Q&A with Salesforce Research Intern Kevin (Chih-Yao) Ma on Self-Monitoring Navigation Agent via Auxiliary Progress Estimation

By: Alexandria Murray

For some, the world we live in today can be represented as data in high-dimensional spatiotemporal space—which we humans typically use language to describe, interpret, and reason about. For Salesforce Research Intern, Kevin (Chih-Yao) Ma, this topic became a key focal point in his research “Self-Monitoring Navigation Agent via Auxiliary Progress Estimation” which he collaborated on with co-authors Jiasen Lu, Zuxuan Wu, Ghassan AlRegib, Zsolt Kira, Richard Socher (Chief Scientist), and Caiming Xiong (Senior Director of Research), during the course of his internship.

This research was one of 6 papers from the Salesforce Research Team recently accepted at the Seventh International Conference on Learning Representations (ICLR).

In the below Q&A, Ma explains why he chose to focus on this, and the benefits it could have for us all.

What is your primary focus area?
My research focus area is the intersection of Computer Vision and Natural Language Processing, where we want to develop visual systems to understand the world we see as well as develop language systems to interpret and reason about the captured visual content.

Tell us about your ICLR research.
We developed an AI agent that can navigate in houses (in virtual environments) by following human navigational commands in natural language format.

How might your research benefit the average human?
We demonstrate the possibility of developing an AI agent that can understand and follow natural language commands to perform a certain task. This has a huge impact on human society in the future, where we are very likely to have robot assistants in every single house for things like "get a glass of water", "take the trash outside", etc.

What was your reaction upon finding out your research had been selected for ICLR?
I was having dinner when my phone went off, with a number of messages from my co-authors. I then received the phone call directly from Caiming sharing this wonderful news. Suddenly, the painful process during the OpenReview rebuttal all became worth it.

What would you tell someone interested doing a PhD internship with the research team at Salesforce?
Salesforce is definitely one of the best places to intern at. I was very lucky to have the guidance from a number of researchers as well as to have insightful discussions with other interns. I also got to know many people related to my field. In the end, I was fortunate to have one ICLR paper and one CVPR oral paper (extension of our ICLR paper) accepted.

Learn more about our Salesforce Research Team and how you can have an impact as a PhD intern at https://einstein.ai/careers.