/>
X

The most exciting, innovative MIT projects in 2016

It's been an exciting year for the MIT CSAIL lab with ingestible origami robots, AI acting as oracles and cyberattack detection research only the tip of the iceberg.

|
charlie-osborne.jpg
|
Topic: Innovation
mit1.jpg
1 of 9 Charlie Osborne/ZDNet

Ingestible origami robots

MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) has been busy this year with advances in computer science, artificial intelligence, networking, robotics and space taking center stage.

One project that caught our attention this year is the lab's ingestible "origami" robots, which once swallowed, can move through human digestive systems to perform tasks such as patching up wounds or administering drugs -- potentially reducing the need for invasive surgeries.

mit2.jpg
2 of 9 Bill Stafford - NASA - JSC

NASA's Valkyrie delivery

In May, MIT lent a helping hand to NASA to assist in making the space agency's Valkyrie robot perform tasks autonomously.

MIT accepted a delivery of a Valkyrie -- a six-foot-tall, 300-pound human-like robot -- in 2016, alongside a $250,000 cash injection to make the robot more suitable for long-haul missions.

mit3.jpg
3 of 9 Jason Dorfman, MIT CSAIL

3D-printed robots with soft skins

Robots are now used in every industry from manufacturing to the military but are also prone to breaking if components are damaged through accidental drops. To combat this, MIT developed a new 3D-printing technique which gives robots soft skins that make robots "safer, more resilient and more precise in their movements."

mit4.jpg
4 of 9 Kalyan Veeramachaneni/MIT CSAIL

An AI platform which detects most cyberattacks

Today's cybersecurity professionals face the challenge of escalating cyberattacks and the increasing frequency of threats. Simply keeping an eye on what is happening around a corporate network can be hard enough, but MIT hopes to take some of the burden away with the development of a new AI system which is able to use algorithms and predictive technologies to detect up to "85 percent" of cyberattacks.

mit5.jpg
5 of 9 Jason Dorfman, MIT CSAIL

Designing your own drones

Companies looking to implement drones in their businesses are limited to very specific designs, capacities, and speeds in today's market.

However, MIT has developed a new program which takes the guesswork out of building custom drones -- whether for carrying specific weights, accessories or reaching faster speeds for longer -- which may allow for the creation of custom drones which retain full functionality.

mit6.jpg
6 of 9 M.Weiss/NASA/CXC

A new method to image black holes

Black holes have remained a fascinating concept for scientists and the many members of the general public, but learning more about this phenomenon -- in which a gravity-heavy region of space does not allow light to escape -- is difficult.

To improve our understanding of black holes, MIT researchers launched a new project this year to bring together all of our telescopic data and algorithms to create the first "true" image of a black hole.

mit7.jpg
7 of 9 Christine Daniloff/MIT

Making computers explain themselves

An important breakthrough for artificial intelligence and machine learning, MIT researchers submitted fresh research in 2016 on how neural networking can be used to make AI systems explain the decisions they make; effectively acting as "black boxes" which can further advance AI models.

mit8.jpg
8 of 9 MIT

Slicing webpage load times by 34 percent

As consumer demand for data and streaming grows, the small amount of time it takes to load up a web page has also come under scrutiny. However, MIT has come up with a way to shave up to 30 percent off the time it takes to load the average page by fetching online files more efficiently.

mit9.jpg
9 of 9 Carl Vondrick/MIT CSAIL

Teaching machines to predict the future

There are situations in life when you instinctively know what is about to happen next -- but this can be given to AI, too.

This year, MIT revealed the development of an algorithm which allows AI to predict human interaction based on YouTube videos without the advantage of experience, which one day, could be expanded for applications such as military use or alerting emergency responders.

Related Galleries

First Look: Samsung's Galaxy Z Flip 3 [pictures]
01-001ef-gf711t-001-front-navy-210719-h.jpg

Related Galleries

First Look: Samsung's Galaxy Z Flip 3 [pictures]

26 Photos
Remote working, programming languages, artificial intelligence, and more: ZDNet's research roundup
gartnerhypeai2020.png

Related Galleries

Remote working, programming languages, artificial intelligence, and more: ZDNet's research roundup

10 Photos
Pitfalls to Avoid when Interpreting Machine Learning Models
1-bad-model-generalization.jpg

Related Galleries

Pitfalls to Avoid when Interpreting Machine Learning Models

8 Photos
Toshiba laptops through the ages
screenshot-2020-08-11-at-11-52-27.png

Related Galleries

Toshiba laptops through the ages

11 Photos
Innovative projects now online to combat coronavirus outbreak
7.jpg

Related Galleries

Innovative projects now online to combat coronavirus outbreak

7 Photos
Top programming languages, 5G worries, cloud computing, and more: Research round-up
gsa-status-of-5g-by-country-end-2019.jpg

Related Galleries

Top programming languages, 5G worries, cloud computing, and more: Research round-up

12 Photos
Samsung Galaxy S20 first look: All the models and colors up close
dsc-9870.jpg

Related Galleries

Samsung Galaxy S20 first look: All the models and colors up close

15 Photos