 |
SINGAPORE, June 22, 2020 /PRNewswire/ -- Kantar, the world's leading data, insights and consulting company, announced today the first patent on Quantum Machine Learning as part of AI/ML advancement in Singapore. With the continuous support and partnership of Singapore's Economic Development Board (EDB), Kantar established its Brand Growth Lab in Singapore in 2018 to develop AI/ML solutions. The Lab, an advanced analytics hub, is dedicated to discovering new ways to leverage big data to drive strategic decision-making for business.
Kantar awarded patent in the Quantum AI field
On January 2nd of this year, Kantar was granted its first patent by the Intellectual Property Office of Singapore for a method of optimising AI/ML predictions from a classical data feed with a hybrid simulator generated from classical and quantum model structures. Some of the other organizations with a patented invention in the Quantum technology field in Singapore are Oxford University Innovation, D-Wave, IBM and Google.
"Quantum technology will revolutionize Artificial Intelligence and Machine Learning. This patent indicates our commitment to lead in this field. We are proud to have been awarded this patent as it demonstrates our advancement in the field of data science," commented Hernan Sanchez, Managing Director, Kantar Brand Growth Lab.
Kantar Brand Growth Lab experiments in the Quantum Machine Learning field
In collaboration with Professor Angelakis, Principal Investigator and the leader of the Quantum Simulation and Computation Group at the National University of Singapore's Centre of Quantum Technologies, two quantum experiments using real consumer behavioral data from Kantar's panels were conducted during the last 6 months.
Experiment 1: Customer segmentation using quantum machine learning
In the first experiment, the goal was to develop quantum-inspired machine learning segmentation algorithm that exploits the concept of quantum interference and work in the classical hardware to improve the results of the current machine learning approach. Comparing the traditional technique with a quantum and a quantum genetic algorithm developed for this experiment, we observed that the Quantum versions showed better results.
Experiment 2: Customer Segmentation based on Media consumption patterns using an IBM quantum computer
The objective of this experiment was to establish the feasibility of using a quantum computer to address a real-life segmentation problem. The quantum algorithm was run using 2 qubits on the IBM's 5-qubit quantum computer. 4 relevant consumer segments were identified. Next steps in our research will be about proving the superiority of the quantum approach and explore the potential of more advanced quantum hardware.
"In the field of Data Science, it is always crucial to be aware of the new techniques and methodologies in order to stay relevant and have that competitive edge. By venturing into Quantum field early and experimenting with different Quantum machine learning techniques, we hope to have an early mover advantage that could bring great business value in the long run," commented Shilpa Jain, Principal Data Scientist, Kantar Brand Growth Lab.
"With the support of EDB, the Lab keeps on researching and developing advanced analytics solutions to help institutions and corporations maximize the productivity of their efforts. The post COVID-19 economy requires a new level of innovation and in today's data-driven economy, AI will play a key role," added Yee Mei Chan, co-managing director of the Kantar Brand Growth Lab.
About Kantar
Kantar is the world's leading evidence-based insights and consulting company. We have a complete, unique and rounded understanding of how people think, feel and act; globally and locally in over 90 markets. By combining the deep expertise of our people, our data resources and benchmarks, our innovative analytics and technology, we help our clients understand people and inspire growth.
Further information about Kantar can be found at www.kantar.com
Logo - https://photos.prnasia.com/prnh/20190611/2492962-1logo?lang=0
