Company News

Paper accepted at NIPS2017 and another paper at CoR2017

September 6, 2017 | Project

Panasonic R&D Center Singapore, collaborating with Assistant Professor Feng Jiashi of the National University of Singapore (NUS), co-authored a paper1 titled “Dual-Agent GANs for Photorealistic and Identity Preserving Profile Face Synthesis” that has been accepted at the Conference and Workshop on Neural Information Processing Systems (NIPS) 2017, a machine learning and computational neuroscience conference. This year, NIPS, a top-rated annual conference with a 22% acceptance rate in most cases, will be held in Long Beach, USA, from December 4th through December 9th, 2017.

Moreover, Panasonic R&D Center Singapore, in collaboration with Professor David Hsu of NUS, co-authored a paper2 titled “Intention-Net: Integrated Planning and Deep Learning for Autonomous Navigation” that has been accepted at the Conference on Robot Learning (CoRL) 2017, a new annual international conference focusing on the intersection of robotics and machine learning. This year’s CoRL will be held in Mountain View, California, USA, from November 13th through November 15th, 2017. Despite being its first-ever event, CoRL has invited a lot of interest from reputable sponsors (with Panasonic being one of the Gold sponsors), famous speakers (such as Yann LeCun, a computer scientist well known for Deep Learning) as well as keen attendees in the fields of robotics, autonomous perception and control.

The publication of the two papers upholds Panasonic R&D Center Singapore’s mission to promote Panasonic’s technologies at top conferences of respective fields.

<Supplemental information>

paper¹: proposes a Dual-Agent Generative Adversarial Network (DA-GAN) model which can improve the realism of a face simulator’s output using unlabelled real faces, while preserving the identity information during realism refinement. The dual agents are specifically designed for distinguishing real vs. fake and identities simultaneously.

paper²: proposes a single unified framework that connects model-based planning and deep learning for autonomous navigation, via an interface named intention to encode where robots would go in the near future.