University of Cambridge > Talks.cam > Machine Learning @ CUED > Extending the Affinity Propagation Model

  Browse A-Z

Log in

University Account External (via Google)

🔐 Log In

Information on

Subscribing to talksdetails Finding a talkdetails Adding a talkdetails Disseminating talksdetails Help and Documentationdetails

ℹī¸ Information

Extending the Affinity Propagation Model

Download to your calendar using vCal

If you have a question about this talk, please contact Zoubin Ghahramani .

Affinity Propagation (AP, Frey & Dueck, 2007) is an exemplar-based clustering algorithm that is based on loopy belief-propagation on a particular factor-graph. In my talk I will present an alternative formulation and derivation of AP that lends itself nicely to modifications and extensions of the basic algorithm. In particular, I will discuss a semi-supervised variant of Affinity Propagation that uses equivalence constraints, and demonstrate its applicability to user interactive image segmentation. Finally, I will describe some on-going work regarding interesting links between the alternative AP formulation and some graph-theoretical problems such as bi-partite weighted matching, which lead us to believe it may be possible to establish convergence criteria for AP in terms of solutions obtained by linear programming relaxations.

This talk is part of the Machine Learning @ CUED series.

This talk is included in these lists:

Note that ex-directory lists are not shown.

Other lists

[tbc] M3 Second Court, St John's College FS06

Other talks

Title to be confirmed Necessary mechanisms for super AI and stopping hallucinations: The consistent reasoning paradox and the indeterminacy function Diasporic Divas: Performing Race, Gender, and Sexuality from the 18th Century to Today Unveiling the curtain with ukubuyisa or go lata: complicated roles of museums and nation states in the cultural repatriation discourse and practice in southern Africa Studying viruses to understand what makes pandemic viruses special and how innate immunity works CodeScaler: Scaling Code LLM Training and Test-Time Inference via Execution-Free Reward Models
 

Š 2006-2025 Talks.cam, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity