Emerging Work in Open Set Recognition for Vision and Language

• Quem: Walter Schierer
• Onde: FGV - Praia de Botafogo, 190, room 317
• Quando: 28 de Agosto de 2014 às 15:30h

To date, almost all experimental evaluations of machine learning-based recognition algorithms have taken the form of “closed set” recognition, whereby all testing classes are known at training time. A more realistic scenario for real-world applications is “open set” recognition, where incomplete knowledge of the world is present at training time, and unknown classes can be submitted to an algorithm during testing. The first part of this talk introduces the problem of open set recognition, and formalizes its definition as a constrained minimization problem. The open set recognition problem is not well addressed by existing algorithms because it requires strong generalization. As a step towards a solution, the “1-vs-Set Machine” algorithm is proposed, which sculpts a decision space from the marginal distances of a 1-class or binary SVM with a linear kernel. This methodology has been applied to several different applications in computer vision where open set recognition is a challenging problem, including object recognition and face verification.

The second part of this talk looks at a different aspect of the open set recognition problem related to stylometry in computational linguistics. The quantification of style in any text cannot proceed in isolation – all texts must be examined in a broader context where even unknown influences can place constraints upon an author. To address this, a new feature set is introduced to capture sound and lexical information. A learning methodology is then developed for the task of open set influence recognition. The talk will examine a case study related to philological evidence that an eighth century CE Latin poem by Paul the Deacon was influenced by the works of the classical Roman poet Catullus.

Speaker

Walter J. Scheirer, Ph.D. is a research fellow at Harvard University, with affiliations in the School of Engineering and Applied Sciences, Dept. of Molecular and Cellular Biology and Center for Brain Science. He is also an Assistant Professor Adjoint at the University of Colorado Colorado Springs. Previously, he was the director of research & development at Securics, Inc., an early stage company producing innovative computer vision-based solutions. He received his Ph.D. from the University of Colorado and his M.S. and B.A. degrees from Lehigh University. Dr. Scheirer has extensive experience in the areas of computer vision and human biometrics, with an emphasis on advanced learning techniques. His overarching research interest is the fundamental problem of recognition, including the representations and algorithms supporting solutions to it.

Important

• Not required a confirmation.
• FGV not allow the entrance of people wearing shorts and/or slippers.

Tags: