School of Engineering

To solve complex problems, data scientists must shepherd their raw data through a series of steps, each one requiring many human-driven decisions. The last step in the process, deciding on a modeling technique, is particularly crucial.

Auto-tuning data science: New research streamlines machine learning

A new automated machine-learning system performs as well or better than its human counterparts — and works 100 times faster.

December 19, 2017

Members of the MIT delegation to the 2017 UN Climate Change Conference gather outside the venue in Bonn, Germany (l-r): Jonas Knapp, Laur Hesse Fisher, Michael Casey, Kathleen Kennedy, Thomas Malone, Jessika Trancik, Stephen Lee, Morgan Edwards, Tom Kiley, and Erick Pinos.

Exploring ethical dimensions of climate negotiation

MIT researchers focus on empathy, transparency, and empowerment at this year's United Nations Climate Change Conference.

December 18, 2017

Class 2.680 (Unmanned Marine Vehicle Autonomy, Sensing and Communications), which is offered during spring semester, is structured around the presence of ice on the Charles. While the river is covered by a thick sheet of ice in February and into March, students are taught to code and program a remotely-piloted marine vehicle for a given mission.

Unlocking marine mysteries with artificial intelligence

Students put their AI software for underwater vehicles to the test on the Charles River.

December 14, 2017

A molten hanging droplet, or pendant (at center), forms from an aluminum oxide rod under intense heat and light from xenon lamps. Using this process, MIT researchers have demonstrated a new electrochemical method to study thermodynamic processes in a molten oxide melt at temperatures above 2,000 degrees Celsius.

Capturing the properties of very hot compounds

MIT researchers demonstrate a new electrochemical method to study thermodynamic processes in an ultra-high temperature molten oxide.

December 14, 2017

A Nov. 20 reception honored the MIT faculty receiving this year’s Professor Amar G. Bose Research Grants. The program is named for the late Amar Bose ’51, SM ’52, ScD ’56, a longtime MIT faculty member and the founder of the Bose Corporation. This year’s reception also honored his son, Vanu Bose ’87, SM ’94, PhD ’99, who passed away last month. In his opening remarks, President L. ...

Bose Grants for 2017 reward bold and unconventional research visions

Six potentially paradigm-shifting research projects will make strides with funding from Professor Amar G. Bose Research Grants.

December 14, 2017

As part of an investigation into the nature of humans' physical intuitions, MIT researchers trained a neural network to predict how unstably stacked blocks would respond to the force of gravity.

Computer systems predict objects’ responses to physical forces

Results may help explain how humans do the same thing.

December 13, 2017

More than 1,000 people packed MIT’s Kresge Auditorium for the 2.009 final product presentations, a major fall-semester event known not only for innovative products presented by students, but also for non-stop fun, music, and confetti.

Students launch products that help users harness their superpowers

In final presentation for mechanical engineering class, students unveil new products based on superhero theme.

December 13, 2017

Illumination of a book (“Paradise Lost,” by John Milton) with the nanobionic light-emitting plants (two 3.5-week-old watercress plants). The book and the light-emitting watercress plants were placed in front of a reflective paper to increase the influence from the light emitting plants to the book pages.

Engineers create plants that glow

Illumination from nanobionic plants might one day replace some electrical lighting.

December 12, 2017

Clockwise from top left: Shafi Goldwasser, Tomás Lozano-Pérez, Michael Sipser, and Silvio Micali.

Four from MIT named 2017 Association for Computing Machinery Fellows

Professors Goldwasser, Lozano-Perez, Micali, and Sipser honored for "providing key knowledge" to computing.

December 11, 2017

Neural nets are so named because they roughly approximate the structure of the human brain. Typically, they’re arranged into layers, and each layer consists of many simple processing units — nodes — each of which is connected to several nodes in the layers above and below. Data is fed into the lowest layer, whose nodes process it and pass it to the next layer. The connections between layers ...

Reading a neural network’s mind

Technique illuminates the inner workings of artificial-intelligence systems that process language.

December 10, 2017