?url_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Adc&rft.title=InfoRank%3A+Unbiased+Learning-to-Rank+via+Conditional+Mutual+Information+Minimization&rft.creator=Jin%2C+J&rft.creator=He%2C+Z&rft.creator=Yang%2C+M&rft.creator=Zhang%2C+W&rft.creator=Yu%2C+Y&rft.creator=Wang%2C+J&rft.creator=McAuley%2C+J&rft.description=Ranking+items+regarding+individual+user+interests+is+a+core+technique+of+multiple+downstream+tasks+such+as+recommender+systems.+Learning+such+a+personalized+ranker+typically+relies+on+the+implicit+feedback+from+users'+past+click-through+behaviors.+However%2C+collected+feedback+is+biased+toward+previously+highly-ranked+items+and+directly+learning+from+it+would+result+in+%22rich-get-richer%22phenomena.+In+this+paper%2C+we+propose+a+simple+yet+sufficient+unbiased+learning-to-rank+paradigm+named+InfoRank+that+aims+to+simultaneously+address+both+position+and+popularity+biases.+We+begin+by+consolidating+the+impacts+of+those+biases+into+a+single+observation+factor%2C+thereby+providing+a+unified+approach+to+addressing+bias-related+issues.+Subsequently%2C+we+minimize+the+mutual+information+between+the+observation+estimation+and+the+relevance+estimation+conditioned+on+the+input+features.+By+doing+so%2C+our+relevance+estimation+can+be+proved+to+be+free+of+bias.+To+implement+InfoRank%2C+we+first+incorporate+an+attention+mechanism+to+capture+latent+correlations+within+user-item+features%2C+thereby+generating+estimations+of+observation+and+relevance.+We+then+introduce+a+regularization+term%2C+grounded+in+conditional+mutual+information%2C+to+promote+conditional+independence+between+relevance+estimation+and+observation+estimation.+Experimental+evaluations+conducted+across+three+extensive+recommendation+and+search+datasets+reveal+that+InfoRank+learns+more+precise+and+unbiased+ranking+strategies.&rft.publisher=Association+for+Computing+Machinery+(ACM)&rft.date=2024-05-13&rft.type=Proceedings+paper&rft.language=eng&rft.source=+++++In%3A++WWW+2024+-+Proceedings+of+the+ACM+Web+Conference.++(pp.+pp.+1350-1361).++Association+for+Computing+Machinery+(ACM)%3A+Singapore%2C+Singapore.+(2024)+++++&rft.format=text&rft.identifier=https%3A%2F%2Fdiscovery.ucl.ac.uk%2Fid%2Feprint%2F10194856%2F1%2F2401.12553v1.pdf&rft.identifier=https%3A%2F%2Fdiscovery.ucl.ac.uk%2Fid%2Feprint%2F10194856%2F&rft.rights=open