Brain-controlled robots, AI that predicts recipes, and genius grants

17 in 2017

This past year MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) was at the forefront of many diverse technological innovations covering a breadth of topics, from healthcare and cybersecurity to self-driving cars.

Shape-shifting robots, tone-detecting AIs, and origami-inspired artificial muscles are just a few of the numerous inventive projects that came out of the lab this year. Here are 17 of the highlights from 2017.

Getting inside the robot mind

1. MIT scientists created a feedback system that enables people to correct robot mistakes using brain signals.

2. Researchers created a VR system that lets users teleoperate robots using an Oculus Rift.

3. A shape-shifting robot called “Primer” walks, rolls, sails, and glides using recyclable exoskeletons.


AI for breast cancer, detecting emotions, and recipes

4. AI system uses machine learning to predict if lesions will upgrade to cancer at surgery, helping reduce false positives.

5. Wearable AI system detects a conversation's tone using audio and vital-sign data and serves as a “social coach.”

6. “Pic2Recipe” AI system, developed with the Qatar Computing Research Initiative (QCRI), uses a deep-learning algorithm to recommend ingredients and recipes based on still images of food.

Enhancing online security

7. A team from MIT creates an encryption system that disguises users’ database queries to  reveal no private information.

8. MIT researchers write a cybersecurity report for the Trump administration urging a plan that warns of hacking risk to electric grid, oil pipelines, and other critical infrastructure.                                                                           

9. New system uses Bitcoin’s security machinery to defend against online identity theft.

Muscles, models, and motion-sensing

10. Origami-inspired artificial muscles let soft-robotic structures lift objects up to 1,000 times their own weight using only air or water pressure.

11. Machine learning model can predict outcomes in Intensive Care Units (ICU), while another system uses natural language processing to keep data consistent across different Electronic Health Records (EHRs).

12. “WiGait” device can measure walking speed using wireless signals to help monitor and diagnose health issues like cognitive decline and cardiac disease.

Turing awards, genius grants, and visits from Eric Schmidt

13. MIT professor Sir Tim Berners-Lee, the researcher who invented the World Wide Web, who’s also an influential advocate for online privacy, wins the prestigious ACM A.M. Turing Award.

14. MIT professor Regina Barzilay, who studies natural language processing and machine learning, wins MacArthur “genius grant” for her work in computational linguistics.

15. Eric Schmidt visits MIT to discuss computing, artificial intelligence, and the future of work.

Improving neural nets

16. Researchers come up with a method for producing real-world 3-D objects that can consistently fool neural networks.

17. A team from CSAIL and QCRI develop a technique to analyze how neural networks process language.

For updates on the latest lab news, follow CSAIL on Facebook, Twitter, and Instagram.