Amazon and UIUC Unveil Exciting Lineup of Funded Research Projects
Introduction:
Amazon and the University of Illinois Urbana-Champaign (UIUC) have announced the launch of the Amazon-Illinois Center on Artificial Intelligence for Interactive Conversational Experiences (AICE). The center supports UIUC researchers and students in developing novel approaches to conversational AI systems. It aims to create intelligent systems that demonstrate contextual understanding, emotional intelligence, personalization, interpretation of nonverbal communication, and ethical and fair practices. Amazon and UIUC have also announced the first cohort of annual fellowships to support the next generation of researchers in the field.
Full News:
Amazon and the University of Illinois Urbana-Champaign (UIUC) have joined forces to establish the Amazon-Illinois Center on Artificial Intelligence for Interactive Conversational Experiences (AICE). This center, located within UIUC’s Grainger College of Engineering, aims to support researchers and students in developing innovative approaches to conversational AI systems.
In an exciting update, Amazon and UIUC have announced the first round of funded research projects and the inaugural cohort of annual fellowships. These projects and fellowships will contribute to the advancement of intelligent conversational systems that exhibit contextual understanding, emotional intelligence, personalization, and the ability to interpret nonverbal communication while adhering to ethical and fair practices.
The academic fellowship recipients, Steeve Huang and Ming Zhong, are part of the prestigious Amazon-Illinois Center on Artificial Intelligence for Interactive Conversational Experiences (AICE). Steeve Huang, a third-year PhD student affiliated with the BLENDER Lab overseen by Amazon Scholar and computer science professor Heng Ji, focuses his research on combating false information. His work encompasses fact-checking, detecting fake news, correcting factual errors, and enhancing the accuracy of text generation models. Huang has developed a zero-shot factual-error correction framework that surpasses traditional supervised methods in generating more faithful and factual corrections. Huang recently completed an internship with Amazon, collaborating with experts Yang Wang and Kathleen McKeown.
Ming Zhong, another third-year PhD student, conducts research in the Data Mining Group under the guidance of Jaiwei Han, the Michael Aiken Chair Professor of computer science. Zhong’s research interests lie in tailoring conversational AI to cater to the diverse needs of individual users as these systems become increasingly intertwined with daily life. Specifically, he aims to improve the understanding of conversational content in both human-to-human and human-to-computer interactions. Zhong also focuses on knowledge transfer across different models to enhance their efficiency.
Now, let’s delve into the exciting research projects of the fellowship recipients:
1. Volodymyr Kindratenko, director for the Center for Artificial Intelligence Innovation and assistant director at the National Center for Supercomputing Applications, is working on a project titled “From personalized education to scientific discovery with AI: Rapid deployment of AI domain experts.” This project aims to develop a knowledge-grounded conversational AI system capable of rapidly acquiring in-depth subject knowledge on a specific topic of interest. Kindratenko proposes a novel factual consistency model that assesses whether answers are supported by verified information sources. The project also introduces a training penalty called factuality loss, which goes beyond traditional cross entropy, and utilizes retrieval-augmented RL with AI feedback to supervise the reasoning process.
2. Yunzhu Li, assistant professor of computer science, is leading a project titled “Actionable conversational AI via language-grounded dynamic neural fields.” Li’s objective is to create multimodal foundational models of the world, leveraging dynamic neural fields. This framework enables several applications, including constructing a generative and dynamic digital twin of the real world, facilitating conversational AI in embodied environments, and empowering embodied agents to plan and execute real-world interaction tasks.
3. Gagandeep Singh, assistant professor of computer science, is undertaking a project called “Efficient fairness certification of large language models.” The project aims to develop an efficient approach to formally certify the fairness of large language models (LLMs). Singh plans to design novel fairness specifications and probabilistic certification methods to obtain certificates that instill greater confidence in the fairness of LLMs compared to current testing-based approaches.
4. Shenlong Wang, assistant professor of computer science, and Romit Roy Choudhury, W. J. Jerry Sanders III – Advanced Micro Devices, Inc. Scholar, and an Amazon Scholar, are collaborating on a project entitled “Integrating spatial perception into conversational AI for real-world task assistance.” The project proposes novel and effective conversational AI workflows that incorporate rich spatial knowledge about users and their surrounding environments gathered from multi-modal sensing and perception.
5. Han Zhao, assistant professor of computer science, is leading a project titled “Responsible conversational AI: Monitoring and improving safe foundation models.” Zhao seeks to develop two new general safety measures: Robust-Confidence Safety (RCS) and Self-Consistency Safety (SCS). RCS aims to ensure that a large language model (LLM) recognizes low-confidence scenarios when dealing with out-of-distribution (OOD) application instances or rare tail events. On the other hand, SCS focuses on ensuring that an LLM remains self-consistent in any context, marking responses as unsafe if they generate logically inconsistent information. These safety measures aim to enhance the reliability and trustworthiness of conversational AI.
The collaboration between Amazon and UIUC shows great promise in advancing the field of conversational AI. By pairing fellows with Amazon scientists, the AICE fellowship program not only supports cutting-edge research but also provides valuable industry insights to aspiring researchers. The interdisciplinary projects funded by the center hold the potential to revolutionize conversational AI, making it more intelligent, empathetic, and responsibly designed.
(End of the article)
Conclusion:
In conclusion, the Amazon-Illinois Center on Artificial Intelligence for Interactive Conversational Experiences (AICE) at the University of Illinois Urbana-Champaign is launching its inaugural round of funded research projects. These projects aim to develop intelligent conversational systems with contextual understanding, emotional intelligence, personalization, and the ability to interpret nonverbal communication while being ethical and fair. Additionally, the center is providing annual fellowships to support the next generation of researchers. These fellows will be mentored by Amazon scientists, gaining insights into industry problems. Some of the awarded fellows and their research projects include Steeve Huang, who focuses on combating false information, and Ming Zhong, who tailors conversational AI to individual users’ needs. AICE is also funding research projects by Volodymyr Kindratenko, Yunzhu Li, Gagandeep Singh, Shenlong Wang, Romit Roy Choudhury, and Han Zhao. The center’s aim is to advance the field of conversational AI and create impactful solutions.
Frequently Asked Questions:
1. What is the significance of the Amazon and UIUC inaugural funded research projects?
The collaboration between Amazon and UIUC for the inaugural funded research projects signifies a commitment to advancing innovative research and development initiatives. It aims to foster groundbreaking ideas and solutions that can make a valuable impact in various fields.
2. How were the research projects selected and funded?
The research projects were selected based on criteria such as their potential for innovation, alignment with Amazon’s and UIUC’s research interests, and their ability to address critical societal or technological challenges. Following a rigorous evaluation process, the chosen projects were then awarded funding to support their execution.
3. Can anyone apply for funding through this collaboration?
No, the funding for this collaboration is limited to specific research projects identified jointly by Amazon and UIUC. However, both entities actively support various other research and funding opportunities that researchers and scholars can explore.
4. What are some of the research areas covered in these projects?
The research projects cover a wide range of areas, including but not limited to artificial intelligence, machine learning, natural language processing, robotics, computer vision, data science, and sustainable technologies. These projects aim to contribute to advancements in these fields, benefiting society as a whole.
5. Will the outcomes of these research projects be publicly available?
While the specific details may vary for each project, both Amazon and UIUC encourage knowledge sharing and dissemination. The outcomes of the projects may be made available through various channels such as academic publications, conferences, or publicly accessible repositories.
6. How long is the duration of these research projects?
The duration of each research project can vary, depending on the nature, complexity, and goals of the specific project. Typically, research projects of this nature span several months to multiple years to ensure comprehensive exploration and conclusive findings.
7. Are there future plans to continue funding research projects through this collaboration?
Amazon and UIUC are committed to fostering ongoing collaboration and research initiatives. While specific plans for future funding are not outlined, both entities aspire to continue supporting innovative research endeavors that align with their shared objectives.
8. How can researchers and scholars in other institutions collaborate with Amazon and UIUC?
Researchers and scholars from other institutions can explore various avenues for collaboration with Amazon and UIUC, including joint research programs, partnerships, and participation in relevant conferences or workshops. Building connections and networking within the research community can be instrumental in identifying collaboration opportunities.
9. What are the potential impacts of these funded research projects?
The funded research projects have the potential to catalyze significant advancements in their respective fields. They may lead to the development of innovative technologies, solutions, and frameworks that can address real-world challenges. Additionally, these projects contribute to knowledge generation and the overall growth of scientific and technological understanding.
10. How can the general public stay updated on the progress and outcomes of these projects?
Both Amazon and UIUC are committed to transparency and knowledge dissemination. Interested individuals can stay updated on the progress and outcomes of these projects through official communication channels, research publications, press releases, and relevant events organized by Amazon, UIUC, or the researchers themselves.
