MIT Laboratory for Information and Decision Systems (LIDS)

Can #LLM’s replace classical machine learning techniques for ML tasks? MIT LIDS PI Kalyan Veeramachaneni and Sarah Alnegheimish share their recent findings in VentureBeat, along with thoughts on best practices and guardrails for #AI. Read the article: https://bit.ly/3BZTqVV

Artificial intelligence is a top priority for businesses everywhere as leaders start to realize its transformative value in improving how their companies operate, make decisions, and deliver results.   To truly reap the benefits of AI, especially generative AI, executives must embrace the need to develop a deep understanding of the technology, its use cases, ethics, and more, according to Eric So, an MIT Sloan professor of global economics and management. He and Sertac Karaman, an MIT professor of aeronautics and astronautics and director of the MIT Laboratory for Information and Decision Systems (LIDS), lead the AI Executive Academy, a new MIT Sloan Executive Education course offered jointly by MIT Sloan and the MIT Schwarzman College of Computing that looks at the technical and business aspects of AI and its impact across industries.   At a recent webinar about the potential of AI for executives, So and Karaman discussed how executives can navigate the new technological landscape, including six ways leaders can embrace AI.

A T-RO paper from researchers at MIT Laboratory for Information and Decision Systems (LIDS) and Industrial & Enterprise Systems Engineering at University of Illinois Urbana-Champaign have proposed a hybrid system stability analysis framework that offers a principled explanation to the traffic breaks (rolling roadblocks) emergent behavior for a single autonomous vehicle to simultaneously stabilize multiple lanes. https://lnkd.in/ga9BXTVY Paper title: Hybrid System Stability Analysis of Multi-Lane Mixed-Autonomy Traffic Authors: Sirui Li, Roy Dong, and Cathy Wu #IntelligentTransportationSystems #AutonomousVehicle #TrafficBreak #HybridSystem #StabilityAnalysis

In a 5Q with the Center for Data Innovation, LIDS PI Devavrat Shah talks about the future of #AI and how his MIT spinout, Ikigai Labs, helps companies apply AI to more effectively leverage their data. Read the 5Q: https://bit.ly/48bgYTS

Congrats to LIDS PI Marija Ilic, recipient of the 2024 IEEE PES Prabha S. Kundur Power System Dynamics and Control Award! Acknowledged specifically for her pioneering contributions to hierarchical and distributed modeling and control of large power systems, this award recognizes Ilic’s outstanding contributions to the understanding and control of the dynamics of the power system. Read more: https://bit.ly/488rNWD

For those interested, I will be sharing work presented at the 2024 IEEE ITSC on rural electromobility in Iceland at the MIT Laboratory for Information and Decision Systems (LIDS) Climate Tea Talk tomorrow, from 4:00-4:30 PM ET. I will go over the CTEC simulator we created and the results from the Icelandic case study that I detailed in my post from about a week ago. Consider stopping by!

Helping robots zero in on the objects that matter LIDS PI Luca Carlone, together with LIDS postdoc Lukas Schmid and graduate students Yun Chang, Nathan Hughes, and Dominic Maggio, and MIT Lincoln Laboratory collaborators, have developed a new method called Clio that enables robots to quickly map a scene and to make intuitive, task-relevant decisions. Learn more at MIT News and read their paper in the journal Robotics and Automation Letters: https://bit.ly/3Bxhnnq MIT AeroAstro MIT Schwarzman College of Computing MIT School of Engineering

Should we label #AI systems like we do prescription drugs? AI systems are increasingly being deployed in safety-critical health care situations. Yet these models sometimes hallucinate incorrect information, make biased predictions, or fail for unexpected reasons, which could have serious consequences for patients and clinicians. In Nature Computational Science, researchers Marzyeh Ghassemi and Elaine Nsoesie argue that in healthcare settings, “responsible use” labels could ensure AI systems are deployed appropriately. In a recent 3Q, LIDS PI Marzyeh Ghassemi discusses the need for such labels, the information they should convey, and how labeling procedures could be implemented.

Could AI lead to inconsistent outcomes in home surveillance? A new study from LIDS researchers Ashia Wilson and Shomik J. with MIT Media Lab alumna Dana Calacci PhD ’23, shows that #AI makes inconsistent decisions about when to call the police when analyzing surveillance videos. https://bit.ly/4ez740n

Congrats to LIDS PhD candidate Brice Huang, recipient of the Machtey Award for Best Student Paper at FOCS 2024 for his paper, “Capacity threshold for the Ising perceptron.” Learn more and read the paper: https://bit.ly/3B3UiII MIT Department of Mathematics MIT EECS MIT School of Engineering MIT Schwarzman College of Computing