Creating and sharing knowledge for telecommunications

Integration of hard and soft classification using SVM

on 16-10-2015

... Date: 16/10/2105 (next Friday), 14H00
Room: LT2 (Torre Norte, Piso 4)

Speaker: Xiuping Jia (University of New South Wales, Australia)

Hard classification by labeling each pixel in a multispectral or hyperspectral image as one of the defined classes may present high uncertainty. One of the reason is that there are many mixed pixels due to limited spatial resolution. Spectral unmixing techniques have been developed to map each pixel as a weighted sum of the endmember classes it contains. While Support Vector Machine has been widely used for hard classification, it has been extended for spectral unmixing. In this talk, an integrated hard and soft classification using SVM is presented, including the concept of unmixing resolution and fuzzy assessment. More Information..

Quantum Machine Learning

on 09-10-2015

... Masoud Mohseni (Google Quantum Artificial Intelligence Lab)
Friday 09/10/2015, at 16:30 Physics Seminar Room (2.8.3), Physics Building, IST
Seminar via video-conference. Please note exceptional time and place.

Over the past 30 years, two diferente computational paradigms have been developed based on the premise that the laws of quantum mechanics could provide radically new and more powerful methods of information processing. One of these approaches is to encode the solution of a computational problem into the ground state of a programmable many-body quantum Hamiltonian system. Although, there is empirical evidence for quantum enhancement in certain problem instances, there is not a full theoretical understanding of the conditions for quantum speed up for problems of practical interest, especially hard combinatorial optimization and inference tasks in machine learning. In his talk, I will provide an overview of quantum computing paradigms and discuss the progress at the Google Quantum Artificial Intelligence Lab towards developing the general theory and overcoming practical limitations. Furthermore, I will briefly discuss two recent quantum machine learning primitives that we have developed known as Quantum Principal Component Analysis and Multiqubit Quantum Tunneling.

Joint session with the Physics of Information Seminar.

Quantum Computation and Information Seminar hp.en?action=next

Support: Phys-Info (IT), SQIG (IT), CFIF and CAMGSD, with support from FCT, FEDER and EU FP7, namely via the Doctoral Programme in the Physics and Mathematics of Information (DP-PMI), and projects PEst-OE/EEI/LA0008/2013, QuSim, QUTE-EUROPE (GA 600788), Landauer (318287) and PAPETS (323901). More Information..