Toward deep-learning models that can reason about code more like humans

Researchers propose a method for finding and fixing weaknesses in automated programming tools.

A robot that senses hidden objects

System uses penetrative radio frequency to pinpoint items, even when they’re hidden from view.

Faster drug discovery through machine learning

New technique speeds up calculations of drug molecules’ binding affinity to proteins.

Algorithm helps artificial intelligence systems dodge “adversarial” inputs

Method builds on gaming techniques to help autonomous vehicles navigate in the real world, where signals may be imperfect.

Facebook’s New AI Teaches Itself to See With Less Human Help

Quest Corporate director Aude Oliva comments on a new method for training image-recognition models that eliminates the need for most labels.

A language learning system that pays attention — more efficiently than ever before

MIT researchers’ new hardware and software system streamlines state-of-the-art sentence analysis.

“Liquid” machine-learning system adapts to changing conditions

The new type of neural network could aid decision making in autonomous driving and medical diagnosis.

Design progresses for MIT Schwarzman College of Computing building on Vassar Street

New building will create a hub for computing research and education at MIT, including spaces designed to be inviting to members of the campus community and the public.

Designing customized “brains” for robots

A new system devises hardware architectures to hasten robots’ response time.

An intro to the fast-paced world of artificial intelligence

In two years, the MIT Quest for Intelligence has allowed hundreds of students to explore AI in its many applications.

Model analyzes how viruses escape the immune system

Using this computational system, researchers can identify viral protein sequences that could make better vaccine targets.

Want cheaper nuclear energy? Turn the design process into a game

Researchers show that deep reinforcement learning can be used to design more efficient nuclear reactors.

AI Algorithms Are Slimming Down to Fit in Your Fridge

Artificial intelligence programs typically are power guzzlers. New research from Song Han's lab at MIT shows how to generate computer vision from a simple, low-power chip.

Model could help determine quarantine measures needed to reduce Covid-19’s spread

A key finding: Early reopening last spring led to a dramatic drop in “quarantine strength” in southern and west-central U.S. states.

IBM and MIT researchers find a new way to prevent deep learning hacks

In a new study, researchers from MIT and the MIT-IBM Watson AI Lab explore whether the human brain can offer clues on how to make deep neural networks even more powerful and secure. Turns out it can.

Researchers fit more AI than ever onto IoT microchips

Researchers from the MIT-IBM Watson AI Lab unveil a technique for squezing more AI than ever onto the simple chips that power connected devices, from medical wearables to coffeemakers.

Shrinking massive neural networks used to model language

A new approach could lower computing costs and increase accessibility to state-of-the-art natural language processing.

MIT forum examines the rise of automation in the workplace

Technologies like robots and artificial intelligence could partner with humans, not oust them from work, research and business leaders say.

Why we shouldn’t fear the future of work

MIT task force wraps up with a final conference, sounds note of optimism that new ideas, better policies can help sustain good careers.

Vibrations of coronavirus proteins may play a role in infection

Study suggests mechanical properties of spike proteins can predict infectivity and lethality of different coronaviruses.

Using machine learning to track the pandemic’s impact on mental health

Textual analysis of social media posts finds users’ anxiety and suicide-risk levels are rising, among other negative trends.

Artificial intelligence model detects asymptomatic Covid-19 infections through cellphone-recorded coughs

Results might provide a convenient screening tool for people who may not suspect they are infected.

Testing sewage to home in on Covid-19

Letting an algorithm decide which maintenance holes to test for evidence of coronavirus could improve pandemic containment efforts.

Stressed on the job? An AI teammate may know how to help

Researchers are working toward intelligent machines that can sense cognitive fatigue and suggest interventions to help a human improve performance.

Bringing construction projects to the digital world

The startup OpenSpace is using 360-degree cameras and computer vision to create comprehensive digital replicas of construction sites.

A global collaboration to move artificial intelligence principles to practice

Convened by the MIT Schwarzman College of Computing, the AI Policy Forum will develop frameworks and tools for governments and companies to implement concrete policies.

Have You Heard of Neurosymbolic AI?

Artificial intelligence has been criticized for lacking common sense. Could a hybrid form of AI, merging deep neural networks and symbolic programs, change that? David Cox, IBM director of the MIT-IBM Watson AI Lab, weighs in.

Machine learning uncovers potential new TB drugs

Computational method for screening drug compounds can help predict which ones will work best against tuberculosis or other diseases.

Anticipating heart failure with machine learning

Many health issues are tied to excess fluid in the lungs. A new algorithm can detect the severity by looking at a single X-ray.

Milo Phillips-Brown receives inaugural MAC3 Society and Ethics in Computing Research Award

The award supports promising PhD candidates or postdocs conducting interdisciplinary research on the societal and ethical dimensions of computing.

Provably exact artificial intelligence for nuclear and particle physics

MIT-led team uses AI and machine learning to explore fundamental forces.

Toward a machine learning model that can reason about everyday actions

Researchers train a model to reach human-level performance at recognizing abstract concepts in video.

National Science Foundation announces MIT-led Institute for Artificial Intelligence and Fundamental Interactions

IAIFI will advance physics knowledge — from the smallest building blocks of nature to the largest structures in the universe — and galvanize AI research innovation.

Data systems that learn to be better

Storage tool developed at MIT CSAIL adapts to what its datasets’ users want to search.

Shrinking deep learning’s carbon footprint

Through innovation in software and hardware, researchers move to reduce the financial and environmental costs of modern artificial intelligence.

How the Coronavirus Pandemic Is Breaking AI and How to Fix It

Machine learning models leverage correlations between different variables to make predictions, but the pandemic has shown how ephemeral and context-driven these correlations can be. “Without an understanding of the underlying causes and effects that drive those correlations, your predictions will be wrong,” says David Cox, IBM director of the MIT-IBM Watson AI Lab.

Looking into the black box

Recent advances give theoretical insight into why deep learning networks are successful.

Have we squeezed as much out of deep learning as we can?

In a paper on the pre-print server arXiv, MIT and IBM researchers argue that it may no longer be economically or environmentally feasible to continue scaling deep learning systems.

How MIT and IBM Are Fighting Covid-19 with AI

David Cox, IBM Director of the MIT-IBM Watson AI lab, speaks with Tonya Hall about ten new projects MIT and IBM have launched to combat COVID-19.

Neurosymbolic AI to Give Us Machines With True Common Sense

Researchers want AI not to just recognize objects, but to be able to understand what it sees and apply reasoning to navigate new situations. In a project funded by DARPA, researchers at MIT, IBM, Harvard and Stanford aim to train computers to learn more like humans.

MIT researchers claim augmentation technique can train GANs with less data

A research team led by MIT's Song Han says it's developed a method that improves the efficiency of generative adversarial networks (GANs) by augmenting both real and fake data samples.

Cynthia Breazeal named Media Lab associate director

Expert in personal social robots will work with lab faculty and researchers to develop strategic research initiatives, and to explore new funding mechanisms.

What jumps out in a photo changes the longer we look

Researchers capture our shifting gaze in a model that suggests how to prioritize visual information based on viewing duration.

Engineers put tens of thousands of artificial brain synapses on a single chip

The design could advance the development of small, portable AI devices.

Giving soft robots feeling

In a pair of papers from MIT CSAIL, two teams enable better sense and perception for soft robotic grippers.

Undergraduates develop next-generation intelligence tools

UROP students explore applications in robotics, health care, language understanding, and nuclear engineering.

The World Has Changed. AI Is Struggling to Cope.

One way to close the gap between AI's promise and its usefulness for business is to make its decision-making process explainable, says David Cox, IBM director of the MIT-IBM Watson AI Lab. "If people don't understand or trust those tools, it's going to be a lost cause."

Study finds stronger links between automation and inequality

Job-replacing tech has directly driven the income gap since the late 1980s, economists report.

Robots help some firms, even while workers across industries struggle

Study finds manufacturing companies that are quick to automate can thrive, but overall employment drops.

How many jobs do robots really replace?

MIT economist Daron Acemoglu’s new research puts a number on the job costs of automation.

A foolproof way to shrink deep learning models

​Researchers unveil a pruning algorithm to make artificial intelligence applications run faster.

A Conversation with the New Computing Dean: Alumnus Daniel Huttenlocher

The MIT Technology Review sat down with Daniel Huttenlocher, the first dean of the MIT Schwarzman College of Computing, to find out what it’s been like to return to campus as an alum in order to lead one of the greatest structural changes in MIT’s history.

MIT researchers release Clevrer to advance visual reasoning and neurosymbolic AI

A new dataset released by researchers at MIT, Harvard and IBM at the International Conference on Learning Representations is aimed at training machines to understand causal relationships among objects.

MIT presents AI frameworks that compress models and encourage agents to explore

In a pair of papers accepted to the International Conference on Learning Representations, MIT researchers investigate new ways to motivate software agents to explore their environment and they unveil a new compression method for making AI models run faster.

The status of artificial adversarial intelligence

Journalist Tonya Hall speaks with MIT's Una-May O'Reilly about the current state of artificial adversarial intelligence and the strengths of adversarial dynamics over model-based malware detectors.

Muscle signals can pilot a robot

CSAIL's Conduct-A-Bot system uses muscle signals to cue a drone’s movement, enabling more natural human-robot communication.

MIT aims for energy efficiency in AI model training

In a new study, researchers at MIT and IBM propose a system for training and running AI models that saves energy and produces fewer carbon emissions.

Reducing the carbon footprint of artificial intelligence

MIT system cuts the energy required for training and running neural networks.

AI Could Save the World, If It Doesn’t Ruin the Environment First

The growing use of energy-intensive AI applications has led to a push for smaller, more efficient AI models. The MIT Quest for Intelligence and MIT-IBM Watson AI Lab recently sponsored a "Green AI" hackathon to spur innovation in this area.

Reducing delays in wireless networks

Congestion control system could help streaming video, mobile games, and other applications run more smoothly.

Computers Already Learn From Us. But Can They Teach Themselves?

The future of AI depends on systems that learn on their own, with little to no human supervision. David Cox, IBM director of the MIT-IBM Watson AI Lab, offers ideas for how to get there.

Learning about artificial intelligence: A hub of MIT resources for K-12 students

New website offers a combination of learning units, hands-on activities, and mentor guides to foster AI literacy.

AI Is Changing Work — and Leaders Need to Adapt

Working with AI doesn't necessarily mean people will lose their jobs — but they will need to work differently. IBM chief economist Martin Fleming discusses the results of an MIT-IBM Watson AI Lab study on AI and the future of work.

Neural networks facilitate optimization in the search for new materials

Sorting through millions of possibilities, a search for battery materials delivered results in five weeks instead of 50 years.

“Inactive” pill ingredients could raise the dose of your medication

With help from artificial intelligence, researchers identify hidden power of vitamin A and ordinary chewing gum glaze.

If AI’s So Smart, Why Can’t It Grasp Cause and Effect?

Causal reasoning would be useful for almost any AI system. Here's how researchers at the MIT-IBM Watson AI Lab and their colleagues are trying to embed an intuitive sense of physics in machines.

The elephant in the server room

Catherine D'Ignazio's new book, Data Feminism, examines problems of bias and power that beset modern information.

A Hybrid AI Model Lets It Reason about the World’s Physics like a Child

A new dataset co-released by the MIT-IBM Watson AI Lab reveals just how bad AI is at reasoning — and suggests that a new hybrid approach might be the best way forward.

Machine learning picks out hidden vibrations from earthquake data

Technique may help scientists more accurately map vast underground geologic structures.

To self-drive in the snow, look under the road

Weather’s a problem for autonomous cars. MIT’s new system shows promise by using “ground-penetrating radar” instead of cameras or lasers.

New AI Curriculum Designed for Middle School Students

An open-source AI curriculum developed in professor Cynthia Breazeal's lab at MIT focuses on how AI systems are designed, how they can influence the public, and how AI may change the workplace.

AI could help design better drugs that don’t clash with other medication

A new system that can predict a proposed drug’s chemical structure could help prevent adverse drug interactions, a leading cause of patient death.

Hey Alexa! Sorry I fooled you …

MIT’s new system TextFooler can trick the types of natural-language-processing systems that Google uses to help power its search results, including audio for Google Home.

A college for the computing age

With the initial organizational structure in place, the MIT Schwarzman College of Computing moves forward with implementation.

Demystifying artificial intelligence

Doctoral candidate Natalie Lao wants to show that anyone can learn to use AI to make a better world.

Technique reveals whether models of patient risk are accurate

Computer scientists’ new method could help doctors avoid ineffective or unnecessarily risky treatments.

Top minds in machine learning predict where AI is going

AI experts look at recent trends and explain where the field is headed in 2020.

Bringing artificial intelligence and MIT to middle school classrooms

MIT researchers and collaborators have developed an open-source curriculum to teach young students about ethics and artificial intelligence.

Finding a good read among billions of choices

As natural language processing techniques improve, suggestions are getting speedier and more relevant.

When machine learning packs an economic punch

Study: After eBay improved its translation software, international commerce increased sharply.

The mind-bending confusion of ‘hammer on a bed’ shows computer vision is far from solved

The ObjectNet dataset compiled by scientists from MIT and IBM is testing the limits of AI vision.

An Obstacle Course to Make AI Better

MIT and IBM researchers have developed a new dataset aimed at improving how AI systems identify objects. "We don’t want them to only recognize what is very common," says MIT's Boris Katz. "We want [a robot] to recognize a chair that is upside down on the floor and not say it is a backpack."

Helping machines perceive some laws of physics

Model registers “surprise” when objects in a scene do something unexpected, which could be used to build smarter AI.

Predicting people’s driving personalities

System from MIT CSAIL sizes up drivers as selfish or selfless. Could this help self-driving cars navigate in traffic?

Visualizing an AI model’s blind spots

New tool highlights what generative models leave out when reconstructing a scene.

AI May Not Kill Your Job—Just Change It

Don't fear the robots, according to a report by the MIT-IBM Watson AI Lab. Worry about algorithms replacing any task that can be automated.

Robust Robotics and MIT’s Quest for Intelligence

Nick Roy, director of the Bridge at MIT Quest, explains how cognitive science pushes artificial intelligence, enabling new engineering tools and services, and speaks about the importance of explainable and ethical AI.

Self-transforming robot blocks jump, spin, flip, and identify each other

Developed at MIT’s Computer Science and Artificial Intelligence Laboratory, robots can self-assemble to form various structures with applications including inspection.

Helping autonomous vehicles see around corners

By sensing tiny changes in shadows, a new system identifies approaching objects that may cause a collision.

Supercomputer analyzes web traffic across entire internet

Modeling web traffic could aid cybersecurity, computing infrastructure design, Internet policy, and more.

This Technique Can Make It Easier for AI to Understand Videos

A research team led by MIT's Song Han is teaching artificial intelligence to process more videos while using less power, potentially making it easier to apply AI to large amounts of video.

Faster video recognition for the smartphone era

MIT and IBM researchers offer a new method to train and run deep learning models more efficiently.

Lincoln Laboratory’s new artificial intelligence supercomputer is the most powerful at a university

TX-GAIA is tailor-made for crunching through deep neural network operations.

Computing and the search for new planets

MIT planetary scientists partner with computer scientists to find exoplanets.

Using machine learning to estimate risk of cardiovascular death

CSAIL system uses a patient's ECG signal to estimate potential for cardiovascular death.

IBM gives artificial intelligence computing at MIT a lift

Nearly $12 million machine will let MIT researchers run more ambitious AI models.

Boosting computing power for the future of particle physics

Prototype machine-learning technology co-developed by MIT scientists speeds processing by up to 175 times over traditional methods.

DeepMind’s Losses Display the Challenges of the AI Industry

Industry-academia collaborations like the MIT-IBM Watson AI Lab are one way of attempting to make AI tools and technologies accessible to all.

Guided by AI, robotic platform automates molecule manufacture

New system could free bench chemists from time-consuming tasks, may help inspire new molecules.

Professor Patrick Winston, former director of MIT’s Artificial Intelligence Laboratory, dies at 76

Beloved professor conducted pioneering research on imbuing machines with human-like intelligence, including the ability to understand stories.

Professor Emeritus Fernando Corbató, MIT computing pioneer, dies at 93

Longtime MIT professor developed early “time-sharing” operating systems and is widely credited as the creator of the world’s first computer password.

GAN Paint Studio uses AI to add, delete, and modify objects in photos

A new GAN-painting tool, developed at the MIT-IBM Watson AI Lab, takes a common-sense approach to adding objects and scenery to photos.

Teaching artificial intelligence to create visuals with more common sense

An MIT/IBM system could help artists and designers make quick tweaks to visuals while also helping researchers identify “fake” images.

New AI programming language goes beyond deep learning

General-purpose language works for computer vision, robotics, statistics, and more.

Translating proteins into music, and back

By turning molecular structures into sounds, researchers gain insight into protein structures and create new variations.

Want to learn how to train an artificial intelligence model? Ask a friend.

MIT Machine Intelligence Community introduces students to nuts and bolts of machine learning.

Algorithm tells robots where nearby humans are headed

A new tool for predicting a person’s movement trajectory may help humans and robots work together in close proximity.

Cracking open the black box of automated machine learning

Interactive tool lets users see and control how automated model searches work.

Bringing human-like reasoning to driverless car navigation

Autonomous control system “learns” to use simple maps and image data to navigate new, complex routes.

Using AI to predict breast cancer and personalize care

MIT/MGH's image-based deep learning model can predict breast cancer up to five years in advance.

Putting vision models to the test

Study shows that artificial neural networks can be used to drive brain activity.

Liberty Mutual Insurance joins MIT’s Quest for Intelligence

Company announces $25 million, five-year collaboration.

For better deep neural network vision, just add feedback (loops)

The DiCarlo lab finds that a recurrent architecture helps both artificial intelligence and our brains to better identify objects.

Vivienne Sze wins Edgerton Faculty Award

EECS faculty member is recognized for technical innovation, educational excellence, and efforts to advance women and under-represented minorities in her field.

3 Questions: Provost Martin Schmidt on building a new college

Working groups of faculty, students, staff are meeting regularly to develop ideas for the MIT Schwarzman College of Computing.

Kicking neural network automation into high gear

Algorithm designs optimized machine-learning models up to 200 times faster than traditional methods.

“Particle robot” works as a cluster of simple units

Loosely connected disc-shaped “particles” can push and pull one another, moving en masse to transport objects.

Computing the future

Fireside chat brings together six Turing Award winners to reflect on their field and the MIT Stephen A. Schwarzman College of Computing.

The future of education in a world of pervasive computing

Conference celebrating new MIT Schwarzman College of Computing explores the changing face of higher ed.

Taking the lead in shaping the future of computing and artificial intelligence

Popular expo highlights student creativity and ambition as celebration of the MIT Schwarzman College of Computing gets underway.

Student challenges kick off celebration of MIT Stephen A. Schwarzman College of Computing

“We see an incredible future where the College of Computing channels the collaborative potential of disciplines,” says grad student Matthew Claudel.

Exploring the nature of intelligence

Undergraduate research projects show how students are advancing research in human and artificial intelligence, and applying intelligence tools to other disciplines.

Using artificial intelligence to engineer materials’ properties

New system of “strain engineering” can change a material’s optical, electrical, and thermal properties.

3Q: Hamsa Balakrishnan on reimagining the future of transportation

MIT associate professor of aeronautics and astronautics describes the seamless flow of people, things, and materials.

Peering under the hood of fake-news detectors

Study uncovers language patterns that AI models link to factual and false articles; underscores need for further testing.

MIMIC Chest X-Ray database to provide researchers access to over 350,000 patient radiographs

A new database of images could pave a path for algorithmic models that ensure accurate diagnoses of conditions like pneumonia.

Engineers program marine robots to take calculated risks

Algorithm could help autonomous underwater vehicles explore risky but scientifically-rewarding environments.

Identifying artificial intelligence “blind spots”

Model identifies instances when autonomous systems have learned from examples that may cause dangerous errors in the real world.

Building site identified for MIT Stephen A. Schwarzman College of Computing

Headquarters would replace Building 44, forming an “entrance to computing” near the intersection of Vassar and Main streets.

3Q: Aleksander Madry on building trustworthy artificial intelligence

A recent MIT symposium explores methods for making artificial intelligence systems more reliable, secure, and transparent.

Deep-learning technique reveals “invisible” objects in the dark

Method could illuminate features of biological tissues in low-exposure images.

I think, therefore I code

Senior Jessy Lin, a double major in EECS and philosophy, is programming for social good.

Artificial intelligence summit adresses impact of technology on jobs and global economy

Speakers at the summit included Massachusetts Secretary of Labor Rosalin Acosta and former Google chairman Eric Schmidt.

The many interfaces of computing

A roundup of MIT student research projects offers a glimpse of where computing is going next.

Machines that learn language more like kids do

Computer model could improve human-machine interaction, provide insight into how children learn language.

Model paves way for faster, more efficient translations of more languages

New system may open up the world’s roughly 7,000 spoken languages to computer-based translation.

How should autonomous vehicles be programmed?

Massive global survey reveals ethics preferences and regional differences.

Tang family gift supports MIT.nano, MIT Quest for Intelligence

$20 million gift names imaging suite in new facility, creates catalyst fund.

Automated system identifies dense tissue, a risk factor for breast cancer, in mammograms

Deep-learning model has been used successfully on patients, may lead to more consistent screening procedures.

Ideas abound at Quest for Intelligence workshop

Community event generates ideas for sparking innovative and ambitious plans to advance research in human and machine intelligence.

Software finds the best way to stick a Mars landing

Program users can tinker with landing and path planning scenarios to identify optimal landing sites for Mars rovers.

Helping computers fill in the gaps between video frames

Machine learning system efficiently recognizes activities by observing how objects change in only a few key frames.

MIT undergraduates to embark on The Quest

Up to 100 Quest-funded UROP projects aim to crack the code of human and machine intelligence.

MIT-SenseTime Alliance funds projects from all five schools

Interdisciplinary work will advance research in human and machine intelligence.

MIT announces leadership of its Quest for Intelligence

Faculty from across the Institute tapped to lead new initiative in human and machine intelligence.

AI-based method could speed development of specialized nanoparticles

Neural network could expedite complex physics simulations.

Artificial intelligence in action

At the MIT-IBM Watson AI Lab, researchers are training computers to recognize dynamic events.

MIT and SenseTime announce effort to advance artificial intelligence research

Alliance will be part of new MIT Intelligence Quest.

New MIT-IBM Watson AI Lab: 5 things to know

IBM plans to make a 10-year, $240 million investment to create the MIT–IBM Watson AI Lab in partnership with MIT, where fundamental AI research will be conducted to unlock the potential of AI. Here are 5 key things to know about the new Lab.