Fostering Innovative Talents through Partnership with University of Toronto’s Applied AI Program

[News Reporter]

LG is rooted in the philosophy of open innovation – an open approach to innovation and facilitating collaboration to create a better life for all. These collaborative partnerships involve LG affiliate companies and extend to government, universities and research labs, with the goal to establish an open ecosystem of innovation that will benefit and empower all involved.

Recently, the LG Toronto AI Lab welcomed 30 graduate students from various universities in South Korea to the

link hidden, please login to view

(CARTE) at the University of Toronto. The event – initiated and organized by the Toronto AI Lab in collaboration with CARTE – brought these talented students together to explore applied research partnerships in Canada and introduce them to career opportunities at LG.

Sponsored by the South Korean government and managed through the University of Toronto’s Faculty of Applied Science & Engineering, the 6-month CARTE program accommodates students from a wide range of disciplines, including technology, engineering, and the natural and mathematical sciences and aims to inspire students to leverage AI innovatively in their respective fields.

The selection for this program was competitive, drawing participants from among South Korea’s brightest. These selected students will engage in 10 industry projects, including three spearheaded by the LG Toronto AI Lab.

Kevin Ferreira, Senior Director and Head of the LG Toronto AI Lab, shared, “Our collaboration with the University of Toronto and the South Korean government on the CARTE program aims to enhance our research capabilities and strengthen our talent pool. This event enabled us to present the extensive work undertaken at the LG Toronto AI Lab to an exceptional group of Korean students, laying the groundwork for deeper research collaboration.”

Showcasing some of the Toronto AI Lab’s recent work, Manasa Bharadwaj, Senior AI Research Scientist, presented on behalf of the LG Toronto AI Lab’s Conversational AI team and their research into Responsible Conversational AI. The talk covered the motivation, technical details and results of the team’s 2023 published research titled ‘COUNT’ and ‘GPT-Detox’. The two paraphrasing approaches leverage unlikelihood training and prompt-based approaches respectively to convert unsafe text into safe text.

Mathieu Tuli, AI Research Scientist, presented on the work that the LG Toronto AI Lab Digital Human team conducts, highlighting FlowFace, the team’s recent CVPR) 2024 accepted paper. The talk provided detail on why FlowFace is an important innovation for the community at large, and how it affects the rest of the Digital Human pipeline.

Thi Ha Kyaw, Staff AI Research Scientist, presented an introduction to quantum computing and quantum algorithms in general, and why quantum is worth pursuing in the current quantum utility era. He also presented his team’s recent result on super-exponential catastrophe of quantum algorithms in noisy quantum hardware.

During their time in the program, students will participate in graduate-level intensive courses in Machine Learning and AI, delivered by the University of Toronto’s Department of Mechanical & Industrial Engineering. They will also benefit from tailored AI support services provided by CARTE, including research and training in AI for academic and industry partners.

Furthermore, students will have the opportunity to attend several applied AI seminars at the University of Toronto and engage with AI and machine learning projects across various sectors, including industry, government, and non-profits. A dedicated workspace will also be available to them, fostering collaboration and allowing them to apply their knowledge and skills within Toronto’s thriving AI ecosystem.

The program will conclude in late June, where selected students will present their projects to industry partners, showcasing their achievements and the potential for future innovations in AI.

Contributed by LG Canada

# # #