tag:blogger.com,1999:blog-4999686744433629981.post6100150285460428291..comments2024-02-14T22:06:08.851-08:00Comments on Augmented Social Cognition Research Blog from PARC: User Needs during Social SearchEd H. Chihttp://www.blogger.com/profile/06360447323238002978noreply@blogger.comBlogger7125tag:blogger.com,1999:blog-4999686744433629981.post-38638054913488693202009-01-09T19:23:00.000-08:002009-01-09T19:23:00.000-08:00Brynn and I are working failure cases now; so we w...Brynn and I are working failure cases now; so we want to see if failures have different characteristics.Ed H. Chihttps://www.blogger.com/profile/06360447323238002978noreply@blogger.comtag:blogger.com,1999:blog-4999686744433629981.post-37348714494057695102009-01-09T14:07:00.000-08:002009-01-09T14:07:00.000-08:00Hi Brynn,Very interesting search model! Are you pl...Hi Brynn,<BR/><BR/>Very interesting search model! <BR/><BR/>Are you planning another research in this field (social search)?<BR/><BR/>-- MotiAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-4999686744433629981.post-39845004453260369472008-10-18T13:05:00.000-07:002008-10-18T13:05:00.000-07:00Ed,Definitely there is a meeting of minds and line...Ed,<BR/><BR/>Definitely there is a meeting of minds and line of similar thoughts here. <BR/>Our team was also so glad to know ASC's works.<BR/><BR/>Very briefly, in IR(information retrieval) perspective, <BR/>using tag for index/lexicon was very unconventional and bumpy roads during the course of a year long beta development.<BR/>As Mr. Heyman investigated, there are both good and bad news for using them. <BR/><BR/>A couple of things we have learned is <BR/><BR/> - tags is incomplete to build lexicon; for IR purposes, it is messy in raw format. <BR/> So it requires us to do fairly significant works to process them. i.e., cleansing, tokenizing and so forth<BR/> - sparsity and scale are both big issues. It was challenging to tackle mitigation<BR/> - Once you scale and supplement social annotated data, <BR/> we couldn't rely on lexical approach for tags and associated URLs<BR/> but statistical methods.<BR/><BR/>But the fact that matter is <BR/>social search engines will provide "sharing domain knowledge and expertise" as you can try on Tusavvy. <BR/>We believe it returns "really unusual, but accurate results."<BR/>That is why this space is so intriguing. <BR/><BR/>If there is a chance to share more findings in private setting,<BR/>we're happy to do so.<BR/><BR/>Looking forward to seeing what you are building<BR/><BR/>Best,<BR/>JaeSungAnonymoushttps://www.blogger.com/profile/05032123526545413531noreply@blogger.comtag:blogger.com,1999:blog-4999686744433629981.post-38533930906901974462008-10-17T17:18:00.000-07:002008-10-17T17:18:00.000-07:00JaeSung,We have been building a social exploration...JaeSung,<BR/><BR/>We have been building a social exploration engine that uses social tags that was somewhat informed by the findings in the study. We're encouraged to see that you have a social search engine based on tags, so there is definitely a meeting of minds here. Stay tuned to our release soon.Ed H. Chihttps://www.blogger.com/profile/06360447323238002978noreply@blogger.comtag:blogger.com,1999:blog-4999686744433629981.post-43062538697971821622008-10-17T07:58:00.000-07:002008-10-17T07:58:00.000-07:00Brynn,We're hoping to continue a dialogue as well....Brynn,<BR/><BR/>We're hoping to continue a dialogue as well. Especially, one of the topics we like to have a dialogue is - haven't been covered in the paper we think - how to use social graph in demographic sense, not individual manner.<BR/><BR/>Best,<BR/>JaeSungAnonymoushttps://www.blogger.com/profile/05032123526545413531noreply@blogger.comtag:blogger.com,1999:blog-4999686744433629981.post-50683055542507324232008-10-16T17:20:00.000-07:002008-10-16T17:20:00.000-07:00Hi JaeSung! I haven't heard of Tusavvy but I am ch...Hi JaeSung! I haven't heard of Tusavvy but I am checking it out.<BR/><BR/>It sounds very promising that the suggestions in the paper align with the thoughts on your beta implementation. <BR/><BR/>I'm glad you found our post and hope we can continue the conversation about social search.Brynn Evanshttps://www.blogger.com/profile/14339567920724336199noreply@blogger.comtag:blogger.com,1999:blog-4999686744433629981.post-27222914598765139732008-10-16T12:16:00.000-07:002008-10-16T12:16:00.000-07:00As the founder of a social search engine, Tusavvyd...As the founder of a social search engine, Tusavvy<BR/>during a couple of my reading on the post and full paper, <BR/>I have felt that the survey and its design suggestion with 'canonical social model' is very Intriguing. <BR/><BR/>Moreover, I was stunned by the similarity of the line of thoughts to our beta implementaion <BR/><BR/>I strongly encourage you to visit our public beta to explore.<BR/>http://www.tusavvy.com<BR/><BR/>Seems to me that our query construction helper is very much overlapped with the "before search" suggestion.<BR/>Moreover, Tusavvy provides built-in web interface with search results for post sharing purposes.<BR/><BR/>Lastly, note that Tusavvy search results are pulled out from human expertise.<BR/><BR/>JaeSung Ro | Founder | zSoupAnonymoushttps://www.blogger.com/profile/05032123526545413531noreply@blogger.com