Learning algorithm rewarded for building confidence over time
Researchers at Disney Research and Boston University have found that a machine learning program can be trained to detect human activity in a video sooner and more accurately than other methods by rewarding the program for gaining confidence in its prediction the longer it observes the activity.
It seems intuitive that the program would grow more confident that it is detecting, say, a person changing a tire, the longer it observes the person loosening lugnuts, jacking up the car and subsequently removing the wheel, but that’s not the way most computer models have been trained to detect activity, said Leonid Sigal, senior research scientist at Disney Research.
“Most training techniques are happy if the computer model gets 60 percent of the video frames correct, even if the errors occur late in the process, when the activity should actually be more apparent,” Sigal said. “That doesn’t make much sense. If the model predicts a person is making coffee even after it sees the person put pasta into boiling water, it should be penalized more than if it made the same incorrect prediction when the person was still just boiling water.”
Shugao Ma, a Ph.D. student in computer science at Boston University and a former intern at Disney Research, found that this change in training methods resulted in more accurate predictions of activities. The computer also was often able to accurately predict the activity early in the process, even after seeing only 20 to 30 percent of the video. Likewise, the program can detect that an activity is finished if its confidence that it is observing that activity begins to drop.
A novel, inexpensive method for detecting the Zika virus could help slow spread of outbreak, and potentially other future pandemic diseases
An international, multi-institutional team of researchers led by synthetic biologist James Collins, Ph.D., at the Wyss Institute for Biologically Inspired Engineering at Harvard University, has developed a low-cost, rapid paper-based diagnostic system for strain-specific detection of the Zika virus, with the goal that it could soon be used in the field to screen blood, urine, or saliva samples.
“The growing global health crisis caused by the Zika virus propelled us to leverage novel technologies we have developed in the lab and use them to create a workflow that could diagnose a patient with Zika, in the field, within 2-3 hours,” said Collins, who is a Wyss Core Faculty member, and Termeer Professor of Medical Engineering & Science and Professor of Biological Engineering at the Massachusetts Institute of Technology (MIT)’s Department of Biological Engineering.
Building off previous work done at Harvard’s Wyss Institute by Collins and his team, along with collaborators from Massachusetts Institute of Technology (MIT), the Broad Institute of Harvard and MIT, Harvard Medical School (HMS), University of Toronto, Arizona State University (ASU), University of Wisconsin-Madison (UW-Madison), Boston University (BU), Cornell University, and Addgene, joined their efforts to quickly prototype the rapid diagnostic test and describe their methods in a study published online May 6 in the journal Cell, all within a matter of six weeks. Collins is the paper’s corresponding author.
Emerging innovation during the Ebola health crisis
In October 2014, Collins’ team developed a breakthrough method for embedding synthetic gene networks — which could be used as programmable diagnostics and sensors – on portable, small discs of ordinary paper.
Boston University (most commonly referred to as BU) is a private, nonsectarian research university located in Boston, Massachusetts.
The university has more than 3,800 faculty members and 33,000 students, and is one of Boston’s largest employers. It offers bachelor’s degrees, master’s degrees, and doctorates, and medical, dental, business, and law degrees through eighteen schools and colleges on two urban campuses. The main campus is situated along the Charles River in Boston’s Fenway-Kenmore and Allston neighborhoods, while the Boston University Medical Campus is in Boston’s South End neighborhood. BU also operates 75 study abroad programs in over 33 cities in over twenty countries and has internship opportunities in ten different countries (including the United States).
BU is categorized as an RU/VH Research University (very high research activity) in the Carnegie Classification of Institutions of Higher Education. In 2009–2010, BU had research expenditures of $407.8 million, or $553 million if the research led by the Medical School faculty at Boston Medical Center is included. BU is a member of the Boston Consortium for Higher Education and the Association of American Universities.
The Latest Updated Research News:
Boston University research articles from Innovation Toronto
- Finding Zika one paper disc at a time in 2 to 3 hours – May 7, 2016
- A programming language for living cells with many applications – April 2, 2016
- Researchers Develop Techniques to Bypass Blood-Brain Barrier, Deliver Drugs to Brain and Nervous System – October 24, 2015
- Can light therapy help with brain injuries and PTSD? – April 4, 2015
- Starscraper aims at cheaper, gentler suborbital rocket – December 31, 2014
- Bionic pancreas may revolutionize diabetes treatment – June 16, 2014
- Privately-funded science university plan
- Treepods air-scrubbers could clean up Boston
- Water-Resistant Hydrophobic Materials Breakthrough Makes a Splash
- Computer-Brain Interfaces Making Big Leaps
- Dodging antibiotic side effects
- Algorithm Finds Best Routes for One-Way Car Sharing
- New fiber optic technology could ease Internet congestion, video streaming
- Researchers Identify Novel Therapy to Treat Muscular Dystrophy
- A shot in the arm for old antibiotics
- Oil for the Joints: Grinstaff Advances New Osteoarthritis Treatment
- Researchers discover mushrooms can provide as much vitamin D as supplements
- Bacterial byproduct offers route to avoiding antibiotic resistance
- Should business be allowed to patent mathematics?
- Scientists Notch a Win in War Against Antibiotic-Resistant Bacteria
- The Meteoric Ascent of the Patent Troll and the Devastating Consequences for Innovation
- Crowd-sourcing brain research leads to breakthrough
- Matrix-style instant learning could be one step closer
- Which Technologies Get Better Faster?
- New biosensor quickly detects viruses such as Ebola
- Plant-based vaccine factory enables production in just weeks
- Non-invasive infrared scanner to detect deadly melanoma
- Creating Wireless Network Using Visible Light
- VIDEO: Knife-Wielding Robot Trains for Grocery Checkout Job Using New Coactive Learning Technique
- Pentagon-funded Atlas robot refuses to be knocked over
- Researchers Use Machine Learning to Boil Down the Stories that Wearable Cameras Are Telling
- Cross-Disciplinary Team From Harvard and Dana-Farber Brings Novel Therapeutic Cancer Vaccine to Human Clinical Trials
- Research Creates New Opportunities From Waste Heat
- Lifelike cooling for sunbaked windows
- New Software Gives Robots the Gift of Hearing
- Will the Robot Uprising Be Squishy?
- Stem cell research yields Down syndrome breakthrough
- Injectable ‘Smart Sponge’ Holds Promise for Controlled Drug Delivery
- A new form of carbon: Grossly warped ‘nanographene’
- DARPA’s ATLAS humanoid robot gears up for disaster response
- Carnegie Mellon Method Uses Network of Cameras to Track People in Complex Indoor Settings
- Facial Recognition Technology Proves Itself
- Iowa State Professor Weighs Benefit vs. Risk of Facial Recognition Technology
- Injectable Nano-Network Controls Blood Sugar in Diabetics for Days at a Time
- Vitamin E Identified as Potential Weapon Against Obesity
- Low-Dose Aspirin Stymies Proliferation of Two Breast Cancer Lines
- Fish Gill-Inspired Design Wins 1st Prize in Biomimicry Design Competition
- Could a computer on the police beat prevent violence?
- Scientists Detail Severe Future Impacts of Climate Change
- The World’s New Second Fastest Robot Is A Tiny Cardboard Cockroach
- Synthetic Biofilter Removes Estrogens and other Medicine Residues from Drinking Water
- “Lung-on-a-chip” sets stage for next wave of research to replace animal testing
- Eco-friendly Optics: Spider Silk’s Hidden Talents Brought to Light for Applications in Biosensors, Lasers, Microchips
- Researchers grow cyborg tissues with embedded nanoelectronics
- U.S. Army Recruiting an Array of Animal-Inspired Robots to Assist Battlefield Troops
- Exotic Material Boosts Electromagnetism Safely
- Let the games begin
- ‘Fantastic Voyage’ Through the Body, With Precision Control
- Far From Any Lab, Paper Bits Find Illness
- Game theory in practice
- Free the H-1Bs, Free the Economy
- Danish Eco City Proves Waste Management Can Reverse Greenhouse Trend
- Anonymity and the Dark Side of the Internet
- Chile’s Grand Innovation Experiment
- Household Robots Do Not Protect Users’ Security And Privacy, Researchers Say
- The Silk Renaissance
- Researchers regenerate nerve connections after spinal injury
- Robot Ranger Sets Untethered ‘Walking’ Record at 14.3 Miles
- Manipulation of the Crowd: How Trustworthy Are Online Ratings?
- Enterprise PCs Work While They Sleep – Saving Energy and Money – With New Software
- Leading Lights
- Robot Pack Mule to Carry Loads for G.I.s on the Move
- Cell Phones Become Handheld Tools For Global Development
New language lets researchers design novel biological circuits
MIT biological engineers have devised a programming language that can be used to give new functions to E. coli bacteria. Using this language, anyone can write a program for the function they want, such as detecting and responding to certain environmental conditions. They can then generate a DNA sequence that will achieve it.
Another advantage of this technique is its speed. Until now, “it would take years to build these types of circuits. Now you just hit the button and immediately get a DNA sequence to test,” Voigt says.