Amazon and MIT have appointed the item in the 2023 gift project and the Fellowship Awards Grantd as part of the Science Hub.
Science Hub-Administer at MIT by Schwarzman College of Computing-is a collaboration between Amazon and MIT, which leading research supports, education and abroad in the development of technologies such as machine learning and robotics.
“This year’s Science Hub projects address some of the most challenging problems in machine learning and robotics. OU scientists are for a hard look at receiving that hat appeared, includes them with autonomous systems, large language models and data and practical solutions,” said Sertac Karaman, MIT -Leader of Science Shub Aerona and Astrona. “I am also excited to welcome our new fellows, which will lead their independent research projects within AI and Robotics.
“We are pleased to support research in the next generation’s big-Langa model Technologies,” said Shiv Vitaladevuni, director of Applied Science in Alexa. “These gifts will support the development of more reliable LLMs and utilize LLMs for data curation. Utilization of LLMs for datauration will increase the developer’s productivity to summarize the content of databases and address from such as noisy, lack of or party data. Back to Academy to long -term benefit to the community.”
The research projects the goals of Huben, which is intended to ensure that the benefits of new technology are largely shared through education and research, as well as expanding participation in research from a wide range of scholars and other innovators.
“We are obliged to support the best robotics and machine learning science,” said Jeremy Wyatt, director of Applied Science at Amazon Robotics. “These gifts will support breakthrough Tork in the use of machine learning to robotics.
A selection of members from both Amazon and MIT chose the four research projects described below.
“Online Security Monitoring For AI-Enabled Robot Car”- Navid AzizanEsther & Harold E. Edgerton Career Development Assistant Professor
“Deep neural networks in autonomous systems, such as robots, can be unreliable on inputs that are significantly different from their training data set – shape occurrence when a robot encounter that an anxiety objects that block their path or interact with a human being who deviates from the robot’s expectation. Such scenarios can lead to breeders or jumps in operation. System.
“Control of large language models with symbolic structures” – Yoon KimProfessor of Electrical Engineering and Computer Science
“Despite their impressive capabilities, large language models are still difficult to control. This proposal seeks to develop mechanisms to control LLMS through symbolic grammar. Given a lured model defines our approval a probabilistic grammar whose non-symbols are explicitly related to lifting the model over partial output Semantic parsing, translation of medical notes).
“Datacore: A Basic Model of Business Datakuration” – Samuel MaddenMy College of Computing Distinguished Professor of Computing
“Modern data -intensive applications are characterized by a need to combine and query a number of data sets, from internal documentation, to partially structured data such as logs, to table data in databases, to machine learning states and broadcast with advanced or better performance, which, like chatgpt, does not require a great love for domain security, but may be interactive.
“Activating memory-efficient sludge for energy-limited devices”- Vivienne SzeAssociate Professor of Electrical Engineering and Computer Science
“We offer co-design algorithms and hardware for simultaneous location and mapping (sludge), which are efficient, robust and accurate all at the same time. Energy-limited devices such as smartphones, AR/VR headset and pill-sizing medicine robots are set to make sets of applications. (Mapping).
2023 Scholarships
The following doctorate students will receive funding to pursue independent research projects in robotics and AI. Students have an appropriate to participate in paid summer internships at Amazon, where they can work directly with Amazon researchers to gain valuable industrial insight and experience.
Sirui Li,,,,,,,, PhD. -Scordable, social and technical systems and statistics
Li received his bachelor in computer science and math from Washington University in 2019 and was one of the five classes in Védictori’s 2019 from Washington University School of Engineering. LI is particularly interested in Bayesian methods, social networks and uses of machine learning to economy and political science. Li collaborates with Cathy Wu, Gilbert W. Winslow Career Development Assistant Professor, Civil and Environmental Technology, by solving the riding sharing problem using graph neural networks and reinforcement learning. LI’s long-term research target is to improve algorithms/heuristics that solve severe social problem in urban planning and mechanism design.
Yue Meng,,,,,,,, PhD. -Candidate, aviation and astronautics
Meng is a fourth-year PhD student at Mit Aeroastro, who works in the reliable autonomous system laboratory. Meng’s research topic uses machine learning techniques for secure and robust robotic control. Before it was Meng an AI resident of IBM Thomas J. Watson Research Center. He won a master’s degree in electric and computer technology at the University of California, San Diego, and received a Bachelor of Science from Tsinghua University in the Department of Automation.