Our Algorithm Engineer, Henry Qu, and Royal Society Industry Fellow, Dr Jianbo Jiao, recently presented at the International Conference on Computer Vision in Paris, which is considered one of the top conferences in computer vision in the world.
This was an incredible opportunity to expand the knowledge and resources of the Vieunite team. Here is what Henry had to say about his experience at this prestigious event and the project they presented:
“Attending the International Conference on Computer Vision (ICCV2023) was not just a great opportunity but also a monumental milestone in my professional journey. Last week, as I showcased “Multi-view Self-supervised Disentanglement for General Image Denoising,” I was pleasantly overwhelmed by the camaraderie, intellectual rigor, and the exchange of ground-breaking ideas that reverberated through the conference halls.
The uniqueness of our research lies in how we tackled the challenge of denoising. Instead of conventional methods, we ventured into the realm of disentanglement, harnessing the power of self-supervised training on synthetic datasets. To our delight, our methods translated seamlessly to real-world datasets, pushing the boundaries of what was previously achievable. This paradigm shift in approach garnered a lot of attention, making our booth a hub for in-depth discussions and brainstorming.
The sheer diversity of researchers and professionals present at ICCV2023 was staggering. Engaging in discussions with brilliant minds from illustrious institutions like ETH Zürich, Tsinghua University, Nanyang Technological University, Microsoft Research Asia, and Adobe was both enlightening and humbling. Each interaction, whether it was a casual chat over coffee or a structured panel discussion, enriched my understanding and widened my horizon.
Yet, one encounter that will remain etched in my memory is meeting Yann LeCun. The aura around this titan of artificial intelligence is undeniable, but what struck me most was his approachability and willingness to listen. Presenting our work to him and receiving words of encouragement was surreal. His musings about the world models, the potential trajectories AI could take, and the challenges that lie ahead were both fascinating and invigorating. Conversations like these are a reminder of why we delve into research – to continuously push the envelope and envision a future brimming with possibilities.
Another trend that was palpable throughout the conference was the shift in focus. The industry seems to be leaning towards uncovering the foundational aspects of AI. Topics around the role of frequency in training, its impact on model efficacy, and its interplay with adversarial challenges were rampant. Such discussions underscore the fact that we’re striving for a more profound, nuanced understanding of the subject, which is a promising direction.
To encapsulate, ICCV2023 was much more than a conference; it was a confluence of ideas, innovation, and inspiration. I came back not just with business cards and notes but with a renewed sense of purpose, fresh perspectives, and the motivation to forge ahead in the ever-evolving landscape of computer vision.”
Henry Qu is an Algorithm Engineer for Vieunite. His background is in Artificial Intelligence and Machine Learning.