MIT LIDS (@MITLIDS) header
MIT LIDS (@MITLIDS) profile picture

LIDS is an interdepartmental research lab in @MIT_SCC Affiliations include @MITEECS @MITAeroAstro @MITMechE @MIT_CEE @ORCenter @MITIDSS @MITSloan @eapsMIT

Cambridge, MA
Joined May 2024
Helping AI text detectors keep up with changing conditions. @manish_raghavan, @NikhGarg, and Kevin Ren show how test-time adaptation can help detectors identify AI-generated writing, even as new models emerge or text is disguised to appear human-written. bit.ly/48n7Igm
70
Why do robots need both classical engineering and AI? @MITLIDS and @MIT_CEE's @GioeleZardini co-authored "Principles of Robot Autonomy," w/ @MIT_alumni @drmapavone PhD ‘10, Daniele Gammelli, Joseph Lorenzetti, and Katie Lu. Read a 3Q and explore the book: bit.ly/4AvpRoa
1
5
307
Can robots learn to correct their own fuzzy memories? UQ-DAAAM, a new system from @MITLIDS students and PI @lucacarlone1 identifies and corrects unreliable object descriptions in robot memory, helping robots perform complex tasks more reliably. bit.ly/3WS7MC5
1
3
963
Five LIDS PIs named 2026 #AI2050 Fellows! Congrats to Senior Fellow @m_sendhil and Early Career Fellows @alex_damian_, @baileyflan, @Micro_Yunha, and @manish_raghavan! The @schmidtsciences initiative supports research tackling #AI's toughest challenges. bit.ly/4Ai7gfj
1
5
18
2,137
Congrats! LIDS researchers Kaveh Alimohammadi, @NavidAzizan, and collaborators, have received the 2026 MIT Prize for Open Data for their work on generating scientifically useful synthetic data while protecting individual privacy. bit.ly/4ixDiN5 @ScienceMIT @mitlibraries
1
6
418
Gunter Stein, a pioneering control systems engineer and longtime @MITLIDS researcher, has died at 84. His work helped shape modern control theory, while his influential essay “Respect the Unstable” emphasized engineers’ responsibility for safety. MIT News: bit.ly/4e40zVv
4
142
How AI came to model the “messy” human mind Benjamin Lindquist explores how engineers introduced randomness into early AI systems, laying the groundwork for today’s deep-learning models—and their tendency to hallucinate. Read the Q&A: bit.ly/3VRrGgi
1
126
How can drones safely navigate unknown environments with moving obstacles? @MITLIDS researchers developed SANDO, a system that plans efficient flight paths with mathematical guarantees against collisions. The tech could advance disaster-response robotics. bit.ly/4jQxIri
5
139
What happens when companies use the same #AI to screen job applicants? A new study from @MITLIDS PIs Brian Hedden and Manish Raghavan finds that “algorithmic monoculture” isn't always harmful to job seekers — and may even benefit them in some cases. More: bit.ly/4hSEFpg
1
2
194
How can AI help improve the complex systems that shape everyday life? @MITLIDS PI and associate prof @MIT_CEE and @mitidss, @wucathy uses Reinforcement Learning to develop safer, more efficient transportation systems. Read her @MITNews profile: bit.ly/4AStaGR @MIT_SCC
5
133