Trending...
- Submissions of Short Films & Videos are Needed for the 10th Annual Indigenous Film Retreats - 174
- Boomin Reunion Band Itsy Bitsy Teenie Weenie Yellow Polka Dot Bikini Has 10 Million You Tube Streams - 162
- Sunmed | Your CBD Store Responds to Governor Newsom's Industry-Killing Hemp Regulation
LOS ANGELES - Californer -- Uncertainty estimation is critical to improving the reliability of deep neural networks. A research team led by Aydogan Ozcan at the University of California, Los Angeles, has introduced an uncertainty quantification method that uses cycle consistency to enhance the reliability of deep neural networks in solving inverse imaging problems.
This research was published on Dec. 21 in Intelligent Computing, a Science Partner Journal.
Deep neural networks have been used to solve inverse imaging problems, such as image denoising, super-resolution imaging and medical image reconstruction, in which the goal is to create an ideal image using the raw image data actually captured, often after some degradation. However, deep neural networks sometimes produce unreliable results, and in some contexts, incorrect predictions can have severe consequences. Models that can quantitatively estimate how certain they are about their output can perform better at detecting abnormal situations, such as anomalous data and outright attacks.
The new method for estimating network uncertainty uses a physical forward model, which serves as a computational representation of the underlying processes governing the input–output relationship. By combining this model with a neural network and executing forward–backward cycles between the input and output data, uncertainty accumulates and can be effectively estimated.
More on The Californer
The theoretical underpinning of the method involves establishing the bounds of cycle consistency, defined as the difference between adjacent outputs in the cycle. The researchers derived upper and lower bounds for cycle consistency, demonstrating its relationship with the uncertainty of the output of the neural network. The study considered cases where cycle outputs diverged and cases where they converged, providing expressions for both scenarios. The derived bounds can be used to estimate uncertainty even without knowledge of the ground truth.
The researchers anticipate that their cycle-consistency-based uncertainty quantification method will significantly contribute to enhancing the reliability of neural network inferences in inverse imaging problems. Additionally, the method could find applications in uncertainty-guided learning. This study marks a significant step toward addressing the challenges associated with uncertainty in neural network predictions, paving the way for more reliable and confident deployment of deep learning models in critical real-world applications.
Article: https://spj.science.org/doi/10.34133/icomputing.0071
This research was published on Dec. 21 in Intelligent Computing, a Science Partner Journal.
Deep neural networks have been used to solve inverse imaging problems, such as image denoising, super-resolution imaging and medical image reconstruction, in which the goal is to create an ideal image using the raw image data actually captured, often after some degradation. However, deep neural networks sometimes produce unreliable results, and in some contexts, incorrect predictions can have severe consequences. Models that can quantitatively estimate how certain they are about their output can perform better at detecting abnormal situations, such as anomalous data and outright attacks.
The new method for estimating network uncertainty uses a physical forward model, which serves as a computational representation of the underlying processes governing the input–output relationship. By combining this model with a neural network and executing forward–backward cycles between the input and output data, uncertainty accumulates and can be effectively estimated.
More on The Californer
- RideBoom Seeks Electric Vehicle Manufacturers to Join Forces in Redefining Last-Mile Transportation
- Fernanda Rocha Launches New YouTube Fitness Channel, Empowering Busy Moms to Prioritize Their Health
- AI-based Tissue Staining to Detect Amyloid Deposits Without Chemical Stains or Microscopy
- California: Governor Newsom welcomes the largest CHP graduating class in over two years
- California's earthquake warning system notified millions ahead of today's Southern California quake
The theoretical underpinning of the method involves establishing the bounds of cycle consistency, defined as the difference between adjacent outputs in the cycle. The researchers derived upper and lower bounds for cycle consistency, demonstrating its relationship with the uncertainty of the output of the neural network. The study considered cases where cycle outputs diverged and cases where they converged, providing expressions for both scenarios. The derived bounds can be used to estimate uncertainty even without knowledge of the ground truth.
The researchers anticipate that their cycle-consistency-based uncertainty quantification method will significantly contribute to enhancing the reliability of neural network inferences in inverse imaging problems. Additionally, the method could find applications in uncertainty-guided learning. This study marks a significant step toward addressing the challenges associated with uncertainty in neural network predictions, paving the way for more reliable and confident deployment of deep learning models in critical real-world applications.
Article: https://spj.science.org/doi/10.34133/icomputing.0071
Source: UCLA ITA
Filed Under: Science
0 Comments
Latest on The Californer
- Long Beach Seeks Volunteers for 40th Annual California Coastal Cleanup Day, September 21
- Vantiva Wins CSI Award for "Best Sustainability Project or Initiative" for Eco-Friendly V7c Android TV Set Top Box
- Skool O' Foolz Drops Rapperz Delite on Gum Drop Hip Hop You Tube Channel From TVS Records
- Ketamine Lead Gen Makes Major Moves in Mental Health and Wellness Market Expansion
- Inframark Announces New Chief Financial Officer
- Hayloft Big Creek Now Pre-Leasing
- SBUX Investors Have Opportunity to Join Starbucks Corporation Fraud Investigation with the Schall Law Firm
- Unique New African American Hosted Classical Music Show Launched
- California: Governor Newsom issues legislative update 9.12.24
- City of Long Beach Launches Backyard Builders Pilot Program
- Green Office Partner Ranked #1 on Crain's 2024 List of Best Places to Work in Chicago
- Snell & Wilmer Welcomes New Associate Tara Giery in the San Diego Office
- Alexandria Real Estate Equities, Inc. Closes Strategic Value-Harvesting Disposition of 1165 Eastlake Avenue East in the Lake Union Submarket of Seattle
- Illuminate Your Spiritual Journey with "Living in Divine Incandescence" by Dr. Angela Butts Chester
- STEED'S Unveils the Ultimate Multi-Bottle Feeder for Your Fur Babies
- Alexandria Bedolla: An Award-Winning Baker with a Passion for Sweets
- JRK Property Holdings Acquires Apartment Communities in San Francisco and Minneapolis
- Long Beach City Council Adopts Fiscal Year 2025 Budget
- New Book Promises to Revolutionize Mindsets and Transform Professional Success for Latinas
- California: Governor Newsom signs "smash-and-grab" bill to strengthen felony sentences for large-scale theft offenses