandits](https://mlsanity.com/paper/2010.01818/) (2020)
- [Multitask Online Mirror Descent](https://mlsanity.com/paper/2106.02393/) (2021)
- [Distributed Online Learning for Joint Regret with Communication Constraints](https://mlsanity.com/paper/2102.07521/) (2021)

## Built on by (most cited first, 1 of 1)

- [Online-to-PAC generalization bounds under graph-mixing dependencies](https://mlsanity.com/paper/2410.08977/) (2024)

## Abstract (arXiv)

We study multitask online learning in a setting where agents can only exchange information with their neighbors on an arbitrary communication network. We introduce textttMT-CO2textttOL, a decentralized algorithm for this setting whose regret depends on the interplay between the task similarities and the network structure. Our analysis shows that the regret of textttMT-CO2textttOL is never worse (up to constants) than the bound obtained when agents do not share information. On the other hand, our bounds significantly improve when neighboring agents operate on similar tasks. In addition, we prove that our algorithm can be made differentially private with a negligible impact on the regret. Finally, we provide experimental support for our theory.