Watch-and-Help: A Challenge to Human-AI Collaboration, Social Perception and Social Perception

Watch-and-help: a multi-agent environment where an AI agent must assist a human-like agent in performing a complex household task. Both agents must understand the task’s underlying goal and coordinate to achieve it. This multi-agent household environment, called VirtualHome-Social, provides learning-based baselines and benchmarks for designing collaborative plans.

Elizabeth Arredondo, a writer, focuses on creating compelling characters that allow for human-machine interaction. She is currently developing the backstory and personality of Mabu, a robotic wellness coach. She received her MFA from USC School of Cinematic Arts in 2005 and was selected for a feature-film writing fellowship. She was also a staff writer for CBS’s COLD CAUSE and participated in the NBC writers-on-the Verge program. She holds a Master’s degree in Computer Science from the University of Southern California and an MBA from the University of Southern California.

Elizabeth Arredondo is the author and VP of User Experience for Sensely, a virtual wellbeing coach. She has a Master of Fine Arts in Screenwriting from the USC School of Cinematic Arts. She also participated in the NBC Writers on the Verge program. In addition to her work with Sensely, she has worked with a number of big brands and has written a conversational iPad app for the company.

In addition to designing engaging user experiences for watch-and-help chatbots, Ajay Chander is leading R&D teams for human-centric technologies, such as artificial intelligence, robotics, and digital health. He is also the director of the Digital Life Lab at Fujitsu Labs of America where he develops solutions to keep humans informed.

Cathy Pearl is the author and VP of User Experience at Sensely. Sensely, her company, is developing virtual nurses to improve healthcare efficiency and effectiveness. For 15 years, she has been working with major brands and was a senior design at Microsoft. She has a master’s degree in Computer Science from the University of California, San Diego, and Indiana.

Among the many challenges in social perception and human-ai collaboration is watch-and-help. DeepMind CEO William Isaac explains how these systems will help people in their daily lives and make them more productive. But what exactly is a “watch-and-help” bot? How can we recognize the intention of a robot?

Social perception and human-ai collaboration face challenges due to the need for effective interaction between humans, machines, and humans. The avatar of Sensely’s virtual nurse avatar communicates with patients and helps them navigate a medical situation. A Sensely virtual nurse, Mabu, is making healthcare more effective. In addition to her work for Sensely, she has previously worked at Microsoft and Nuance.

Cathy Pearl, founder of Sensely (a company that assists healthcare users to get the most from their interactions with its virtual nurse), discussed the potential for watch-and-help in social and human-ai collaboration. During her time at Sensely she created customer-centric solutions that were used by major companies such as Walgreens, Microsoft, Google, and Microsoft.

As a writer, Cathy Pearl is focusing on developing compelling characters for a human-ai collaboration. Sensely, a company that provides virtual nurses to make healthcare more efficient, is also her creation. Previously, she worked at Microsoft, Nuance, and Volio, where she led design and development efforts for a conversational iPad app.

Watch-and-help, in a similar manner to a social agent’s assistance, can support human interaction. A human agent can help with difficult speech input, such as by assigning transcription or semantic interpretation. A human agent can also help with live feedback. The use of a robot assistant in a human-ai collaboration has the potential to make human-ai interactions more effective and more personalized.

Watch-and-Help: A Challenge to Human-AI Collaboration, Social Perception and Social Perception
Scroll to top
%d bloggers like this: