English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Collaborative learning techniques have the potential to enable trainingmachine learning models that are superior to models trained on a singleentity's data. However, in many cases, potential participants in suchcollaborative schemes are competitors on a downstream task, such as firms thateach aim to attract customers by providing the best recommendations. This canincentivize dishonest updates that damage other participants' models,potentially undermining the benefits of collaboration. In this work, weformulate a game that models such interactions and study two learning taskswithin this framework: single-round mean estimation and multi-round SGD onstrongly-convex objectives. For a natural class of player actions, we show thatrational clients are incentivized to strongly manipulate their updates,preventing learning. We then propose mechanisms that incentivize honestcommunication and ensure learning quality comparable to full cooperation.Lastly, we empirically demonstrate the effectiveness of our incentive scheme ona standard non-convex federated learning benchmark. Our work shows thatexplicitly modeling the incentives and actions of dishonest clients, ratherthan assuming them malicious, can enable strong robustness guarantees forcollaborative learning.

Incentivizing Honesty among Competitors in Collaborative Learning and Optimization