An adaptive social network for information access: Theoretical and experimental results
作者:Bin Yu、Mahadevan Venkatraman、Munindar P. Singh
DOI:10.1080/713827056
日期:2003.1
approaches, our architecture is fully distributed and includes agents who preserve the privacy and autonomy of their users. These agents learn models of each other in terms of expertise (ability to produce correct domain answers) and sociability (ability to produce accurate referrals). We study our framework experimentally to study how the social network evolves. Specifically, we find that under our multi-agent
我们考虑了一个软件代理的社交网络,他们互相帮助,帮助用户查找信息。与之前的大多数方法不同,我们的架构是完全分布式的,并且包括保护用户隐私和自主权的代理。这些代理在专业知识(产生正确领域答案的能力)和社交能力(产生准确推荐的能力)方面相互学习模型。我们通过实验研究我们的框架,以研究社交网络如何发展。具体来说,我们发现在我们的多智能体学习启发式下,网络的质量随着交互而提高:当同时考虑专业知识和社交能力时,质量最大化;枢轴代理进一步提高了网络的质量,并对其质量产生催化作用,即使它们最终被移除。而且,