IR and friends
Lexing Xie (ANU)
Monday 19 August 2013Retrieving relevant videos from a large corpus on mobile devices is a vital challenge. We address two key issues for mobile search on user-generated videos. The first is the lack of good relevance measurement, due to the unconstrained nature of online videos, for learning semantic-rich representations. The second is due to the limited resource on mobile devices, stringent bandwidth, and delay requirement between the device and the video server.
We propose a knowledge-embedded sparse projection learning approach. To alleviate the need for expensive annotation for hash learning, we investigate varying approaches for pseudo label mining, where explicit semantic analysis leverages Wikipedia and performs the best. In addition, we propose a novel sparse projection method to address the efficiency challenge. It learns a discriminative compact representation that drastically reduces transmission cost. With less than 10% non-zero element in the projection matrix, it also reduces computational and storage cost.
The experimental results on 100K videos show that our proposed algorithm is competitive in the performance to the prior state-of-the-art hashing methods which are not applicable for mobiles and solely rely on costly manual annotations. The average query time on 100K videos consumes only 0.592 seconds. This is joint work with Guan-Long Wu, Winston Hsu and others in National Taiwan University....
Friday, August 16, 2013
Finding Videos on Mobile Devices
Lexing Xie will speak on "Scalable mobile video retrieval with sparse projection learning and pseudo label mining" at the CSIRO on the Australian National University campus in Canberra, 4pm, 19 August 2013.