116
#1
One of the biggest bottlenecks when adopting new technologies is trust in AI. Now, a new tool generates automatic indicators if data and predictions generated by AI algorithms are trustworthy.
Neural networks are a type of artificial intelligence that is modeled after the brain and generate predictions. But how we can tell these predictions are trusted? One of the key barriers to the adoption of self-driving cars is that the vehicles need to act as independent decision-makers in the road. Can we trust the computer software within the vehicles to make sound decisions within fractions of a second - especially when conflicting information is coming from computer vision from cameras or data from lidar? Even humans can be indecisive in certain cases involving conflicting information.
Knowing which systems to trust and is most accurate would be helpful to determine what decisions
the autopilot should make (Photo: aitrends)
A new tool developed by USC Viterbi Engineering researchers named DeepTrust can quantify the amount of uncertainty, should human intervention is necessary. It took the research team almost two years to employ what is known as subjective logic to assess the architecture of the neural networks. On one of their test cases, DeepTrust found that the prediction pointing towards Clinton winning in 2016 Presidential election had a greater margin for error.
The other significance of this research is that it provides insights on how to test the reliability of AI algorithms that are normally trained on thousands to millions of data points. In other way, how to make sure the architecture of these neural network systems has greater accuracy. This work could also serve as a guidepost as to how much "noise" can be in testing samples if computer scientists want to maximize accuracy and trust.
This model is the first of its kind. To our knowledge, there is no trust quantification model or tool for deep learning, artificial intelligence, and machine learning. DeepTrust has the potential to make AI aware and adaptive.
Source: ScienceDaily
About us: TMA Solutions was established in 1997 to provide quality software outsourcing services to leading companies worldwide. We are one of the largest software outsourcing companies in Vietnam with 2,500 engineers.
Visit us at TMA Solutions | Leading Vietnam Software Outsourcing company
Neural networks are a type of artificial intelligence that is modeled after the brain and generate predictions. But how we can tell these predictions are trusted? One of the key barriers to the adoption of self-driving cars is that the vehicles need to act as independent decision-makers in the road. Can we trust the computer software within the vehicles to make sound decisions within fractions of a second - especially when conflicting information is coming from computer vision from cameras or data from lidar? Even humans can be indecisive in certain cases involving conflicting information.
Knowing which systems to trust and is most accurate would be helpful to determine what decisions
the autopilot should make (Photo: aitrends)
The other significance of this research is that it provides insights on how to test the reliability of AI algorithms that are normally trained on thousands to millions of data points. In other way, how to make sure the architecture of these neural network systems has greater accuracy. This work could also serve as a guidepost as to how much "noise" can be in testing samples if computer scientists want to maximize accuracy and trust.
This model is the first of its kind. To our knowledge, there is no trust quantification model or tool for deep learning, artificial intelligence, and machine learning. DeepTrust has the potential to make AI aware and adaptive.
Source: ScienceDaily
About us: TMA Solutions was established in 1997 to provide quality software outsourcing services to leading companies worldwide. We are one of the largest software outsourcing companies in Vietnam with 2,500 engineers.
Visit us at TMA Solutions | Leading Vietnam Software Outsourcing company
