Orkut
Search results
Cloob.com was a Persian-language social networking website, mainly popular in Iran. After the locally (and internationally) popular social networking website Orkut was blocked by the Iranian government, a series of local sites and networks, including Cloob, emerged to fill the gap.
CLOOB uses InfoLOOB as objective to avoid the saturation of the InfoNCE objective used in CLIP. Modern Hopfield networks exacerbate the saturation problem of the InfoNCE objective caused by the increased similarity of retrieved samples. InfoLOOB does not suffer from saturation.
کلوب (به انگلیسی: cloob) وبگاهی ایرانی بود که در یکم دی ماه ۱۳۸۳ به عنوان یک شبکه اجتماعی راه اندازی شد و ۱۵ مرداد ۱۴۰۰ به فعالیت خود پایان داد. این تارنما امکان ارتباط ایرانیان جهت تبادل ...
Oct 21, 2021 · We propose to use the InfoLOOB objective to mitigate this saturation effect. We introduce the novel "Contrastive Leave One Out Boost" (CLOOB), which uses modern Hopfield networks for covariance enrichment together with the InfoLOOB objective.
Iran’s oldest social media network Cloob announced Monday it is shutting down after years of battling censors
Oct 16, 2017 · Cloob website was launched 12 years ago as the Iranian answer to Facebook and Google’s now-dead Orkut, and at its peak had some two million users in the country.
CLIP and CLOOB were trained for a comparable number of epochs used in CLIP (see Appendix Section A.2.2) while CLIP* and CLOOB* were trained until evaluation performance plateaued (epoch 128). In both cases CLOOB significantly outperforms CLIP on the majority of tasks or matches its performance.
We introduce “Contrastive Leave One Out Boost” (CLOOB) which combines the “InfoLOOB” objective with modern Hopfield networks. CLOOB overcomes CLIP’s problems of (i) the saturation problem of the InfoNCE objective and (ii) the “explaining away” problem when the covariance structure is insufficiently extracted. Our contributions are:
We propose to use the InfoLOOB objective to mitigate this saturation effect. We introduce the novel "Contrastive Leave One Out Boost" (CLOOB), which uses modern Hopfield networks for covariance enrichment together with the InfoLOOB objective.
CLOOB has two major components: (i) modern Hopfield networks that alleviate CLIP’s problem of insufficiently exploiting the covariance structure in the data and (ii) the InfoLOOB objective that does not suffer from InfoNCE’s saturation problem. The next two sections analyze CLOOB’s major components.
